Task stop timeout occurred. Best timestamp to restart task from?

[MoeyE]

Hi Support,

I recently encountered a task stop timeout, and I read this article which suggests that there could have been changes lost,

https://community.qlik.com/t5/Official-Support-Articles/Qlik-Replicate-Task-stop-timeout-occurred/ta...

Would there be any issues with starting the task from the "last source timestamp" that is in the log file when a task is stopped? Would I still need to account for latency if I do this?

My understanding is that this "last source timestamp" is the last point at which the task is reading from the source logs.

Regards,

Mohammed

[Yadunandan]

Hello Mohammed,

Yes, you need to consider the latency at the time task was stopped.

Regards,
Yadunandan

[MoeyE]

Hi Yadunandan, 

What is the recommended action when you don't know the latency at the point when the task was stopped.

Also, I have the "last source timestamp" which is the point at which the task was reading last, which is different from the point at which the task was stopped. My current thought is to simply use the "last source timestamp", but go 30mins to 1 hour back just for good measure.

Regards,

Mohammed

[Heinvandenheuvel]

Just FWIW - 1725230237000000 = GMT Sunday, September 1, 2024 22:37:17 

Courtesy : https://www.epochconverter.com/

You may want to to restart at that time minus the typical latency - or perhaps the latency according to PERFORMANCE LOGGING TRACE - this is a good option to have active all the time, perhaps except fr Oracle sources where it gets a bit verbose.

Of course you should also carefully peruse the last task log for an indication as to why it did not want to stop. Typically this because of a long running bulk target apply - the target DB may have clues as to why that was happening.

Hein

[MoeyE]

Hi Hein,

Thanks for the clarification, much appreciated. Yes the source is Oracle which is why we don't recommend to keep it on like the other source endpoints.

Regards,

Mohammed