IPCC/UCCX/CRS 5 HA cluster in PARTIAL_SERVICE even though all configured services ok

Answered Question
Jul 8th, 2010
User Badges:

I have normally worked on  release 7 IPCC/CRS/UCCX so have not seen this before.


Although all configured services are IN_SERVICE, the 5.0(1)SR02_Build026 HA cluster shows as being in PARTIAL_SERVICE.


On both servers:

The CRS Engine's Subsystem Manager is in PARTIAL_SERVICE but all configured services are IN_SERVICE

i.e. the Database Subsystem, eMail Subsystem, MRCP ASR Subsystem and MRCP TTS Subsystem are not used so are not configured


On the subscriber server

The CRS Cluster View Daemon’s Manager Manager is in PARTIAL_SERVICE but all configured services are up

The Archive Manager is not used



Is this normal?


Thanks,

            Terry

Correct Answer by Aaron Harrison about 6 years 8 months ago

Hi Terry


Yeah, you tend to see that the components under the snappily  titled 'Manager Manager' and other parts either appear as greyed or running - presumably because they aren't specifically stoppable. They can be tricky to figure out why they're not running, not least of all because it's not obvious what functions they each perform.


MRCP/DB is normal to be not running, but that Manager/Manager should be fully running.


Aaron

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
Aaron Harrison Thu, 07/08/2010 - 23:59
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

Hi Terry


It's certainly normal for 'partial_service' to appear, and usually MRCP/DB subsystems are the reason. In 7.0 they have the little 'u' unconfigured state which is new...


The archive manager should be running I believe; I would probably bounce the secondary and see if it comes up (the only thing that restarting the secondary should affect is outbound which will pause until the restart is done, and any HRC clients pointed at the secondary server).


Regards


Aaron


Please rate helpful posts...

terry.Haebich Fri, 07/09/2010 - 00:19
User Badges:

Hi Aaron,

              Thanks for the reply - it is good to get confirmation. I couldn't find anything searching the Cisco support site - I would have expected it to be documented somewhere if this was "the norm"


Both servers were restarted this morning which enabled me to finally find the reason the Unified CM Telephony Subsystem service was in PARTIAL_SERVICE (some problem with a test Trigger so I deleted it)


With regards to the Archive Manager, the > shows as "browned out" which is what I understand as a service being "not configured". Again, not knowing much about 5, I assume this is "the norm" but will look at the MCVD logs from this morning restart.


Rgds,

         Terry

Correct Answer
Aaron Harrison Fri, 07/09/2010 - 00:37
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

Hi Terry


Yeah, you tend to see that the components under the snappily  titled 'Manager Manager' and other parts either appear as greyed or running - presumably because they aren't specifically stoppable. They can be tricky to figure out why they're not running, not least of all because it's not obvious what functions they each perform.


MRCP/DB is normal to be not running, but that Manager/Manager should be fully running.


Aaron

terry.Haebich Fri, 07/09/2010 - 01:24
User Badges:

Hi Aaron,

              You are correct


In between our correspondence, I had a Cisco TAC person connect and use CET to delete the entries in


com.cisco.crs.cluster.config.ArchiveAdminConfig

com.cisco.crs.cluster.config.ArchiveIdConfig

com.cisco.crs.cluster.config.ArchiveLeaseConfig


He said we had to do a Node Manager restart but then Archive Manager on the Subscriber came good. Unfortunately, it then failed again so looks like I do need to do the restart - need to organise an outage


Thanks for your response


---------------------------------------------------------------


Although the above is the correct procedure, this CET fix and Node manager restarts did not resolve the problem as it appears this system is a victim of bug CSCsi78765 - last night I was supplied ES patch crs5.0.1sr2es10 which I have yet to apply

{Defect fixed in CRS 5.0(1) sr2 es08 - CSCsi78765 - ArchiveManager remains OOS after rebooting a node}


Also, it is worth pointing out that Node Manager on the Primary server would not restart because of the following error (seen in the MCVD log) even though the Archive Manager service that is OOS is on the Standby server - I had to reboot the Primary server after waiting several minutes:

Jul 13 06:43:27.475 EST %MCVD-ARCHIVE_MGR-7-UNK:SessionThread.abort(com.cisco.archive.impl.ArchiveFailureException: Archive Manager is shutting down on node #1)

I didn't even try to restart Node Manager on the Standby server due to problem with the Primary -  I just rebooted it.

I wasn't monitoring the logs so I wasn't aware of this problem until later when I had time to look at why Node Manager wouldn't restart

I will provide further update/s when I have applied the patch

smartns04 Sun, 09/23/2012 - 22:36
User Badges:

Hi,


I have exactly the same issue on  a 8.0.2 version, did you finally sort out the issue with the patch?

Actions

This Discussion