[Netarchivesuite-devel] Reconnect NAS to running jobs

[Nicholas Grooss Clarke]

And not if you look directory in the job definition.

Sometimes the job is set to failed but is showing as running in the running jobs pages.

Anyway if the the job is not set to failed yet it should be a bit easier to re-register it.

I'm guessing that progress reports for jobs that are not "known" to be running are silent ignored.

And the running jobs page "only" shows what is in the progress reports table.

So somehow from progress report it should be possible to re-register jobs which appear to still be running.

This could maybe be done fairly easily. Someone would need to look in the code to see if progress reports are just ignored for jobs that the GUI does not think are running.


Best

Nicholas

________________________________
From: Netarchivesuite-devel <netarchivesuite-devel-bounces at ml.sbforge.org> on behalf of aponb at gmx.at <aponb at gmx.at>
Sent: Wednesday, April 11, 2018 11:14:45 AM
To: netarchivesuite-devel at ml.sbforge.org
Subject: Re: [Netarchivesuite-devel] Reconnect NAS to running jobs

Hi Nicholas,

thanks for your quick response!

The job was not set to failed after restarting NAS. Actually the ob is running and also appears in the running jobs overview, but it is not showing the correct progress and queued files and so on.

Yeah, that would be nice if NAS would check for running jobs and could reconnect to it. Doesn't need to be automatically. A manual step would be ok. But I understand that it seems to be a lot of work to redesign the current behavior.

Regards
a.



On 2018-04-11 10:38, Nicholas Grooss Clarke wrote:

Hi Andreas


When you did a restart did the job get set to failed and also does not appear in the running jobs overview?


Basically you would like the GUI to check for running jobs when it is restarted?


One problem with this is that the harvestcontrollers do not listen to JMS when busy.

The H3 monitor does not yet have a list of all hosts/ports where H3 instances are configured to run.

So currently the H3 monitor only read the state of jobs from the database to see want to monitor.


I think Søren changed the GUI code at some point so HarvestControllers reconnected with the GUI when it gets a message from an "unknown" source.


I have no idea how difficult it would be to change the GUI/Jobscheduler to identify running job during a restart.

I would need to have a grace period of 5-10 minuttes for progress reports to "magically" appear to know an orphaned job needs to be registered.


Redesigning the jobscheduler would make this functionality straightforward. If/when that happens.

Theoretically it should also be possible to restart the HarvestController and reconnect to the H3 job.

But not with the existing design.


Best

Nicholas

________________________________
From: Netarchivesuite-devel <netarchivesuite-devel-bounces at ml.sbforge.org><mailto:netarchivesuite-devel-bounces at ml.sbforge.org> on behalf of aponb at gmx.at<mailto:aponb at gmx.at> <aponb at gmx.at><mailto:aponb at gmx.at>
Sent: Wednesday, April 11, 2018 10:04:50 AM
To: netarchivesuite-devel at ml.sbforge.org<mailto:netarchivesuite-devel at ml.sbforge.org>
Subject: [Netarchivesuite-devel] Reconnect NAS to running jobs


I had to restart NAS (5.4) without jmsbroker , just the running Heritrix /HarvestControllerApplications kept alive. After restarting NAS I am not getting the services of "Details and Actions on Running Job"-Page on http://nasurl/History/history/job/xxx/

I am only receiving

Details and Actions on Running Job xxx
NAS Job xxx is in state FAILED.

I only can access the job directly via the heritrix3 WebConsole.

Is there any chance to reconnect the NAS-App with this running job?

Regards
a.




_______________________________________________
Netarchivesuite-devel mailing list
Netarchivesuite-devel at ml.sbforge.org<mailto:Netarchivesuite-devel at ml.sbforge.org>
https://ml.sbforge.org/mailman/listinfo/netarchivesuite-devel


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://ml.sbforge.org/pipermail/netarchivesuite-devel/attachments/20180411/d2cf93c4/attachment-0001.html>