Tiemrs and restore

[Alberto Mancini]

Hello,

according to this stackoverflow response 

https://stackoverflow.com/questions/36306136/will-apache-flink-restore-trigger-timers-after-failure

IIUC we should expect that after a restore the timers will be not executed until a new timer is scheduled. 

I wonder if this is still true and if there is any chance of forcing the restart of the timer task.

Thank you. 

Regards,

   Alberto. 

[Tzu-Li (Gordon) Tai]

Hi Alberto,

Looking at the code, I think the current behavior is that all timers (both processing time and event time) are re-registered on restore, and therefore should be triggered automatically.

So, for processing time timers, on restore all timers that were supposed to be fired while the job was down should fire automatically; for event time timers, they will be triggered once the watermark passes their timestamps.

Also looped in Aljoscha on this, in case I misunderstood anything.

Cheers,

Gordon

On 16 April 2018 at 1:20:00 AM, Alberto Mancini ([hidden email]) wrote:

Hello,

according to this stackoverflow response 

https://stackoverflow.com/questions/36306136/will-apache-flink-restore-trigger-timers-after-failure

IIUC we should expect that after a restore the timers will be not executed until a new timer is scheduled. 

I wonder if this is still true and if there is any chance of forcing the restart of the timer task.

Thank you. 

Regards,

   Alberto. 

[Aljoscha Krettek]

Gordon is correct: there was a bug on a very old version of Flink that caused processing-timers not to be invoked after restore but that was fixed.

Aljoscha

On 16. Apr 2018, at 06:20, Tzu-Li (Gordon) Tai <[hidden email]> wrote:

Hi Alberto,

Looking at the code, I think the current behavior is that all timers (both processing time and event time) are re-registered on restore, and therefore should be triggered automatically.

So, for processing time timers, on restore all timers that were supposed to be fired while the job was down should fire automatically; for event time timers, they will be triggered once the watermark passes their timestamps.

Also looped in Aljoscha on this, in case I misunderstood anything.

Cheers,

Gordon

On 16 April 2018 at 1:20:00 AM, Alberto Mancini ([hidden email]) wrote:

Hello,

according to this stackoverflow response 

https://stackoverflow.com/questions/36306136/will-apache-flink-restore-trigger-timers-after-failure

IIUC we should expect that after a restore the timers will be not executed until a new timer is scheduled. 

I wonder if this is still true and if there is any chance of forcing the restart of the timer task.

Thank you. 

Regards,

   Alberto. 

[Alberto Mancini]

Hi Gordon and Aljoscha,

thank you. 

I updated the stackoverflow response too with a link to this thread to make life easier to anyone with a similar question. 

Thanks.

  A. 

On Mon, Apr 16, 2018 at 6:30 PM Aljoscha Krettek <[hidden email]> wrote:

Gordon is correct: there was a bug on a very old version of Flink that caused processing-timers not to be invoked after restore but that was fixed.

Aljoscha

On 16. Apr 2018, at 06:20, Tzu-Li (Gordon) Tai <[hidden email]> wrote:

Hi Alberto,

Looking at the code, I think the current behavior is that all timers (both processing time and event time) are re-registered on restore, and therefore should be triggered automatically.

So, for processing time timers, on restore all timers that were supposed to be fired while the job was down should fire automatically; for event time timers, they will be triggered once the watermark passes their timestamps.

Also looped in Aljoscha on this, in case I misunderstood anything.

Cheers,

Gordon

On 16 April 2018 at 1:20:00 AM, Alberto Mancini ([hidden email]) wrote:

Hello,

according to this stackoverflow response 

https://stackoverflow.com/questions/36306136/will-apache-flink-restore-trigger-timers-after-failure

IIUC we should expect that after a restore the timers will be not executed until a new timer is scheduled. 

I wonder if this is still true and if there is any chance of forcing the restart of the timer task.

Thank you. 

Regards,

   Alberto.