I have attached sql trace file which contains block process report and query
involved in block process.

In order to resolve these blocks we have tried below solutions till now.

1) SELCT query with NOLOCK.
We Modified source code which will now fire select query with NOLOCK as
stated below.

select REVISION_ID, JOURNAL_ID, PRODUCER_ID, REVISION_DATA from
XX_DMS_JRNL_JOURNAL with (NOLOCK) where REVISION_ID >  301810  order by
REVISION_ID. 

It reduced select blocks but still update blocks are seen.


2) We have changed the revision journal mechanism for cluster from database
journal to file journal but performance for document upload using Webload
was poor.

Total documents uploaded in case of database journal for 30 users load test:
600 (around) documents.
Total documents uploaded in case of file journal for 30 users load test:
416(around) documents.
Hence for removing blocks we used file journal but then performance is
degraded. 

We still are facing these issues .

Kindly suggest.

http://jackrabbit.510166.n4.nabble.com/file/n4189670/Block-Update_Select.txt
Block-Update_Select.txt 

--
View this message in context: 
http://jackrabbit.510166.n4.nabble.com/Database-blocks-are-observed-while-doing-document-upload-in-jackrabbit-with-3-cluster-nodes-tp4189060p4189670.html
Sent from the Jackrabbit - Dev mailing list archive at Nabble.com.

Reply via email to