cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
Veritas.com
Support

The requested journal set could not be started

Paul_T
Level 4

‎08-08-2007 06:20 AM

We are running VRE3.1 between a pair of file servers across our WAN.  Due to various server lock-up problems, we have been advised to set aschedule so that replication is not active during our AV sweep and backup window, and so does not run between 02:00 and 06:00.  Unfortunately, it frequently fails to kick off around 06:00 giving errors that "pair did not start successfully.   This is frustrating but can usually be fixed by restarted the VRE agent on the target server. 
 
QUESTION 1:  Why would this be happening ?
 
Today, having restarted the services again, I have one job that is only "running partially" and then "rallying".  It makes no difference if I restart the services on source and target nor if I delete the job and reconfigure it from scratch.
 
QUESTION 2:  How can i find out why this job is struggling and get it running cleanly ?
 
 
I note quite a few threads re: server lockup.  Ours is typically the source server.  We run McAfee VRE 7.1 and have been told that may be the problem.  I shall read some of these threads with interest.
 
 
0 Kudos
Reply
1 REPLY 1

MKnoll
Level 3

‎09-19-2007 12:27 PM

I've had that problem on a couple of sites too.
 
Some things I've experienced for jobs running partially:
 
First I'd check your rules, makes sure the rules match existing drives on both sides.  If a rule references a bad path, but the pair can still communicate, you will get the "running partially" status.  And the rallying status just means VRE is trying to execute the job over.
 
Another thing I've seen cause these symptoms is flaky network connections causing these symptoms.  Bad nics, or a nic with speed & duplex not matching the settings on the switch will get irregular connections.  So the pair will connect but when it tries to transfer data it errors out.
 
For the journalling issue.
 
Try running rxdbvalidate to check the databases on each end and make sure neither is corrupt.  On the client side run rxdbvalidate -f -v RSA.  Also run that command on the RMS side, then run rxbvalidate -v -f RMS.   


Message Edited by MKnoll on 09-19-2007 12:29 PM

Message Edited by MKnoll on 09-19-2007 12:32 PM
0 Kudos
Reply