Home » Server Options » Replication » multimaster replication blocked (Oracle, 9.2.0.5, Linux Red Hat)
multimaster replication blocked [message #282567] Thu, 22 November 2007 06:47 Go to next message
krok
Messages: 3
Registered: November 2007
Junior Member
Hello,
we have some problem on master node of a 2 nodes multimaster configuration.
(After a veri high update on some replicated tables that caused ORA-1653: unable to extend table SYSTEM.DEF$_AQCALL tabs SYSTEM )
Now I see on node 1 all tables in all repgroup in "Need Generation" state, while on the second node they are all marked as valid.
There is a lot of transations (700000) from 1 that don't go pushed on 2.
Considerig that the application now is using the first node and I can't suspend the application what do you suggest to do to minimize (or avoid) the downtime?

Thanks to all.

FC
Re: multimaster replication blocked [message #282614 is a reply to message #282567] Thu, 22 November 2007 09:42 Go to previous messageGo to next message
prsahe
Messages: 2
Registered: November 2007
Location: Switzerland
Junior Member

Difficult situation indeed. Did you increase the space for the SYSTEM tablespace ? What happens if you try to push the transactions manually ?
Re: multimaster replication blocked [message #283122 is a reply to message #282614] Mon, 26 November 2007 03:04 Go to previous messageGo to next message
krok
Messages: 3
Registered: November 2007
Junior Member
Hi,
yes, I increased the system tablespace.
If I try to push manually the session seems to hang.
The push job is in a broken state.
For now we are surviving using only the master node and I'm waiting on a SR.
Do you now an "official" way to drop and recreare all replication catalog?
Re: multimaster replication blocked [message #283234 is a reply to message #283122] Mon, 26 November 2007 07:28 Go to previous messageGo to next message
prsahe
Messages: 2
Registered: November 2007
Location: Switzerland
Junior Member

Hia, my major problem was always the locks. Might be good the check them. The push job is broken because it tried 16 times to execute the job. Check the alert_log on both database servers. And if worst comes to worst you will need to suspend the replication, fix the objects and then resume.
Re: multimaster replication blocked [message #284716 is a reply to message #283234] Fri, 30 November 2007 11:52 Go to previous message
krok
Messages: 3
Registered: November 2007
Junior Member
Hi,
thanks for the answer.
Can suggest me any refrerence whith practical examples and best practices about advanced replication?
If the slave instance becomes unavailable which is the the procedure to dropp all replication support avoiding the accumulation of transactions on the master?
Can I drop all replica catalog and recreate it in a second time?

Previous Topic: Can't refresh Materialized View
Next Topic: Using Replication for Backup
Goto Forum:
  


Current Time: Thu Mar 28 15:20:49 CDT 2024