[SURVEY] Remove Mesos support

[rmetzger0]

Hi all,

I wanted to discuss if it makes sense to remove support for Mesos in Flink. It seems that nobody is actively maintaining that component (except for necessary refactorings because of interfaces we are changing), and there are almost no users reporting issues or asking for features.

The Apache Mesos project itself seems very inactive: There has been only one message on the dev@ list in the last 3 months. 

In 2020, I found only 3 users mentioning that they are using Mesos on the user@ list.

Maybe it makes sense to add a prominent log warning into the Mesos code in the Flink 1.12 release, that we are planning to remove Mesos support. Users will then have enough opportunity to raise concerns or discuss with us.

Best,

Robert

[Konstantin Knauf-3]

Hi Robert, 

+1 to the plan you outlined. If we were to drop support in Flink 1.13+, we would still support it in Flink 1.12- with bug fixes for some time so that users have time to move on.

It would certainly be very interesting to hear from current Flink on Mesos users, on how they see the evolution of this part of the ecosystem.

Best, 

Konstantin

[Xintong Song]

+1 for adding a warning in 1.12 about planning to remove Mesos support.

On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:

Hi Robert,

+1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
would still support it in Flink 1.12- with bug fixes for some time so that
users have time to move on.

It would certainly be very interesting to hear from current Flink on Mesos
users, on how they see the evolution of this part of the ecosystem.

Best,

Konstantin

[Kostas Kloudas-2]

+1 for adding a warning about the removal of Mesos support and I would
also propose to state explicitly in the warning the version that we
are planning to actually remove it (e.g. 1.13 or even 1.14 if we feel
it is too aggressive).

This will help as a reminder to users and devs about the upcoming
removal and it will avoid future, potentially endless, discussions.

Cheers,
Kostas

On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:

>
> +1 for adding a warning in 1.12 about planning to remove Mesos support.
>
>
> With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
>
>
> Thank you~
>
> Xintong Song
>
>
>
> On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
>>
>> Hi Robert,
>>
>> +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
>> would still support it in Flink 1.12- with bug fixes for some time so that
>> users have time to move on.
>>
>> It would certainly be very interesting to hear from current Flink on Mesos
>> users, on how they see the evolution of this part of the ecosystem.
>>
>> Best,
>>
>> Konstantin

[Till Rohrmann]

Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.

Cheers,

Till

On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:

+1 for adding a warning in 1.12 about planning to remove Mesos support.

With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.

Thank you~

Xintong Song

On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:

Hi Robert,

+1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
would still support it in Flink 1.12- with bug fixes for some time so that
users have time to move on.

It would certainly be very interesting to hear from current Flink on Mesos
users, on how they see the evolution of this part of the ecosystem.

Best,

Konstantin

[Piyush Narang]

Hi folks,

 

We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)

 

Thanks,

 

-- Piyush

 

 

From: Till Rohrmann <[hidden email]>
Date: Friday, October 23, 2020 at 8:19 AM
To: Xintong Song <[hidden email]>
Cc: dev <[hidden email]>, user <[hidden email]>
Subject: Re: [SURVEY] Remove Mesos support

 

Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.

 

Cheers,

Till

 

On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:

+1 for adding a warning in 1.12 about planning to remove Mesos support.

 

With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.

 

Thank you~

Xintong Song

 

 

On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:

Hi Robert,

+1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
would still support it in Flink 1.12- with bug fixes for some time so that
users have time to move on.

It would certainly be very interesting to hear from current Flink on Mesos
users, on how they see the evolution of this part of the ecosystem.

Best,

Konstantin

[Kostas Kloudas-2]

Thanks Piyush for the message.
After this, I revoke my +1. I agree with the previous opinions that we
cannot drop code that is actively used by users, especially if it
something that deep in the stack as support for cluster management
framework.

Cheers,
Kostas

On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:

>
> Hi folks,
>
>
>
> We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
>
>
>
> Thanks,
>
>
>
> -- Piyush
>
>
>
>
>
> From: Till Rohrmann <[hidden email]>
> Date: Friday, October 23, 2020 at 8:19 AM
> To: Xintong Song <[hidden email]>
> Cc: dev <[hidden email]>, user <[hidden email]>
> Subject: Re: [SURVEY] Remove Mesos support
>
>
>
> Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
>
>
>
> Cheers,
>
> Till
>
>
>
> On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
>
> +1 for adding a warning in 1.12 about planning to remove Mesos support.
>
>
>
> With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
>
>
>
> Thank you~
>
> Xintong Song
>
>
>
>
>
> On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
>
> Hi Robert,
>
> +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
> would still support it in Flink 1.12- with bug fixes for some time so that
> users have time to move on.
>
> It would certainly be very interesting to hear from current Flink on Mesos
> users, on how they see the evolution of this part of the ecosystem.
>
> Best,
>
> Konstantin

[Piyush Narang]

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush
 

On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.
   
    Cheers,
    Kostas
   
    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin
   

[rmetzger0]

Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.

On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Lasse Nedergaard-2]

Hi

At Trackunit We have been using Mesos for long time but have now moved to k8s. 

Med venlig hilsen / Best regards

Lasse Nedergaard

Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.

On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Xintong Song]

Thanks for sharing the information with us, Piyush an Lasse.

On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <[hidden email]> wrote:

Hi

At Trackunit We have been using Mesos for long time but have now moved to k8s. 

Med venlig hilsen / Best regards

Lasse Nedergaard

Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.

On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Piyush Narang]

Hi Xintong,

 

Do you have any jiras that cover any of the items on 1 or 2? I can reach out to folks internally and see if I can get some folks to commit to helping out.

 

To cover the other qs:

• Yes, we’ve not got a plan at the moment to get off Mesos. We use Yarn for some our Flink workloads when we can. Mesos is only used when we need streaming capabilities in our WW dcs (as our Yarn is centralized in one DC)
• We’re currently on Flink 1.9 (old planner). We have a plan to bump to 1.11 / 1.12 this quarter.
• We typically upgrade once every 6 months to a year (not every release). We’d like to speed up the cadence but we’re not there yet.
• We’d largely be good with keeping Flink on Mesos as-is and functional while missing out on some of the newer features. We understand the pain on the communities side and we can take on the work if we see some fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in the request to port it over.

 

Thanks,

 

-- Piyush

 

 

From: Xintong Song <[hidden email]>
Date: Sunday, October 25, 2020 at 10:57 PM
To: dev <[hidden email]>, user <[hidden email]>
Cc: Lasse Nedergaard <[hidden email]>, <[hidden email]>
Subject: Re: [SURVEY] Remove Mesos support

 

Thanks for sharing the information with us, Piyush an Lasse.

 

@Piyush

 

Thanks for offering the help. IMO, there are currently several problems that make supporting Flink on Mesos challenging for us.

1. Lack of Mesos experts. AFAIK, there are very few people (if not none) among the active contributors in this community that are familiar with Mesos and can help with development on this component.
2. Absence of tests. Mesos does not provide a testing cluster, like `MiniYARNCluster`, making it hard to test interactions between Flink and Mesos. We have only a few very simple e2e tests running on Mesos deployed in a docker, covering the most fundamental workflows. We are not sure how well those tests work, especially against some potential corner cases.
3. Divergence from other deployment. Because of 1 and 2, the new efforts (features, maintenance, refactors) tend to exclude Mesos if possible. When the new efforts have to touch the Mesos related components (e.g., changes to the common resource manager interfaces), we have to be very careful and make as few changes as possible, to avoid accidentally breaking anything that we are not familiar with. As a result, the component diverges a lot from other deployment components (K8s/Yarn), which makes it harder to maintain.

It would be greatly appreciated if you can help with either of the above issues.

 

Additionally, I have a few questions concerning your use cases at Criteo. IIUC, you are going to stay on Mesos in the foreseeable future, while keeping the Flink version up-to-date? What Flink version are you currently using? How often do you upgrade (e.g., every release)? Would you be good with keeping the Flink on Mesos component as it is (means that deployment and resource management improvements may not be ported to Mesos), while keeping other components up-to-date (e.g., improvements from programming APIs, operators, state backens, etc.)? 

 

Thank you~

Xintong Song

 

 

On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <[hidden email]> wrote:

Hi

 

At Trackunit We have been using Mesos for long time but have now moved to k8s. 

Med venlig hilsen / Best regards

Lasse Nedergaard

 

Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.

 

On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Xintong Song]

Hi Piyush,

Thanks a lot for sharing the information. It would be a great relief that you are good with Flink on Mesos as is.

As for the jira issues, I believe the most essential ones should have already been resolved. You may find some remaining open issues here [1], but not all of them are necessary if we decide to keep Flink on Mesos as is.

At the moment and in the short future, I think helps are mostly needed on testing the upcoming release 1.12 with Mesos use cases. The community is currently actively preparing the new release, and hopefully we could come up with a release candidate early next month. It would be greatly appreciated if you fork as experienced Flink on Mesos users can help with verifying the release candidates. 

[1] https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open

On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]> wrote:

Hi Xintong,

 

Do you have any jiras that cover any of the items on 1 or 2? I can reach out to folks internally and see if I can get some folks to commit to helping out.

 

To cover the other qs:

• Yes, we’ve not got a plan at the moment to get off Mesos. We use Yarn for some our Flink workloads when we can. Mesos is only used when we need streaming capabilities in our WW dcs (as our Yarn is centralized in one DC)
• We’re currently on Flink 1.9 (old planner). We have a plan to bump to 1.11 / 1.12 this quarter.
• We typically upgrade once every 6 months to a year (not every release). We’d like to speed up the cadence but we’re not there yet.
• We’d largely be good with keeping Flink on Mesos as-is and functional while missing out on some of the newer features. We understand the pain on the communities side and we can take on the work if we see some fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in the request to port it over.

 

Thanks,

 

-- Piyush

 

 

From: Xintong Song <[hidden email]>
Date: Sunday, October 25, 2020 at 10:57 PM
To: dev <[hidden email]>, user <[hidden email]>
Cc: Lasse Nedergaard <[hidden email]>, <[hidden email]>
Subject: Re: [SURVEY] Remove Mesos support

 

Thanks for sharing the information with us, Piyush an Lasse.

 

@Piyush

 

Thanks for offering the help. IMO, there are currently several problems that make supporting Flink on Mesos challenging for us.

1. Lack of Mesos experts. AFAIK, there are very few people (if not none) among the active contributors in this community that are familiar with Mesos and can help with development on this component.
2. Absence of tests. Mesos does not provide a testing cluster, like `MiniYARNCluster`, making it hard to test interactions between Flink and Mesos. We have only a few very simple e2e tests running on Mesos deployed in a docker, covering the most fundamental workflows. We are not sure how well those tests work, especially against some potential corner cases.
3. Divergence from other deployment. Because of 1 and 2, the new efforts (features, maintenance, refactors) tend to exclude Mesos if possible. When the new efforts have to touch the Mesos related components (e.g., changes to the common resource manager interfaces), we have to be very careful and make as few changes as possible, to avoid accidentally breaking anything that we are not familiar with. As a result, the component diverges a lot from other deployment components (K8s/Yarn), which makes it harder to maintain.

It would be greatly appreciated if you can help with either of the above issues.

 

Additionally, I have a few questions concerning your use cases at Criteo. IIUC, you are going to stay on Mesos in the foreseeable future, while keeping the Flink version up-to-date? What Flink version are you currently using? How often do you upgrade (e.g., every release)? Would you be good with keeping the Flink on Mesos component as it is (means that deployment and resource management improvements may not be ported to Mesos), while keeping other components up-to-date (e.g., improvements from programming APIs, operators, state backens, etc.)? 

 

Thank you~

Xintong Song

 

 

On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <[hidden email]> wrote:

Hi

 

At Trackunit We have been using Mesos for long time but have now moved to k8s. 

Med venlig hilsen / Best regards

Lasse Nedergaard

 

Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.

 

On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Oleksandr Nitavskyi]

Hello Xintong,

Thanks for the insights and support.

Browsing the Mesos backlog and didn't identify anything critical, which is left there.

I see that there are were quite a lot of contributions to the Flink Mesos in the recent version: https://github.com/apache/flink/commits/master/flink-mesos. 

We plan to validate the current Flink master (or release 1.12 branch) our Mesos setup. In case of any issues, we will try to propose changes.

My feeling is that our test results shouldn't affect the Flink 1.12 release cycle. And if any potential commits will land into the 1.12.1 it should be totally fine.

In the future, we would be glad to help you guys with any maintenance-related questions. One of the highest priorities around this component seems to be the development of the full e2e test.

Kind Regards

Oleksandr Nitavskyi

---

From: Xintong Song <[hidden email]>
Sent: Tuesday, October 27, 2020 7:14 AM
To: dev <[hidden email]>; user <[hidden email]>
Cc: Piyush Narang <[hidden email]>
Subject: [BULK]Re: [SURVEY] Remove Mesos support

 

Hi Piyush,

Thanks a lot for sharing the information. It would be a great relief that you are good with Flink on Mesos as is.

As for the jira issues, I believe the most essential ones should have already been resolved. You may find some remaining open issues here [1], but not all of them are necessary if we decide to keep Flink on Mesos as is.

At the moment and in the short future, I think helps are mostly needed on testing the upcoming release 1.12 with Mesos use cases. The community is currently actively preparing the new release, and hopefully we could come up with a release candidate early next month. It would be greatly appreciated if you fork as experienced Flink on Mesos users can help with verifying the release candidates. 

[1] https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open

On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]> wrote:

Hi Xintong,

 

Do you have any jiras that cover any of the items on 1 or 2? I can reach out to folks internally and see if I can get some folks to commit to helping out.

 

To cover the other qs:

• Yes, we’ve not got a plan at the moment to get off Mesos. We use Yarn for some our Flink workloads when we can. Mesos is only used when we need streaming capabilities in our WW dcs (as our Yarn is centralized in one DC)
• We’re currently on Flink 1.9 (old planner). We have a plan to bump to 1.11 / 1.12 this quarter.
• We typically upgrade once every 6 months to a year (not every release). We’d like to speed up the cadence but we’re not there yet.
• We’d largely be good with keeping Flink on Mesos as-is and functional while missing out on some of the newer features. We understand the pain on the communities side and we can take on the work if we see some fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in the request to port it over.

 

Thanks,

 

-- Piyush

 

 

From: Xintong Song <[hidden email]>
Date: Sunday, October 25, 2020 at 10:57 PM
To: dev <[hidden email]>, user <[hidden email]>
Cc: Lasse Nedergaard <[hidden email]>, <[hidden email]>
Subject: Re: [SURVEY] Remove Mesos support

 

Thanks for sharing the information with us, Piyush an Lasse.

 

@Piyush

 

Thanks for offering the help. IMO, there are currently several problems that make supporting Flink on Mesos challenging for us.

1. Lack of Mesos experts. AFAIK, there are very few people (if not none) among the active contributors in this community that are familiar with Mesos and can help with development on this component.
2. Absence of tests. Mesos does not provide a testing cluster, like `MiniYARNCluster`, making it hard to test interactions between Flink and Mesos. We have only a few very simple e2e tests running on Mesos deployed in a docker, covering the most fundamental workflows. We are not sure how well those tests work, especially against some potential corner cases.
3. Divergence from other deployment. Because of 1 and 2, the new efforts (features, maintenance, refactors) tend to exclude Mesos if possible. When the new efforts have to touch the Mesos related components (e.g., changes to the common resource manager interfaces), we have to be very careful and make as few changes as possible, to avoid accidentally breaking anything that we are not familiar with. As a result, the component diverges a lot from other deployment components (K8s/Yarn), which makes it harder to maintain.

It would be greatly appreciated if you can help with either of the above issues.

 

Additionally, I have a few questions concerning your use cases at Criteo. IIUC, you are going to stay on Mesos in the foreseeable future, while keeping the Flink version up-to-date? What Flink version are you currently using? How often do you upgrade (e.g., every release)? Would you be good with keeping the Flink on Mesos component as it is (means that deployment and resource management improvements may not be ported to Mesos), while keeping other components up-to-date (e.g., improvements from programming APIs, operators, state backens, etc.)? 

 

Thank you~

Xintong Song

 

 

On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <[hidden email]> wrote:

Hi

 

At Trackunit We have been using Mesos for long time but have now moved to k8s. 

Med venlig hilsen / Best regards

Lasse Nedergaard

 

Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.

 

On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]>
    > Cc: dev <[hidden email]>, user <[hidden email]>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Till Rohrmann]

Hi Oleksandr,

yes you are right. The biggest problem is at the moment the lack of test coverage and thereby confidence to make changes. We have some e2e tests which you can find here [1]. These tests are, however, quite coarse grained and are missing a lot of cases. One idea would be to add a Mesos e2e test based on Flink's end-to-end test framework [2]. I think what needs to be done there is to add a Mesos resource and a way to submit jobs to a Mesos cluster to write e2e tests.

[1] https://github.com/apache/flink/tree/master/flink-jepsen

[2] https://github.com/apache/flink/tree/master/flink-end-to-end-tests/flink-end-to-end-tests-common

Cheers,

Till

On Tue, Oct 27, 2020 at 12:29 PM Oleksandr Nitavskyi <[hidden email]> wrote:

Hello Xintong,

Thanks for the insights and support.

Browsing the Mesos backlog and didn't identify anything critical, which is left there.

I see that there are were quite a lot of contributions to the Flink Mesos in the recent version: https://github.com/apache/flink/commits/master/flink-mesos.
We plan to validate the current Flink master (or release 1.12 branch) our Mesos setup. In case of any issues, we will try to propose changes.
My feeling is that our test results shouldn't affect the Flink 1.12 release cycle. And if any potential commits will land into the 1.12.1 it should be totally fine.

In the future, we would be glad to help you guys with any maintenance-related questions. One of the highest priorities around this component seems to be the development of the full e2e test.

Kind Regards
Oleksandr Nitavskyi
________________________________
From: Xintong Song <[hidden email]>
Sent: Tuesday, October 27, 2020 7:14 AM
To: dev <[hidden email]>; user <[hidden email]>
Cc: Piyush Narang <[hidden email]>
Subject: [BULK]Re: [SURVEY] Remove Mesos support

Hi Piyush,

Thanks a lot for sharing the information. It would be a great relief that you are good with Flink on Mesos as is.

As for the jira issues, I believe the most essential ones should have already been resolved. You may find some remaining open issues here [1], but not all of them are necessary if we decide to keep Flink on Mesos as is.

At the moment and in the short future, I think helps are mostly needed on testing the upcoming release 1.12 with Mesos use cases. The community is currently actively preparing the new release, and hopefully we could come up with a release candidate early next month. It would be greatly appreciated if you fork as experienced Flink on Mesos users can help with verifying the release candidates.


Thank you~

Xintong Song

[1] https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues.apache.org%2Fjira%2Fbrowse%2FFLINK-17402%3Fjql%3Dproject%2520%253D%2520FLINK%2520AND%2520component%2520%253D%2520%2522Deployment%2520%252F%2520Mesos%2522%2520AND%2520status%2520%253D%2520Open&data=04%7C01%7Co.nitavskyi%40criteo.com%7C3585e1f25bdf4e091af808d87a3f92db%7C2a35d8fd574d48e3927c8c398e225a01%7C1%7C0%7C637393760750820881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hytJFQE0MCPzMLiQTQTdbg3GVckX5M3r1NPRGrRV8j4%3D&reserved=0>

On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]<mailto:[hidden email]>> wrote:

Hi Xintong,



Do you have any jiras that cover any of the items on 1 or 2? I can reach out to folks internally and see if I can get some folks to commit to helping out.



To cover the other qs:

  *   Yes, we’ve not got a plan at the moment to get off Mesos. We use Yarn for some our Flink workloads when we can. Mesos is only used when we need streaming capabilities in our WW dcs (as our Yarn is centralized in one DC)
  *   We’re currently on Flink 1.9 (old planner). We have a plan to bump to 1.11 / 1.12 this quarter.
  *   We typically upgrade once every 6 months to a year (not every release). We’d like to speed up the cadence but we’re not there yet.
  *   We’d largely be good with keeping Flink on Mesos as-is and functional while missing out on some of the newer features. We understand the pain on the communities side and we can take on the work if we see some fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in the request to port it over.



Thanks,



-- Piyush





From: Xintong Song <[hidden email]<mailto:[hidden email]>>
Date: Sunday, October 25, 2020 at 10:57 PM
To: dev <[hidden email]<mailto:[hidden email]>>, user <[hidden email]<mailto:[hidden email]>>
Cc: Lasse Nedergaard <[hidden email]<mailto:[hidden email]>>, <[hidden email]<mailto:[hidden email]>>
Subject: Re: [SURVEY] Remove Mesos support



Thanks for sharing the information with us, Piyush an Lasse.



@Piyush



Thanks for offering the help. IMO, there are currently several problems that make supporting Flink on Mesos challenging for us.

  1.  Lack of Mesos experts. AFAIK, there are very few people (if not none) among the active contributors in this community that are familiar with Mesos and can help with development on this component.
  2.  Absence of tests. Mesos does not provide a testing cluster, like `MiniYARNCluster`, making it hard to test interactions between Flink and Mesos. We have only a few very simple e2e tests running on Mesos deployed in a docker, covering the most fundamental workflows. We are not sure how well those tests work, especially against some potential corner cases.
  3.  Divergence from other deployment. Because of 1 and 2, the new efforts (features, maintenance, refactors) tend to exclude Mesos if possible. When the new efforts have to touch the Mesos related components (e.g., changes to the common resource manager interfaces), we have to be very careful and make as few changes as possible, to avoid accidentally breaking anything that we are not familiar with. As a result, the component diverges a lot from other deployment components (K8s/Yarn), which makes it harder to maintain.

It would be greatly appreciated if you can help with either of the above issues.



Additionally, I have a few questions concerning your use cases at Criteo. IIUC, you are going to stay on Mesos in the foreseeable future, while keeping the Flink version up-to-date? What Flink version are you currently using? How often do you upgrade (e.g., every release)? Would you be good with keeping the Flink on Mesos component as it is (means that deployment and resource management improvements may not be ported to Mesos), while keeping other components up-to-date (e.g., improvements from programming APIs, operators, state backens, etc.)?



Thank you~

Xintong Song





On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <[hidden email]<mailto:[hidden email]>> wrote:

Hi



At Trackunit We have been using Mesos for long time but have now moved to k8s.

Med venlig hilsen / Best regards

Lasse Nedergaard





Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]<mailto:[hidden email]>>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.



On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]<mailto:[hidden email]>> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]<mailto:[hidden email]>> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]<mailto:[hidden email]>> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]<mailto:[hidden email]>>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]<mailto:[hidden email]>>
    > Cc: dev <[hidden email]<mailto:[hidden email]>>, user <[hidden email]<mailto:[hidden email]>>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]<mailto:[hidden email]>> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]<mailto:[hidden email]>> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Matthias]

Hi everyone,

considering the upcoming release of Flink 1.13, I wanted to revive the discussion about the Mesos support ones more. Mesos is also already listed as deprecated in Flink's overall roadmap [1]. Maybe, it's time to align the documentation accordingly to make it more explicit?

What do you think?

Best,
Matthias

[1] https://flink.apache.org/roadmap.html#feature-radar

On Wed, Oct 28, 2020 at 9:40 AM Till Rohrmann <[hidden email]> wrote:

Hi Oleksandr,

yes you are right. The biggest problem is at the moment the lack of test coverage and thereby confidence to make changes. We have some e2e tests which you can find here [1]. These tests are, however, quite coarse grained and are missing a lot of cases. One idea would be to add a Mesos e2e test based on Flink's end-to-end test framework [2]. I think what needs to be done there is to add a Mesos resource and a way to submit jobs to a Mesos cluster to write e2e tests.

[1] https://github.com/apache/flink/tree/master/flink-jepsen

[2] https://github.com/apache/flink/tree/master/flink-end-to-end-tests/flink-end-to-end-tests-common

Cheers,

Till

On Tue, Oct 27, 2020 at 12:29 PM Oleksandr Nitavskyi <[hidden email]> wrote:

Hello Xintong,

Thanks for the insights and support.

Browsing the Mesos backlog and didn't identify anything critical, which is left there.

I see that there are were quite a lot of contributions to the Flink Mesos in the recent version: https://github.com/apache/flink/commits/master/flink-mesos.
We plan to validate the current Flink master (or release 1.12 branch) our Mesos setup. In case of any issues, we will try to propose changes.
My feeling is that our test results shouldn't affect the Flink 1.12 release cycle. And if any potential commits will land into the 1.12.1 it should be totally fine.

In the future, we would be glad to help you guys with any maintenance-related questions. One of the highest priorities around this component seems to be the development of the full e2e test.

Kind Regards
Oleksandr Nitavskyi
________________________________
From: Xintong Song <[hidden email]>
Sent: Tuesday, October 27, 2020 7:14 AM
To: dev <[hidden email]>; user <[hidden email]>
Cc: Piyush Narang <[hidden email]>
Subject: [BULK]Re: [SURVEY] Remove Mesos support

Hi Piyush,

Thanks a lot for sharing the information. It would be a great relief that you are good with Flink on Mesos as is.

As for the jira issues, I believe the most essential ones should have already been resolved. You may find some remaining open issues here [1], but not all of them are necessary if we decide to keep Flink on Mesos as is.

At the moment and in the short future, I think helps are mostly needed on testing the upcoming release 1.12 with Mesos use cases. The community is currently actively preparing the new release, and hopefully we could come up with a release candidate early next month. It would be greatly appreciated if you fork as experienced Flink on Mesos users can help with verifying the release candidates.


Thank you~

Xintong Song

[1] https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues.apache.org%2Fjira%2Fbrowse%2FFLINK-17402%3Fjql%3Dproject%2520%253D%2520FLINK%2520AND%2520component%2520%253D%2520%2522Deployment%2520%252F%2520Mesos%2522%2520AND%2520status%2520%253D%2520Open&data=04%7C01%7Co.nitavskyi%40criteo.com%7C3585e1f25bdf4e091af808d87a3f92db%7C2a35d8fd574d48e3927c8c398e225a01%7C1%7C0%7C637393760750820881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hytJFQE0MCPzMLiQTQTdbg3GVckX5M3r1NPRGrRV8j4%3D&reserved=0>

On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]<mailto:[hidden email]>> wrote:

Hi Xintong,



Do you have any jiras that cover any of the items on 1 or 2? I can reach out to folks internally and see if I can get some folks to commit to helping out.



To cover the other qs:

  *   Yes, we’ve not got a plan at the moment to get off Mesos. We use Yarn for some our Flink workloads when we can. Mesos is only used when we need streaming capabilities in our WW dcs (as our Yarn is centralized in one DC)
  *   We’re currently on Flink 1.9 (old planner). We have a plan to bump to 1.11 / 1.12 this quarter.
  *   We typically upgrade once every 6 months to a year (not every release). We’d like to speed up the cadence but we’re not there yet.
  *   We’d largely be good with keeping Flink on Mesos as-is and functional while missing out on some of the newer features. We understand the pain on the communities side and we can take on the work if we see some fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in the request to port it over.



Thanks,



-- Piyush





From: Xintong Song <[hidden email]<mailto:[hidden email]>>
Date: Sunday, October 25, 2020 at 10:57 PM
To: dev <[hidden email]<mailto:[hidden email]>>, user <[hidden email]<mailto:[hidden email]>>
Cc: Lasse Nedergaard <[hidden email]<mailto:[hidden email]>>, <[hidden email]<mailto:[hidden email]>>
Subject: Re: [SURVEY] Remove Mesos support



Thanks for sharing the information with us, Piyush an Lasse.



@Piyush



Thanks for offering the help. IMO, there are currently several problems that make supporting Flink on Mesos challenging for us.

  1.  Lack of Mesos experts. AFAIK, there are very few people (if not none) among the active contributors in this community that are familiar with Mesos and can help with development on this component.
  2.  Absence of tests. Mesos does not provide a testing cluster, like `MiniYARNCluster`, making it hard to test interactions between Flink and Mesos. We have only a few very simple e2e tests running on Mesos deployed in a docker, covering the most fundamental workflows. We are not sure how well those tests work, especially against some potential corner cases.
  3.  Divergence from other deployment. Because of 1 and 2, the new efforts (features, maintenance, refactors) tend to exclude Mesos if possible. When the new efforts have to touch the Mesos related components (e.g., changes to the common resource manager interfaces), we have to be very careful and make as few changes as possible, to avoid accidentally breaking anything that we are not familiar with. As a result, the component diverges a lot from other deployment components (K8s/Yarn), which makes it harder to maintain.

It would be greatly appreciated if you can help with either of the above issues.



Additionally, I have a few questions concerning your use cases at Criteo. IIUC, you are going to stay on Mesos in the foreseeable future, while keeping the Flink version up-to-date? What Flink version are you currently using? How often do you upgrade (e.g., every release)? Would you be good with keeping the Flink on Mesos component as it is (means that deployment and resource management improvements may not be ported to Mesos), while keeping other components up-to-date (e.g., improvements from programming APIs, operators, state backens, etc.)?



Thank you~

Xintong Song





On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <[hidden email]<mailto:[hidden email]>> wrote:

Hi



At Trackunit We have been using Mesos for long time but have now moved to k8s.

Med venlig hilsen / Best regards

Lasse Nedergaard





Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]<mailto:[hidden email]>>:



Hey Piyush,

thanks a lot for raising this concern. I believe we should keep Mesos in Flink then in the foreseeable future.

Your offer to help is much appreciated. We'll let you know once there is something.



On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]<mailto:[hidden email]>> wrote:

Thanks Kostas. If there's items we can help with, I'm sure we'd be able to find folks who would be excited to contribute / help in any way.

-- Piyush


On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]<mailto:[hidden email]>> wrote:

    Thanks Piyush for the message.
    After this, I revoke my +1. I agree with the previous opinions that we
    cannot drop code that is actively used by users, especially if it
    something that deep in the stack as support for cluster management
    framework.

    Cheers,
    Kostas

    On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]<mailto:[hidden email]>> wrote:
    >
    > Hi folks,
    >
    >
    >
    > We at Criteo are active users of the Flink on Mesos resource management component. We are pretty heavy users of Mesos for scheduling workloads on our edge datacenters and we do want to continue to be able to run some of our Flink topologies (to compute machine learning short term features) on those DCs. If possible our vote would be not to drop Mesos support as that will tie us to an old release / have to maintain a fork as we’re not planning to migrate off Mesos anytime soon. Is the burden something that can be helped with by the community? (Or are you referring to having to ensure PRs handle the Mesos piece as well when they touch the resource managers?)
    >
    >
    >
    > Thanks,
    >
    >
    >
    > -- Piyush
    >
    >
    >
    >
    >
    > From: Till Rohrmann <[hidden email]<mailto:[hidden email]>>
    > Date: Friday, October 23, 2020 at 8:19 AM
    > To: Xintong Song <[hidden email]<mailto:[hidden email]>>
    > Cc: dev <[hidden email]<mailto:[hidden email]>>, user <[hidden email]<mailto:[hidden email]>>
    > Subject: Re: [SURVEY] Remove Mesos support
    >
    >
    >
    > Thanks for starting this survey Robert! I second Konstantin and Xintong in the sense that our Mesos user's opinions should matter most here. If our community is no longer using the Mesos integration, then I would be +1 for removing it in order to decrease the maintenance burden.
    >
    >
    >
    > Cheers,
    >
    > Till
    >
    >
    >
    > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]<mailto:[hidden email]>> wrote:
    >
    > +1 for adding a warning in 1.12 about planning to remove Mesos support.
    >
    >
    >
    > With my developer hat on, removing the Mesos support would definitely reduce the maintaining overhead for the deployment and resource management related components. On the other hand, the Flink on Mesos users' voices definitely matter a lot for this community. Either way, it would be good to draw users attention to this discussion early.
    >
    >
    >
    > Thank you~
    >
    > Xintong Song
    >
    >
    >
    >
    >
    > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]<mailto:[hidden email]>> wrote:
    >
    > Hi Robert,
    >
    > +1 to the plan you outlined. If we were to drop support in Flink 1.13+, we
    > would still support it in Flink 1.12- with bug fixes for some time so that
    > users have time to move on.
    >
    > It would certainly be very interesting to hear from current Flink on Mesos
    > users, on how they see the evolution of this part of the ecosystem.
    >
    > Best,
    >
    > Konstantin

[Konstantin Knauf-3]

Hi Matthias, 

Thank you for following up on this. +1 to officially deprecate Mesos in the code and documentation, too. It will be confusing for users if this diverges from the roadmap. 

Cheers, 

Konstantin

On Thu, Mar 25, 2021 at 12:23 PM Matthias Pohl <[hidden email]> wrote:

Hi everyone,
considering the upcoming release of Flink 1.13, I wanted to revive the
discussion about the Mesos support ones more. Mesos is also already listed
as deprecated in Flink's overall roadmap [1]. Maybe, it's time to align the
documentation accordingly to make it more explicit?

What do you think?

Best,
Matthias

[1] https://flink.apache.org/roadmap.html#feature-radar

On Wed, Oct 28, 2020 at 9:40 AM Till Rohrmann <[hidden email]> wrote:

> Hi Oleksandr,
>
> yes you are right. The biggest problem is at the moment the lack of test
> coverage and thereby confidence to make changes. We have some e2e tests
> which you can find here [1]. These tests are, however, quite coarse grained
> and are missing a lot of cases. One idea would be to add a Mesos e2e test
> based on Flink's end-to-end test framework [2]. I think what needs to be
> done there is to add a Mesos resource and a way to submit jobs to a Mesos
> cluster to write e2e tests.
>
> [1] https://github.com/apache/flink/tree/master/flink-jepsen
> [2]
> https://github.com/apache/flink/tree/master/flink-end-to-end-tests/flink-end-to-end-tests-common
>
> Cheers,
> Till
>
> On Tue, Oct 27, 2020 at 12:29 PM Oleksandr Nitavskyi <
> [hidden email]> wrote:
>
>> Hello Xintong,
>>
>> Thanks for the insights and support.
>>
>> Browsing the Mesos backlog and didn't identify anything critical, which
>> is left there.
>>
>> I see that there are were quite a lot of contributions to the Flink Mesos
>> in the recent version:
>> https://github.com/apache/flink/commits/master/flink-mesos.
>> We plan to validate the current Flink master (or release 1.12 branch) our
>> Mesos setup. In case of any issues, we will try to propose changes.
>> My feeling is that our test results shouldn't affect the Flink 1.12
>> release cycle. And if any potential commits will land into the 1.12.1 it
>> should be totally fine.
>>
>> In the future, we would be glad to help you guys with any
>> maintenance-related questions. One of the highest priorities around this
>> component seems to be the development of the full e2e test.
>>
>> Kind Regards
>> Oleksandr Nitavskyi
>> ________________________________
>> From: Xintong Song <[hidden email]>
>> Sent: Tuesday, October 27, 2020 7:14 AM
>> To: dev <[hidden email]>; user <[hidden email]>
>> Cc: Piyush Narang <[hidden email]>
>> Subject: [BULK]Re: [SURVEY] Remove Mesos support
>>
>> Hi Piyush,
>>
>> Thanks a lot for sharing the information. It would be a great relief that
>> you are good with Flink on Mesos as is.
>>
>> As for the jira issues, I believe the most essential ones should have
>> already been resolved. You may find some remaining open issues here [1],
>> but not all of them are necessary if we decide to keep Flink on Mesos as is.
>>
>> At the moment and in the short future, I think helps are mostly needed on
>> testing the upcoming release 1.12 with Mesos use cases. The community is
>> currently actively preparing the new release, and hopefully we could come
>> up with a release candidate early next month. It would be greatly
>> appreciated if you fork as experienced Flink on Mesos users can help with
>> verifying the release candidates.
>>
>>
>> Thank you~
>>
>> Xintong Song
>>
>> [1]
>> https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open
>> <
>> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues.apache.org%2Fjira%2Fbrowse%2FFLINK-17402%3Fjql%3Dproject%2520%253D%2520FLINK%2520AND%2520component%2520%253D%2520%2522Deployment%2520%252F%2520Mesos%2522%2520AND%2520status%2520%253D%2520Open&data=04%7C01%7Co.nitavskyi%40criteo.com%7C3585e1f25bdf4e091af808d87a3f92db%7C2a35d8fd574d48e3927c8c398e225a01%7C1%7C0%7C637393760750820881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hytJFQE0MCPzMLiQTQTdbg3GVckX5M3r1NPRGrRV8j4%3D&reserved=0
>> >
>>
>> On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>> Hi Xintong,
>>
>>
>>
>> Do you have any jiras that cover any of the items on 1 or 2? I can reach
>> out to folks internally and see if I can get some folks to commit to
>> helping out.
>>
>>
>>
>> To cover the other qs:
>>
>>   *   Yes, we’ve not got a plan at the moment to get off Mesos. We use
>> Yarn for some our Flink workloads when we can. Mesos is only used when we
>> need streaming capabilities in our WW dcs (as our Yarn is centralized in
>> one DC)
>>   *   We’re currently on Flink 1.9 (old planner). We have a plan to bump
>> to 1.11 / 1.12 this quarter.
>>   *   We typically upgrade once every 6 months to a year (not every
>> release). We’d like to speed up the cadence but we’re not there yet.
>>   *   We’d largely be good with keeping Flink on Mesos as-is and
>> functional while missing out on some of the newer features. We understand
>> the pain on the communities side and we can take on the work if we see some
>> fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in
>> the request to port it over.
>>
>>
>>
>> Thanks,
>>
>>
>>
>> -- Piyush
>>
>>
>>
>>
>>
>> From: Xintong Song <[hidden email]<mailto:[hidden email]>>
>> Date: Sunday, October 25, 2020 at 10:57 PM
>> To: dev <[hidden email]<mailto:[hidden email]>>, user <
>> [hidden email]<mailto:[hidden email]>>
>> Cc: Lasse Nedergaard <[hidden email]<mailto:
>> [hidden email]>>, <[hidden email]<mailto:
>> [hidden email]>>
>> Subject: Re: [SURVEY] Remove Mesos support
>>
>>
>>
>> Thanks for sharing the information with us, Piyush an Lasse.
>>
>>
>>
>> @Piyush
>>
>>
>>
>> Thanks for offering the help. IMO, there are currently several problems
>> that make supporting Flink on Mesos challenging for us.
>>
>>   1.  Lack of Mesos experts. AFAIK, there are very few people (if not
>> none) among the active contributors in this community that are familiar
>> with Mesos and can help with development on this component.
>>   2.  Absence of tests. Mesos does not provide a testing cluster, like
>> `MiniYARNCluster`, making it hard to test interactions between Flink and
>> Mesos. We have only a few very simple e2e tests running on Mesos deployed
>> in a docker, covering the most fundamental workflows. We are not sure how
>> well those tests work, especially against some potential corner cases.
>>   3.  Divergence from other deployment. Because of 1 and 2, the new
>> efforts (features, maintenance, refactors) tend to exclude Mesos if
>> possible. When the new efforts have to touch the Mesos related components
>> (e.g., changes to the common resource manager interfaces), we have to be
>> very careful and make as few changes as possible, to avoid accidentally
>> breaking anything that we are not familiar with. As a result, the component
>> diverges a lot from other deployment components (K8s/Yarn), which makes it
>> harder to maintain.
>>
>> It would be greatly appreciated if you can help with either of the above
>> issues.
>>
>>
>>
>> Additionally, I have a few questions concerning your use cases at Criteo.
>> IIUC, you are going to stay on Mesos in the foreseeable future, while
>> keeping the Flink version up-to-date? What Flink version are you currently
>> using? How often do you upgrade (e.g., every release)? Would you be good
>> with keeping the Flink on Mesos component as it is (means that deployment
>> and resource management improvements may not be ported to Mesos), while
>> keeping other components up-to-date (e.g., improvements from programming
>> APIs, operators, state backens, etc.)?
>>
>>
>>
>> Thank you~
>>
>> Xintong Song
>>
>>
>>
>>
>>
>> On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <
>> [hidden email]<mailto:[hidden email]>>
>> wrote:
>>
>> Hi
>>
>>
>>
>> At Trackunit We have been using Mesos for long time but have now moved to
>> k8s.
>>
>> Med venlig hilsen / Best regards
>>
>> Lasse Nedergaard
>>
>>
>>
>>
>>
>> Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]
>> <mailto:[hidden email]>>:
>>
>> 
>>
>> Hey Piyush,
>>
>> thanks a lot for raising this concern. I believe we should keep Mesos in
>> Flink then in the foreseeable future.
>>
>> Your offer to help is much appreciated. We'll let you know once there is
>> something.
>>
>>
>>
>> On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>> Thanks Kostas. If there's items we can help with, I'm sure we'd be able
>> to find folks who would be excited to contribute / help in any way.
>>
>> -- Piyush
>>
>>
>> On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]<mailto:
>> [hidden email]>> wrote:
>>
>>     Thanks Piyush for the message.
>>     After this, I revoke my +1. I agree with the previous opinions that we
>>     cannot drop code that is actively used by users, especially if it
>>     something that deep in the stack as support for cluster management
>>     framework.
>>
>>     Cheers,
>>     Kostas
>>
>>     On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > Hi folks,
>>     >
>>     >
>>     >
>>     > We at Criteo are active users of the Flink on Mesos resource
>> management component. We are pretty heavy users of Mesos for scheduling
>> workloads on our edge datacenters and we do want to continue to be able to
>> run some of our Flink topologies (to compute machine learning short term
>> features) on those DCs. If possible our vote would be not to drop Mesos
>> support as that will tie us to an old release / have to maintain a fork as
>> we’re not planning to migrate off Mesos anytime soon. Is the burden
>> something that can be helped with by the community? (Or are you referring
>> to having to ensure PRs handle the Mesos piece as well when they touch the
>> resource managers?)
>>     >
>>     >
>>     >
>>     > Thanks,
>>     >
>>     >
>>     >
>>     > -- Piyush
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > From: Till Rohrmann <[hidden email]<mailto:
>> [hidden email]>>
>>     > Date: Friday, October 23, 2020 at 8:19 AM
>>     > To: Xintong Song <[hidden email]<mailto:
>> [hidden email]>>
>>     > Cc: dev <[hidden email]<mailto:[hidden email]>>, user <
>> [hidden email]<mailto:[hidden email]>>
>>     > Subject: Re: [SURVEY] Remove Mesos support
>>     >
>>     >
>>     >
>>     > Thanks for starting this survey Robert! I second Konstantin and
>> Xintong in the sense that our Mesos user's opinions should matter most
>> here. If our community is no longer using the Mesos integration, then I
>> would be +1 for removing it in order to decrease the maintenance burden.
>>     >
>>     >
>>     >
>>     > Cheers,
>>     >
>>     > Till
>>     >
>>     >
>>     >
>>     > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > +1 for adding a warning in 1.12 about planning to remove Mesos
>> support.
>>     >
>>     >
>>     >
>>     > With my developer hat on, removing the Mesos support would
>> definitely reduce the maintaining overhead for the deployment and resource
>> management related components. On the other hand, the Flink on Mesos users'
>> voices definitely matter a lot for this community. Either way, it would be
>> good to draw users attention to this discussion early.
>>     >
>>     >
>>     >
>>     > Thank you~
>>     >
>>     > Xintong Song
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > Hi Robert,
>>     >
>>     > +1 to the plan you outlined. If we were to drop support in Flink
>> 1.13+, we
>>     > would still support it in Flink 1.12- with bug fixes for some time
>> so that
>>     > users have time to move on.
>>     >
>>     > It would certainly be very interesting to hear from current Flink
>> on Mesos
>>     > users, on how they see the evolution of this part of the ecosystem.
>>     >
>>     > Best,
>>     >
>>     > Konstantin
>>
>

--

[Till Rohrmann]

+1 for officially deprecating this component for the 1.13 release.

Cheers,

Till

On Thu, Mar 25, 2021 at 1:49 PM Konstantin Knauf <[hidden email]> wrote:

Hi Matthias, 

Thank you for following up on this. +1 to officially deprecate Mesos in the code and documentation, too. It will be confusing for users if this diverges from the roadmap. 

Cheers, 

Konstantin

On Thu, Mar 25, 2021 at 12:23 PM Matthias Pohl <[hidden email]> wrote:

Hi everyone,
considering the upcoming release of Flink 1.13, I wanted to revive the
discussion about the Mesos support ones more. Mesos is also already listed
as deprecated in Flink's overall roadmap [1]. Maybe, it's time to align the
documentation accordingly to make it more explicit?

What do you think?

Best,
Matthias

[1] https://flink.apache.org/roadmap.html#feature-radar

On Wed, Oct 28, 2020 at 9:40 AM Till Rohrmann <[hidden email]> wrote:

> Hi Oleksandr,
>
> yes you are right. The biggest problem is at the moment the lack of test
> coverage and thereby confidence to make changes. We have some e2e tests
> which you can find here [1]. These tests are, however, quite coarse grained
> and are missing a lot of cases. One idea would be to add a Mesos e2e test
> based on Flink's end-to-end test framework [2]. I think what needs to be
> done there is to add a Mesos resource and a way to submit jobs to a Mesos
> cluster to write e2e tests.
>
> [1] https://github.com/apache/flink/tree/master/flink-jepsen
> [2]
> https://github.com/apache/flink/tree/master/flink-end-to-end-tests/flink-end-to-end-tests-common
>
> Cheers,
> Till
>
> On Tue, Oct 27, 2020 at 12:29 PM Oleksandr Nitavskyi <
> [hidden email]> wrote:
>
>> Hello Xintong,
>>
>> Thanks for the insights and support.
>>
>> Browsing the Mesos backlog and didn't identify anything critical, which
>> is left there.
>>
>> I see that there are were quite a lot of contributions to the Flink Mesos
>> in the recent version:
>> https://github.com/apache/flink/commits/master/flink-mesos.
>> We plan to validate the current Flink master (or release 1.12 branch) our
>> Mesos setup. In case of any issues, we will try to propose changes.
>> My feeling is that our test results shouldn't affect the Flink 1.12
>> release cycle. And if any potential commits will land into the 1.12.1 it
>> should be totally fine.
>>
>> In the future, we would be glad to help you guys with any
>> maintenance-related questions. One of the highest priorities around this
>> component seems to be the development of the full e2e test.
>>
>> Kind Regards
>> Oleksandr Nitavskyi
>> ________________________________
>> From: Xintong Song <[hidden email]>
>> Sent: Tuesday, October 27, 2020 7:14 AM
>> To: dev <[hidden email]>; user <[hidden email]>
>> Cc: Piyush Narang <[hidden email]>
>> Subject: [BULK]Re: [SURVEY] Remove Mesos support
>>
>> Hi Piyush,
>>
>> Thanks a lot for sharing the information. It would be a great relief that
>> you are good with Flink on Mesos as is.
>>
>> As for the jira issues, I believe the most essential ones should have
>> already been resolved. You may find some remaining open issues here [1],
>> but not all of them are necessary if we decide to keep Flink on Mesos as is.
>>
>> At the moment and in the short future, I think helps are mostly needed on
>> testing the upcoming release 1.12 with Mesos use cases. The community is
>> currently actively preparing the new release, and hopefully we could come
>> up with a release candidate early next month. It would be greatly
>> appreciated if you fork as experienced Flink on Mesos users can help with
>> verifying the release candidates.
>>
>>
>> Thank you~
>>
>> Xintong Song
>>
>> [1]
>> https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open
>> <
>> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues.apache.org%2Fjira%2Fbrowse%2FFLINK-17402%3Fjql%3Dproject%2520%253D%2520FLINK%2520AND%2520component%2520%253D%2520%2522Deployment%2520%252F%2520Mesos%2522%2520AND%2520status%2520%253D%2520Open&data=04%7C01%7Co.nitavskyi%40criteo.com%7C3585e1f25bdf4e091af808d87a3f92db%7C2a35d8fd574d48e3927c8c398e225a01%7C1%7C0%7C637393760750820881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hytJFQE0MCPzMLiQTQTdbg3GVckX5M3r1NPRGrRV8j4%3D&reserved=0
>> >
>>
>> On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>> Hi Xintong,
>>
>>
>>
>> Do you have any jiras that cover any of the items on 1 or 2? I can reach
>> out to folks internally and see if I can get some folks to commit to
>> helping out.
>>
>>
>>
>> To cover the other qs:
>>
>>   *   Yes, we’ve not got a plan at the moment to get off Mesos. We use
>> Yarn for some our Flink workloads when we can. Mesos is only used when we
>> need streaming capabilities in our WW dcs (as our Yarn is centralized in
>> one DC)
>>   *   We’re currently on Flink 1.9 (old planner). We have a plan to bump
>> to 1.11 / 1.12 this quarter.
>>   *   We typically upgrade once every 6 months to a year (not every
>> release). We’d like to speed up the cadence but we’re not there yet.
>>   *   We’d largely be good with keeping Flink on Mesos as-is and
>> functional while missing out on some of the newer features. We understand
>> the pain on the communities side and we can take on the work if we see some
>> fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in
>> the request to port it over.
>>
>>
>>
>> Thanks,
>>
>>
>>
>> -- Piyush
>>
>>
>>
>>
>>
>> From: Xintong Song <[hidden email]<mailto:[hidden email]>>
>> Date: Sunday, October 25, 2020 at 10:57 PM
>> To: dev <[hidden email]<mailto:[hidden email]>>, user <
>> [hidden email]<mailto:[hidden email]>>
>> Cc: Lasse Nedergaard <[hidden email]<mailto:
>> [hidden email]>>, <[hidden email]<mailto:
>> [hidden email]>>
>> Subject: Re: [SURVEY] Remove Mesos support
>>
>>
>>
>> Thanks for sharing the information with us, Piyush an Lasse.
>>
>>
>>
>> @Piyush
>>
>>
>>
>> Thanks for offering the help. IMO, there are currently several problems
>> that make supporting Flink on Mesos challenging for us.
>>
>>   1.  Lack of Mesos experts. AFAIK, there are very few people (if not
>> none) among the active contributors in this community that are familiar
>> with Mesos and can help with development on this component.
>>   2.  Absence of tests. Mesos does not provide a testing cluster, like
>> `MiniYARNCluster`, making it hard to test interactions between Flink and
>> Mesos. We have only a few very simple e2e tests running on Mesos deployed
>> in a docker, covering the most fundamental workflows. We are not sure how
>> well those tests work, especially against some potential corner cases.
>>   3.  Divergence from other deployment. Because of 1 and 2, the new
>> efforts (features, maintenance, refactors) tend to exclude Mesos if
>> possible. When the new efforts have to touch the Mesos related components
>> (e.g., changes to the common resource manager interfaces), we have to be
>> very careful and make as few changes as possible, to avoid accidentally
>> breaking anything that we are not familiar with. As a result, the component
>> diverges a lot from other deployment components (K8s/Yarn), which makes it
>> harder to maintain.
>>
>> It would be greatly appreciated if you can help with either of the above
>> issues.
>>
>>
>>
>> Additionally, I have a few questions concerning your use cases at Criteo.
>> IIUC, you are going to stay on Mesos in the foreseeable future, while
>> keeping the Flink version up-to-date? What Flink version are you currently
>> using? How often do you upgrade (e.g., every release)? Would you be good
>> with keeping the Flink on Mesos component as it is (means that deployment
>> and resource management improvements may not be ported to Mesos), while
>> keeping other components up-to-date (e.g., improvements from programming
>> APIs, operators, state backens, etc.)?
>>
>>
>>
>> Thank you~
>>
>> Xintong Song
>>
>>
>>
>>
>>
>> On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <
>> [hidden email]<mailto:[hidden email]>>
>> wrote:
>>
>> Hi
>>
>>
>>
>> At Trackunit We have been using Mesos for long time but have now moved to
>> k8s.
>>
>> Med venlig hilsen / Best regards
>>
>> Lasse Nedergaard
>>
>>
>>
>>
>>
>> Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]
>> <mailto:[hidden email]>>:
>>
>> 
>>
>> Hey Piyush,
>>
>> thanks a lot for raising this concern. I believe we should keep Mesos in
>> Flink then in the foreseeable future.
>>
>> Your offer to help is much appreciated. We'll let you know once there is
>> something.
>>
>>
>>
>> On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>> Thanks Kostas. If there's items we can help with, I'm sure we'd be able
>> to find folks who would be excited to contribute / help in any way.
>>
>> -- Piyush
>>
>>
>> On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]<mailto:
>> [hidden email]>> wrote:
>>
>>     Thanks Piyush for the message.
>>     After this, I revoke my +1. I agree with the previous opinions that we
>>     cannot drop code that is actively used by users, especially if it
>>     something that deep in the stack as support for cluster management
>>     framework.
>>
>>     Cheers,
>>     Kostas
>>
>>     On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > Hi folks,
>>     >
>>     >
>>     >
>>     > We at Criteo are active users of the Flink on Mesos resource
>> management component. We are pretty heavy users of Mesos for scheduling
>> workloads on our edge datacenters and we do want to continue to be able to
>> run some of our Flink topologies (to compute machine learning short term
>> features) on those DCs. If possible our vote would be not to drop Mesos
>> support as that will tie us to an old release / have to maintain a fork as
>> we’re not planning to migrate off Mesos anytime soon. Is the burden
>> something that can be helped with by the community? (Or are you referring
>> to having to ensure PRs handle the Mesos piece as well when they touch the
>> resource managers?)
>>     >
>>     >
>>     >
>>     > Thanks,
>>     >
>>     >
>>     >
>>     > -- Piyush
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > From: Till Rohrmann <[hidden email]<mailto:
>> [hidden email]>>
>>     > Date: Friday, October 23, 2020 at 8:19 AM
>>     > To: Xintong Song <[hidden email]<mailto:
>> [hidden email]>>
>>     > Cc: dev <[hidden email]<mailto:[hidden email]>>, user <
>> [hidden email]<mailto:[hidden email]>>
>>     > Subject: Re: [SURVEY] Remove Mesos support
>>     >
>>     >
>>     >
>>     > Thanks for starting this survey Robert! I second Konstantin and
>> Xintong in the sense that our Mesos user's opinions should matter most
>> here. If our community is no longer using the Mesos integration, then I
>> would be +1 for removing it in order to decrease the maintenance burden.
>>     >
>>     >
>>     >
>>     > Cheers,
>>     >
>>     > Till
>>     >
>>     >
>>     >
>>     > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > +1 for adding a warning in 1.12 about planning to remove Mesos
>> support.
>>     >
>>     >
>>     >
>>     > With my developer hat on, removing the Mesos support would
>> definitely reduce the maintaining overhead for the deployment and resource
>> management related components. On the other hand, the Flink on Mesos users'
>> voices definitely matter a lot for this community. Either way, it would be
>> good to draw users attention to this discussion early.
>>     >
>>     >
>>     >
>>     > Thank you~
>>     >
>>     > Xintong Song
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > Hi Robert,
>>     >
>>     > +1 to the plan you outlined. If we were to drop support in Flink
>> 1.13+, we
>>     > would still support it in Flink 1.12- with bug fixes for some time
>> so that
>>     > users have time to move on.
>>     >
>>     > It would certainly be very interesting to hear from current Flink
>> on Mesos
>>     > users, on how they see the evolution of this part of the ecosystem.
>>     >
>>     > Best,
>>     >
>>     > Konstantin
>>
>

--

Konstantin Knauf

https://twitter.com/snntrable

https://github.com/knaufk

[Xintong Song]

+1

It's already a matter of fact for a while that we no longer port new features to the Mesos deployment.

On Fri, Mar 26, 2021 at 10:37 PM Till Rohrmann <[hidden email]> wrote:

+1 for officially deprecating this component for the 1.13 release.

Cheers,

Till

On Thu, Mar 25, 2021 at 1:49 PM Konstantin Knauf <[hidden email]> wrote:

Hi Matthias, 

Thank you for following up on this. +1 to officially deprecate Mesos in the code and documentation, too. It will be confusing for users if this diverges from the roadmap. 

Cheers, 

Konstantin

On Thu, Mar 25, 2021 at 12:23 PM Matthias Pohl <[hidden email]> wrote:

Hi everyone,
considering the upcoming release of Flink 1.13, I wanted to revive the
discussion about the Mesos support ones more. Mesos is also already listed
as deprecated in Flink's overall roadmap [1]. Maybe, it's time to align the
documentation accordingly to make it more explicit?

What do you think?

Best,
Matthias

[1] https://flink.apache.org/roadmap.html#feature-radar

On Wed, Oct 28, 2020 at 9:40 AM Till Rohrmann <[hidden email]> wrote:

> Hi Oleksandr,
>
> yes you are right. The biggest problem is at the moment the lack of test
> coverage and thereby confidence to make changes. We have some e2e tests
> which you can find here [1]. These tests are, however, quite coarse grained
> and are missing a lot of cases. One idea would be to add a Mesos e2e test
> based on Flink's end-to-end test framework [2]. I think what needs to be
> done there is to add a Mesos resource and a way to submit jobs to a Mesos
> cluster to write e2e tests.
>
> [1] https://github.com/apache/flink/tree/master/flink-jepsen
> [2]
> https://github.com/apache/flink/tree/master/flink-end-to-end-tests/flink-end-to-end-tests-common
>
> Cheers,
> Till
>
> On Tue, Oct 27, 2020 at 12:29 PM Oleksandr Nitavskyi <
> [hidden email]> wrote:
>
>> Hello Xintong,
>>
>> Thanks for the insights and support.
>>
>> Browsing the Mesos backlog and didn't identify anything critical, which
>> is left there.
>>
>> I see that there are were quite a lot of contributions to the Flink Mesos
>> in the recent version:
>> https://github.com/apache/flink/commits/master/flink-mesos.
>> We plan to validate the current Flink master (or release 1.12 branch) our
>> Mesos setup. In case of any issues, we will try to propose changes.
>> My feeling is that our test results shouldn't affect the Flink 1.12
>> release cycle. And if any potential commits will land into the 1.12.1 it
>> should be totally fine.
>>
>> In the future, we would be glad to help you guys with any
>> maintenance-related questions. One of the highest priorities around this
>> component seems to be the development of the full e2e test.
>>
>> Kind Regards
>> Oleksandr Nitavskyi
>> ________________________________
>> From: Xintong Song <[hidden email]>
>> Sent: Tuesday, October 27, 2020 7:14 AM
>> To: dev <[hidden email]>; user <[hidden email]>
>> Cc: Piyush Narang <[hidden email]>
>> Subject: [BULK]Re: [SURVEY] Remove Mesos support
>>
>> Hi Piyush,
>>
>> Thanks a lot for sharing the information. It would be a great relief that
>> you are good with Flink on Mesos as is.
>>
>> As for the jira issues, I believe the most essential ones should have
>> already been resolved. You may find some remaining open issues here [1],
>> but not all of them are necessary if we decide to keep Flink on Mesos as is.
>>
>> At the moment and in the short future, I think helps are mostly needed on
>> testing the upcoming release 1.12 with Mesos use cases. The community is
>> currently actively preparing the new release, and hopefully we could come
>> up with a release candidate early next month. It would be greatly
>> appreciated if you fork as experienced Flink on Mesos users can help with
>> verifying the release candidates.
>>
>>
>> Thank you~
>>
>> Xintong Song
>>
>> [1]
>> https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open
>> <
>> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues.apache.org%2Fjira%2Fbrowse%2FFLINK-17402%3Fjql%3Dproject%2520%253D%2520FLINK%2520AND%2520component%2520%253D%2520%2522Deployment%2520%252F%2520Mesos%2522%2520AND%2520status%2520%253D%2520Open&data=04%7C01%7Co.nitavskyi%40criteo.com%7C3585e1f25bdf4e091af808d87a3f92db%7C2a35d8fd574d48e3927c8c398e225a01%7C1%7C0%7C637393760750820881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hytJFQE0MCPzMLiQTQTdbg3GVckX5M3r1NPRGrRV8j4%3D&reserved=0
>> >
>>
>> On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>> Hi Xintong,
>>
>>
>>
>> Do you have any jiras that cover any of the items on 1 or 2? I can reach
>> out to folks internally and see if I can get some folks to commit to
>> helping out.
>>
>>
>>
>> To cover the other qs:
>>
>>   *   Yes, we’ve not got a plan at the moment to get off Mesos. We use
>> Yarn for some our Flink workloads when we can. Mesos is only used when we
>> need streaming capabilities in our WW dcs (as our Yarn is centralized in
>> one DC)
>>   *   We’re currently on Flink 1.9 (old planner). We have a plan to bump
>> to 1.11 / 1.12 this quarter.
>>   *   We typically upgrade once every 6 months to a year (not every
>> release). We’d like to speed up the cadence but we’re not there yet.
>>   *   We’d largely be good with keeping Flink on Mesos as-is and
>> functional while missing out on some of the newer features. We understand
>> the pain on the communities side and we can take on the work if we see some
>> fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in
>> the request to port it over.
>>
>>
>>
>> Thanks,
>>
>>
>>
>> -- Piyush
>>
>>
>>
>>
>>
>> From: Xintong Song <[hidden email]<mailto:[hidden email]>>
>> Date: Sunday, October 25, 2020 at 10:57 PM
>> To: dev <[hidden email]<mailto:[hidden email]>>, user <
>> [hidden email]<mailto:[hidden email]>>
>> Cc: Lasse Nedergaard <[hidden email]<mailto:
>> [hidden email]>>, <[hidden email]<mailto:
>> [hidden email]>>
>> Subject: Re: [SURVEY] Remove Mesos support
>>
>>
>>
>> Thanks for sharing the information with us, Piyush an Lasse.
>>
>>
>>
>> @Piyush
>>
>>
>>
>> Thanks for offering the help. IMO, there are currently several problems
>> that make supporting Flink on Mesos challenging for us.
>>
>>   1.  Lack of Mesos experts. AFAIK, there are very few people (if not
>> none) among the active contributors in this community that are familiar
>> with Mesos and can help with development on this component.
>>   2.  Absence of tests. Mesos does not provide a testing cluster, like
>> `MiniYARNCluster`, making it hard to test interactions between Flink and
>> Mesos. We have only a few very simple e2e tests running on Mesos deployed
>> in a docker, covering the most fundamental workflows. We are not sure how
>> well those tests work, especially against some potential corner cases.
>>   3.  Divergence from other deployment. Because of 1 and 2, the new
>> efforts (features, maintenance, refactors) tend to exclude Mesos if
>> possible. When the new efforts have to touch the Mesos related components
>> (e.g., changes to the common resource manager interfaces), we have to be
>> very careful and make as few changes as possible, to avoid accidentally
>> breaking anything that we are not familiar with. As a result, the component
>> diverges a lot from other deployment components (K8s/Yarn), which makes it
>> harder to maintain.
>>
>> It would be greatly appreciated if you can help with either of the above
>> issues.
>>
>>
>>
>> Additionally, I have a few questions concerning your use cases at Criteo.
>> IIUC, you are going to stay on Mesos in the foreseeable future, while
>> keeping the Flink version up-to-date? What Flink version are you currently
>> using? How often do you upgrade (e.g., every release)? Would you be good
>> with keeping the Flink on Mesos component as it is (means that deployment
>> and resource management improvements may not be ported to Mesos), while
>> keeping other components up-to-date (e.g., improvements from programming
>> APIs, operators, state backens, etc.)?
>>
>>
>>
>> Thank you~
>>
>> Xintong Song
>>
>>
>>
>>
>>
>> On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <
>> [hidden email]<mailto:[hidden email]>>
>> wrote:
>>
>> Hi
>>
>>
>>
>> At Trackunit We have been using Mesos for long time but have now moved to
>> k8s.
>>
>> Med venlig hilsen / Best regards
>>
>> Lasse Nedergaard
>>
>>
>>
>>
>>
>> Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]
>> <mailto:[hidden email]>>:
>>
>> 
>>
>> Hey Piyush,
>>
>> thanks a lot for raising this concern. I believe we should keep Mesos in
>> Flink then in the foreseeable future.
>>
>> Your offer to help is much appreciated. We'll let you know once there is
>> something.
>>
>>
>>
>> On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>> Thanks Kostas. If there's items we can help with, I'm sure we'd be able
>> to find folks who would be excited to contribute / help in any way.
>>
>> -- Piyush
>>
>>
>> On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]<mailto:
>> [hidden email]>> wrote:
>>
>>     Thanks Piyush for the message.
>>     After this, I revoke my +1. I agree with the previous opinions that we
>>     cannot drop code that is actively used by users, especially if it
>>     something that deep in the stack as support for cluster management
>>     framework.
>>
>>     Cheers,
>>     Kostas
>>
>>     On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > Hi folks,
>>     >
>>     >
>>     >
>>     > We at Criteo are active users of the Flink on Mesos resource
>> management component. We are pretty heavy users of Mesos for scheduling
>> workloads on our edge datacenters and we do want to continue to be able to
>> run some of our Flink topologies (to compute machine learning short term
>> features) on those DCs. If possible our vote would be not to drop Mesos
>> support as that will tie us to an old release / have to maintain a fork as
>> we’re not planning to migrate off Mesos anytime soon. Is the burden
>> something that can be helped with by the community? (Or are you referring
>> to having to ensure PRs handle the Mesos piece as well when they touch the
>> resource managers?)
>>     >
>>     >
>>     >
>>     > Thanks,
>>     >
>>     >
>>     >
>>     > -- Piyush
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > From: Till Rohrmann <[hidden email]<mailto:
>> [hidden email]>>
>>     > Date: Friday, October 23, 2020 at 8:19 AM
>>     > To: Xintong Song <[hidden email]<mailto:
>> [hidden email]>>
>>     > Cc: dev <[hidden email]<mailto:[hidden email]>>, user <
>> [hidden email]<mailto:[hidden email]>>
>>     > Subject: Re: [SURVEY] Remove Mesos support
>>     >
>>     >
>>     >
>>     > Thanks for starting this survey Robert! I second Konstantin and
>> Xintong in the sense that our Mesos user's opinions should matter most
>> here. If our community is no longer using the Mesos integration, then I
>> would be +1 for removing it in order to decrease the maintenance burden.
>>     >
>>     >
>>     >
>>     > Cheers,
>>     >
>>     > Till
>>     >
>>     >
>>     >
>>     > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > +1 for adding a warning in 1.12 about planning to remove Mesos
>> support.
>>     >
>>     >
>>     >
>>     > With my developer hat on, removing the Mesos support would
>> definitely reduce the maintaining overhead for the deployment and resource
>> management related components. On the other hand, the Flink on Mesos users'
>> voices definitely matter a lot for this community. Either way, it would be
>> good to draw users attention to this discussion early.
>>     >
>>     >
>>     >
>>     > Thank you~
>>     >
>>     > Xintong Song
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>     >
>>     > Hi Robert,
>>     >
>>     > +1 to the plan you outlined. If we were to drop support in Flink
>> 1.13+, we
>>     > would still support it in Flink 1.12- with bug fixes for some time
>> so that
>>     > users have time to move on.
>>     >
>>     > It would certainly be very interesting to hear from current Flink
>> on Mesos
>>     > users, on how they see the evolution of this part of the ecosystem.
>>     >
>>     > Best,
>>     >
>>     > Konstantin
>>
>

--

Konstantin Knauf

https://twitter.com/snntrable

https://github.com/knaufk

[Yangze Guo]

+1

Best,
Yangze Guo

On Mon, Mar 29, 2021 at 11:31 AM Xintong Song <[hidden email]> wrote:

>
> +1
> It's already a matter of fact for a while that we no longer port new features to the Mesos deployment.
>
> Thank you~
>
> Xintong Song
>
>
>
> On Fri, Mar 26, 2021 at 10:37 PM Till Rohrmann <[hidden email]> wrote:
>>
>> +1 for officially deprecating this component for the 1.13 release.
>>
>> Cheers,
>> Till
>>
>> On Thu, Mar 25, 2021 at 1:49 PM Konstantin Knauf <[hidden email]> wrote:
>>>
>>> Hi Matthias,
>>>
>>> Thank you for following up on this. +1 to officially deprecate Mesos in the code and documentation, too. It will be confusing for users if this diverges from the roadmap.
>>>
>>> Cheers,
>>>
>>> Konstantin
>>>
>>> On Thu, Mar 25, 2021 at 12:23 PM Matthias Pohl <[hidden email]> wrote:
>>>>
>>>> Hi everyone,
>>>> considering the upcoming release of Flink 1.13, I wanted to revive the
>>>> discussion about the Mesos support ones more. Mesos is also already listed
>>>> as deprecated in Flink's overall roadmap [1]. Maybe, it's time to align the
>>>> documentation accordingly to make it more explicit?
>>>>
>>>> What do you think?
>>>>
>>>> Best,
>>>> Matthias
>>>>
>>>> [1] https://flink.apache.org/roadmap.html#feature-radar
>>>>
>>>> On Wed, Oct 28, 2020 at 9:40 AM Till Rohrmann <[hidden email]> wrote:
>>>>
>>>> > Hi Oleksandr,
>>>> >
>>>> > yes you are right. The biggest problem is at the moment the lack of test
>>>> > coverage and thereby confidence to make changes. We have some e2e tests
>>>> > which you can find here [1]. These tests are, however, quite coarse grained
>>>> > and are missing a lot of cases. One idea would be to add a Mesos e2e test
>>>> > based on Flink's end-to-end test framework [2]. I think what needs to be
>>>> > done there is to add a Mesos resource and a way to submit jobs to a Mesos
>>>> > cluster to write e2e tests.
>>>> >
>>>> > [1] https://github.com/apache/flink/tree/master/flink-jepsen
>>>> > [2]
>>>> > https://github.com/apache/flink/tree/master/flink-end-to-end-tests/flink-end-to-end-tests-common
>>>> >
>>>> > Cheers,
>>>> > Till
>>>> >
>>>> > On Tue, Oct 27, 2020 at 12:29 PM Oleksandr Nitavskyi <
>>>> > [hidden email]> wrote:
>>>> >
>>>> >> Hello Xintong,
>>>> >>
>>>> >> Thanks for the insights and support.
>>>> >>
>>>> >> Browsing the Mesos backlog and didn't identify anything critical, which
>>>> >> is left there.
>>>> >>
>>>> >> I see that there are were quite a lot of contributions to the Flink Mesos
>>>> >> in the recent version:
>>>> >> https://github.com/apache/flink/commits/master/flink-mesos.
>>>> >> We plan to validate the current Flink master (or release 1.12 branch) our
>>>> >> Mesos setup. In case of any issues, we will try to propose changes.
>>>> >> My feeling is that our test results shouldn't affect the Flink 1.12
>>>> >> release cycle. And if any potential commits will land into the 1.12.1 it
>>>> >> should be totally fine.
>>>> >>
>>>> >> In the future, we would be glad to help you guys with any
>>>> >> maintenance-related questions. One of the highest priorities around this
>>>> >> component seems to be the development of the full e2e test.
>>>> >>
>>>> >> Kind Regards
>>>> >> Oleksandr Nitavskyi
>>>> >> ________________________________
>>>> >> From: Xintong Song <[hidden email]>
>>>> >> Sent: Tuesday, October 27, 2020 7:14 AM
>>>> >> To: dev <[hidden email]>; user <[hidden email]>
>>>> >> Cc: Piyush Narang <[hidden email]>
>>>> >> Subject: [BULK]Re: [SURVEY] Remove Mesos support
>>>> >>
>>>> >> Hi Piyush,
>>>> >>
>>>> >> Thanks a lot for sharing the information. It would be a great relief that
>>>> >> you are good with Flink on Mesos as is.
>>>> >>
>>>> >> As for the jira issues, I believe the most essential ones should have
>>>> >> already been resolved. You may find some remaining open issues here [1],
>>>> >> but not all of them are necessary if we decide to keep Flink on Mesos as is.
>>>> >>
>>>> >> At the moment and in the short future, I think helps are mostly needed on
>>>> >> testing the upcoming release 1.12 with Mesos use cases. The community is
>>>> >> currently actively preparing the new release, and hopefully we could come
>>>> >> up with a release candidate early next month. It would be greatly
>>>> >> appreciated if you fork as experienced Flink on Mesos users can help with
>>>> >> verifying the release candidates.
>>>> >>
>>>> >>
>>>> >> Thank you~
>>>> >>
>>>> >> Xintong Song
>>>> >>
>>>> >> [1]
>>>> >> https://issues.apache.org/jira/browse/FLINK-17402?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Deployment%20%2F%20Mesos%22%20AND%20status%20%3D%20Open
>>>> >> <
>>>> >> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues.apache.org%2Fjira%2Fbrowse%2FFLINK-17402%3Fjql%3Dproject%2520%253D%2520FLINK%2520AND%2520component%2520%253D%2520%2522Deployment%2520%252F%2520Mesos%2522%2520AND%2520status%2520%253D%2520Open&data=04%7C01%7Co.nitavskyi%40criteo.com%7C3585e1f25bdf4e091af808d87a3f92db%7C2a35d8fd574d48e3927c8c398e225a01%7C1%7C0%7C637393760750820881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hytJFQE0MCPzMLiQTQTdbg3GVckX5M3r1NPRGrRV8j4%3D&reserved=0
>>>> >> >
>>>> >>
>>>> >> On Tue, Oct 27, 2020 at 2:58 AM Piyush Narang <[hidden email]
>>>> >> <mailto:[hidden email]>> wrote:
>>>> >>
>>>> >> Hi Xintong,
>>>> >>
>>>> >>
>>>> >>
>>>> >> Do you have any jiras that cover any of the items on 1 or 2? I can reach
>>>> >> out to folks internally and see if I can get some folks to commit to
>>>> >> helping out.
>>>> >>
>>>> >>
>>>> >>
>>>> >> To cover the other qs:
>>>> >>
>>>> >>   *   Yes, we’ve not got a plan at the moment to get off Mesos. We use
>>>> >> Yarn for some our Flink workloads when we can. Mesos is only used when we
>>>> >> need streaming capabilities in our WW dcs (as our Yarn is centralized in
>>>> >> one DC)
>>>> >>   *   We’re currently on Flink 1.9 (old planner). We have a plan to bump
>>>> >> to 1.11 / 1.12 this quarter.
>>>> >>   *   We typically upgrade once every 6 months to a year (not every
>>>> >> release). We’d like to speed up the cadence but we’re not there yet.
>>>> >>   *   We’d largely be good with keeping Flink on Mesos as-is and
>>>> >> functional while missing out on some of the newer features. We understand
>>>> >> the pain on the communities side and we can take on the work if we see some
>>>> >> fancy improvement in Flink on Yarn / K8s that we want in Mesos to put in
>>>> >> the request to port it over.
>>>> >>
>>>> >>
>>>> >>
>>>> >> Thanks,
>>>> >>
>>>> >>
>>>> >>
>>>> >> -- Piyush
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >> From: Xintong Song <[hidden email]<mailto:[hidden email]>>
>>>> >> Date: Sunday, October 25, 2020 at 10:57 PM
>>>> >> To: dev <[hidden email]<mailto:[hidden email]>>, user <
>>>> >> [hidden email]<mailto:[hidden email]>>
>>>> >> Cc: Lasse Nedergaard <[hidden email]<mailto:
>>>> >> [hidden email]>>, <[hidden email]<mailto:
>>>> >> [hidden email]>>
>>>> >> Subject: Re: [SURVEY] Remove Mesos support
>>>> >>
>>>> >>
>>>> >>
>>>> >> Thanks for sharing the information with us, Piyush an Lasse.
>>>> >>
>>>> >>
>>>> >>
>>>> >> @Piyush
>>>> >>
>>>> >>
>>>> >>
>>>> >> Thanks for offering the help. IMO, there are currently several problems
>>>> >> that make supporting Flink on Mesos challenging for us.
>>>> >>
>>>> >>   1.  Lack of Mesos experts. AFAIK, there are very few people (if not
>>>> >> none) among the active contributors in this community that are familiar
>>>> >> with Mesos and can help with development on this component.
>>>> >>   2.  Absence of tests. Mesos does not provide a testing cluster, like
>>>> >> `MiniYARNCluster`, making it hard to test interactions between Flink and
>>>> >> Mesos. We have only a few very simple e2e tests running on Mesos deployed
>>>> >> in a docker, covering the most fundamental workflows. We are not sure how
>>>> >> well those tests work, especially against some potential corner cases.
>>>> >>   3.  Divergence from other deployment. Because of 1 and 2, the new
>>>> >> efforts (features, maintenance, refactors) tend to exclude Mesos if
>>>> >> possible. When the new efforts have to touch the Mesos related components
>>>> >> (e.g., changes to the common resource manager interfaces), we have to be
>>>> >> very careful and make as few changes as possible, to avoid accidentally
>>>> >> breaking anything that we are not familiar with. As a result, the component
>>>> >> diverges a lot from other deployment components (K8s/Yarn), which makes it
>>>> >> harder to maintain.
>>>> >>
>>>> >> It would be greatly appreciated if you can help with either of the above
>>>> >> issues.
>>>> >>
>>>> >>
>>>> >>
>>>> >> Additionally, I have a few questions concerning your use cases at Criteo.
>>>> >> IIUC, you are going to stay on Mesos in the foreseeable future, while
>>>> >> keeping the Flink version up-to-date? What Flink version are you currently
>>>> >> using? How often do you upgrade (e.g., every release)? Would you be good
>>>> >> with keeping the Flink on Mesos component as it is (means that deployment
>>>> >> and resource management improvements may not be ported to Mesos), while
>>>> >> keeping other components up-to-date (e.g., improvements from programming
>>>> >> APIs, operators, state backens, etc.)?
>>>> >>
>>>> >>
>>>> >>
>>>> >> Thank you~
>>>> >>
>>>> >> Xintong Song
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >> On Sat, Oct 24, 2020 at 2:48 AM Lasse Nedergaard <
>>>> >> [hidden email]<mailto:[hidden email]>>
>>>> >> wrote:
>>>> >>
>>>> >> Hi
>>>> >>
>>>> >>
>>>> >>
>>>> >> At Trackunit We have been using Mesos for long time but have now moved to
>>>> >> k8s.
>>>> >>
>>>> >> Med venlig hilsen / Best regards
>>>> >>
>>>> >> Lasse Nedergaard
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >> Den 23. okt. 2020 kl. 17.01 skrev Robert Metzger <[hidden email]
>>>> >> <mailto:[hidden email]>>:
>>>> >>
>>>> >> 
>>>> >>
>>>> >> Hey Piyush,
>>>> >>
>>>> >> thanks a lot for raising this concern. I believe we should keep Mesos in
>>>> >> Flink then in the foreseeable future.
>>>> >>
>>>> >> Your offer to help is much appreciated. We'll let you know once there is
>>>> >> something.
>>>> >>
>>>> >>
>>>> >>
>>>> >> On Fri, Oct 23, 2020 at 4:28 PM Piyush Narang <[hidden email]
>>>> >> <mailto:[hidden email]>> wrote:
>>>> >>
>>>> >> Thanks Kostas. If there's items we can help with, I'm sure we'd be able
>>>> >> to find folks who would be excited to contribute / help in any way.
>>>> >>
>>>> >> -- Piyush
>>>> >>
>>>> >>
>>>> >> On 10/23/20, 10:25 AM, "Kostas Kloudas" <[hidden email]<mailto:
>>>> >> [hidden email]>> wrote:
>>>> >>
>>>> >>     Thanks Piyush for the message.
>>>> >>     After this, I revoke my +1. I agree with the previous opinions that we
>>>> >>     cannot drop code that is actively used by users, especially if it
>>>> >>     something that deep in the stack as support for cluster management
>>>> >>     framework.
>>>> >>
>>>> >>     Cheers,
>>>> >>     Kostas
>>>> >>
>>>> >>     On Fri, Oct 23, 2020 at 4:15 PM Piyush Narang <[hidden email]
>>>> >> <mailto:[hidden email]>> wrote:
>>>> >>     >
>>>> >>     > Hi folks,
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > We at Criteo are active users of the Flink on Mesos resource
>>>> >> management component. We are pretty heavy users of Mesos for scheduling
>>>> >> workloads on our edge datacenters and we do want to continue to be able to
>>>> >> run some of our Flink topologies (to compute machine learning short term
>>>> >> features) on those DCs. If possible our vote would be not to drop Mesos
>>>> >> support as that will tie us to an old release / have to maintain a fork as
>>>> >> we’re not planning to migrate off Mesos anytime soon. Is the burden
>>>> >> something that can be helped with by the community? (Or are you referring
>>>> >> to having to ensure PRs handle the Mesos piece as well when they touch the
>>>> >> resource managers?)
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > Thanks,
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > -- Piyush
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > From: Till Rohrmann <[hidden email]<mailto:
>>>> >> [hidden email]>>
>>>> >>     > Date: Friday, October 23, 2020 at 8:19 AM
>>>> >>     > To: Xintong Song <[hidden email]<mailto:
>>>> >> [hidden email]>>
>>>> >>     > Cc: dev <[hidden email]<mailto:[hidden email]>>, user <
>>>> >> [hidden email]<mailto:[hidden email]>>
>>>> >>     > Subject: Re: [SURVEY] Remove Mesos support
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > Thanks for starting this survey Robert! I second Konstantin and
>>>> >> Xintong in the sense that our Mesos user's opinions should matter most
>>>> >> here. If our community is no longer using the Mesos integration, then I
>>>> >> would be +1 for removing it in order to decrease the maintenance burden.
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > Cheers,
>>>> >>     >
>>>> >>     > Till
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > On Fri, Oct 23, 2020 at 2:03 PM Xintong Song <[hidden email]
>>>> >> <mailto:[hidden email]>> wrote:
>>>> >>     >
>>>> >>     > +1 for adding a warning in 1.12 about planning to remove Mesos
>>>> >> support.
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > With my developer hat on, removing the Mesos support would
>>>> >> definitely reduce the maintaining overhead for the deployment and resource
>>>> >> management related components. On the other hand, the Flink on Mesos users'
>>>> >> voices definitely matter a lot for this community. Either way, it would be
>>>> >> good to draw users attention to this discussion early.
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > Thank you~
>>>> >>     >
>>>> >>     > Xintong Song
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     >
>>>> >>     > On Fri, Oct 23, 2020 at 7:53 PM Konstantin Knauf <[hidden email]
>>>> >> <mailto:[hidden email]>> wrote:
>>>> >>     >
>>>> >>     > Hi Robert,
>>>> >>     >
>>>> >>     > +1 to the plan you outlined. If we were to drop support in Flink
>>>> >> 1.13+, we
>>>> >>     > would still support it in Flink 1.12- with bug fixes for some time
>>>> >> so that
>>>> >>     > users have time to move on.
>>>> >>     >
>>>> >>     > It would certainly be very interesting to hear from current Flink
>>>> >> on Mesos
>>>> >>     > users, on how they see the evolution of this part of the ecosystem.
>>>> >>     >
>>>> >>     > Best,
>>>> >>     >
>>>> >>     > Konstantin
>>>> >>
>>>> >
>>>
>>>
>>>
>>> --
>>>
>>> Konstantin Knauf
>>>
>>> https://twitter.com/snntrable
>>>
>>> https://github.com/knaufk