DX NetOps

Since we went to Spectrum 10.2.1 we've had issue after issue with Archive Managers on multiple landscapes:

1) After a an auto-reboot after WSUS has done it's thing, despite "Auto Stop/Start Archive Manager" being checked, once SS has restarted, Archive Manager remains down.  There's no .DDMDB.LOCK file hanging over and it appears to have shut down correctly.  Even when we do a manual reboot for whatever reason, this setting is seeminly ignored, although I do see "Can't connect to MySQL server" in the tail of the log so I am suspecting that the process is attempting to start ArchMgr before all processes have come up. Is there a way to delay this?

2) On some Archive Managers, they have a habit of randomly stopping.  On some servers we've tracked this down to DEP being re-enabled by WSUS patching, but I had one this morning (and have had in the past) where our Service Desk has logged a ticket for Archive Manager being down.  The log doesn't always give much info either, save for "A stop request has been received."

Anyone else had issues with ArchMgr on 10.2.1 (Windows 2012 R2)

Thanks!

The first point is not something very important. It will keep retrying.

For the second one: did you check the database is not corrupted? Did it stop properly previously? Did you try to check it with mysqlcheck?

First point *is* important as the scheduled reboot is at 4:00am and our Service Desk then log a ticket about 30 minutes later when ArchMgr still isn't up.  I then start work at 9:00am and it's still down. If it's retrying, it's not trying very hard.

Yes - there are no database corruptions and the DDMDB shuts down cleanly before reboot.

Is mysql up and running after a reboot or do you have to start it manually?

Check the $SPECROOT/SS/lib/SDPM/pprocessd_log file for any errors.

Any changes made to any of the idb files located in the $SPECROOT/SS/lib/SDPM/partlist directory?

If mysql is coming up but taking a while, try modifying the ARCHMGR.idb file and increase the RETRYTIMEOUT and RETRYMAX parameters.

Joe

Hello David,

Did the above suggestions help? Please let us know the latest update with Archive Manager.

Thanks.

Yan

Hi David,

I would suggest to open a support issue and request PTF_10.2.156, which contains the latest fixes for Archive Manager for Spectrum 10.2.1.

The patch includes two Archive Manager fixes:

Symptom: Archive Manager crash after client disconnect.
Resolution: Corrected code faults causing disconnected clients to be cleaned up prematurely.
(DE280357, 00672723 )

Symptom: Excessive Archive Manager memory usage, due to high numbers of attached event clients.
Resolution: Improved Archive Manager handling of event attachments, resulting in significantly
reduced memory consumption.

(DE304156, 00783125 )

Best regards,

Glenn