Replication Queue stuck

Post by **mark** » Tue Jul 24, 2007 3:38 pm

Are either or both appliances under heavy load? Check in Maintenance->tech support for the "load" and for any processes using a lot of cpu or memory.

Jit · Post by **Jit** » Tue Jul 24, 2007 3:43 pm

They shouldn't be as I'm the only person using them.

As mentioned above, the /replsend.txt for the profile is using up ~60% CPU on #1 (others are 7% and 2x 3%). "monitor/Check index" is taking up about 30% CPU time on #2. The rest of the processes are trivial for #2.

Jit · Post by **Jit** » Wed Jul 25, 2007 11:31 am

Any other suggestions?
Should I open a support ticket?

Post by **John** » Wed Jul 25, 2007 12:47 pm

Yes, if you open a ticket we should be able to determine the exact cause and get a fix.

michel.weber · Post by **michel.weber** » Thu Jul 26, 2007 5:53 pm

We have had similar problems.

One item suggested by support was to set verbosity to é (if it is set higher) but that did not help that much.

The indexing itself was not that slow in our place, but once you have several 10000 items in your replication queue it empties very slowly (several weeks for 200 000 items).

I have noticed that if a site is 'clean' (ie no errors) the replication queue tends to stay small. But if there are a lot of errors the queue size explodes.

Hope this helps

Post by **John** » Fri Jul 27, 2007 9:56 am

We have found the issue, and the next script update will address large replication queues, and 300,000 items now can clear out in about 1 day.