ImpExImportJob stuck in RUNNING state

[Former Member]

Hello

I got an ImpExImport job which is created from a publication from Datahub. According to the following lines, the job has been started:

 12:01:46,629 26-06-2018 [[8797928555446]] INFO  [mpex.impl.DefaultImportService]  Starting import synchronous using cronjob with code=0000AFVU
 12:01:46,669 26-06-2018 [[8797928555446]] INFO  [ex.jalo.cronjob.ImpExImportJob] (0000AFVU)  Starting multi-threaded ImpEx cronjob "ImpEx-Import" (4 threads)

The job is now still in RUNNING state after 3 hours without any additional logging. When I use the impex-media from the job and run it locally or on a second server node, the duration is around 5min.

So, my questions are the following:

1. How do I see if the cronjob is still doing something?

2. How can I make him stop?

3. What might be reasons it takes so long?

ImpEx-Importjob is still hybris-default.

Hybris/Datahub version is: 6.0.0.1

Regards

Andy

[Former Member · Accepted solution]

The problem seems to be solved after patching hybris to the latest 6.0.0.27 version. I suspect one of those patches solved the issue:

https://jira.hybris.com/browse/PATCH-4405

https://jira.hybris.com/browse/ECP-1915

https://jira.hybris.com/browse/ECP-1534

https://jira.hybris.com/browse/PATCH-2993

[former_member633554]

To make it stop you probably have to restart hybris.

i've also been able to remove it with an impex

[Marko_salonen]

Yes, as writes, check the JDBC logs, these will give you a lot of information. And take thread dumps, there you can see where the threads are stuck. If threads are stuck in the sql layer, then you know that you have index problems.

[Former Member]

Try setting the threads to 1 to see if it's a concurrency/deadlock issue. Also turn on JDBC logging to see what's happening on the node executing the job.

There is a pretty short list of things that happen in ImpEx import jobs' workload: maybe translators, maybe interceptors, SQL select/update/insert statements.