Full index build

[kaya]

Hi,

I am currently trying a full index build and Alfresco is stuck at :

2010-04-21 14:05:12,879 INFO  [node.index.FullIndexRecoveryComponent] Index recovery started: 13,815,602 transactions.

It's been almost 3 hours, and it doesn't seem to be any activity on the filesystem.
Before the rebuild, the lucene_indexes folder's size was 8GB. I emptied it and started the rebuild, now it's been at 214 MB for 2 hours.

Is it normal ? Shoul I wait anyway ?
I seem to remember that the waiting part was after the 100% stage ….

[derek]

Hi,
Please provide version details.
There are a few things that can be done to limit the work done by the rebuild.
[edit] Also, you could take a jstack of the running system to see what it is up to.
Regards

[kaya]

Hi,

Thanks for your reply.
We are running an Alfresco 2.1.7 Entreprise Edition with postrgres on a Solaris 10 server
There was progress in the night :

2010-04-21 14:05:12,879 INFO  [node.index.FullIndexRecoveryComponent] Index recovery started: 13,815,602 transactions.
2010-04-21 23:22:55,077 INFO  [node.index.FullIndexRecoveryComponent]   10 % complete.
2010-04-22 03:18:04,075 INFO  [node.index.FullIndexRecoveryComponent]   20 % complete.

and we ve been stuck here for more than 7 hours ..

Before starting alfresco, we added an index on alf_transaction as told here :

https://issues.alfresco.com/jira/browse/ETWOTWO-797

It was supposed to be 28 times faster with this index ..