Synchronization unable to establish a peer after purge on loggers completed

[Amer rajai Sha'er]

Hello,

I am having a serious problem that the HDS synchronization is unable to establish peer to the loggers after the purge is completed.

Any idea's.

It's also causing the AW client to not be able to update the database.

I have attached the event viewer snapshots from ICMA , ICMB and the HDS server.

Thanks

Amer

[geoff]

This doesn't look good. The problem is not with your HDS, though it is impacted.

Are you saying that after the nightly purge job (at around 30mins past midnight), the historical logger on one side went down then the other side went down and the whole thing is messed up? That's what the events from RouterB seem to show - first historical logger on side B goes down, then the historical logger on side A. I see the times there.

Are you seeing a minidump in the lb\logs directory? This would indicate a crash.

What are the last lines from the EMS file for the historical logger before it goes down and restarts? There must be more evidence than you are providing. The Event Viewer is not a diagnostic tool - let's see the trace.

Regards,

Geoff

[Amer rajai Sha'er]

Hello Geoff,

Yes , you are right , after the purge (which is 00:30) everyday , the both loggers goes down and they need around 30 minutes to go back.

I am not now at the customer side , i will get the traces and show them to you , do you have even a clue for what is going on?

Amer.

[geoff]

Amer,

I don't have a clue - I've not seen this before. But to be honest, I have not studied the purge mechanism.

Is it possible that this is OK behaviour? Is there some requirement for the logger to stop for the purge? Does the logger stay down for 30 minutes (say till 1am) then just start up perfectly? Is there some setting that controls this? Just saying ....

You would not think that stopping the logger would be required for a purge of historical data, would you? How would 24x7 call centers function if this were the case? I suppose loggerA could purge at 0:30 and go down, but loggerA remains up and the CC remains functional. Then at 1am loggerB does its thing?

As above .... I've not studied purge in detail. If you look at the SQL Agent in SQL Management Studio and select one of the purge jobs and check the job history, have they worked OK?

Regards,

Geoff

[Amer rajai Sha'er]

Hello Geoff,

Yes , they are ok , all of the purges were succeeded .

I tried to dumplog today after enabling the traces yesterday , i got this error message :

“The sideB Logger cannot come online in duplexed mode because its database is out of date.”

and i also found another warning on the cmsnode , check the attached.

Amer