RE: Data Guard Hanging
Date: Fri, 5 Dec 2008 11:30:37 -0500
That's what it looks like to me too. I was doing legacy load on the primary database that started with a sqlloader, a truncate, and another sqlloader. I think the truncate was killing me. I just skipped all dml to that loader table and I got by a ton of archivelog files. Now I am processing the actual load of the data I think. But this is still too slow.
Gotta find another way. Did you try the DBMS_LOGSTDBY.INSTANTIATE_TABLE option to bypass these load problems and just bring the table across intact? Does this make sense?
From: Mark Strickland [mailto:strickland.mark_at_gmail.com]
Sent: Friday, December 05, 2008 11:19 AM
To: Mercadante, Thomas F (LABOR)
Subject: Re: Data Guard Hanging
Tom, in 10.1.0.5, I had a severe problem with very large transactions taking a VERY long time to propagate to the logical standby. For example, an update of 1,000,000 rows with a single update statement that might take 5 minutes to complete in the primary could take days to propagate to the standby. I went around and around with Oracle Support and they eventually acknowledged that there was likely a bug, but there was never a resolution. I don't know about 10gR2. I'm no longer at that company (went under) and we're not using logical standby at my current company. Do a search of Oracle-L for my thread about this in 2006-2007.
On Fri, Dec 5, 2008 at 5:36 AM, Mercadante, Thomas F (LABOR) <Thomas.Mercadante_at_labor.state.ny.us> wrote:
I am running a Data Guard with a Logical Standby on Aix V 10.2.0.4.
This is our first run at this so we are still learning.
During my first large data load in the primary database, I ran out of room in my tablespaces. The load died and I restarted it after fixing the tablespaces. I then noticed that the DDL did not change the tablespaces in the Logical Standby. I realized that this does not happen automatically because the Apply died. I changed the tablespaces in the Standby and restarted Apply. Everything seemed to work fine. The data all got propagated over. But now the Apply process seems to be hung - or very slow. We are about 150 archivelogs behind. All of the monitoring sql's that Oracle Metalink gives us shows that the apply process is just running "EXECUTE TRANSACTION" - I can't see any sql statements being run.
Anybody seen anything like this? We have an SR submitted with Oracle and the have been actively asking questions but nothing so far.
TomReceived on Fri Dec 05 2008 - 10:30:37 CST