[
http://jira.jboss.com/jira/browse/JBMESSAGING-1359?page=comments#action_1... ]
Tim Fox commented on JBMESSAGING-1359:
--------------------------------------
I'm not really understanding why this is a JBM issue.
It's my understanding that the table only contains 5000 rows, and this is verifiable
by doing a select count(*) from JBM_MSG?
Each message is 1K so that should take up around 5MB of data.
Instead you're seeing that the database is taking up 2GB on the disk?
If so, then surely this points towards a MSSQL issue with reclaiming space? JBM only uses
the JDBC api, it seems it has deleted the records ok - since we can see there are only
5000, but MSSQL has not cleaned up the space.
If JBM has successfully executed DELETE statements and the rows are not there any more as
reported by the query, I'm not sure what else it can do.
Unless there are some extra non standard commands the JBM needs to execute with MSSQL, but
I am not aware of those.
JBM is not able to work with MSSQL - MSSQL not cleaning deleted rows
from JBM_MSG table, not reclaiming disk space
------------------------------------------------------------------------------------------------------------------
Key: JBMESSAGING-1359
URL:
http://jira.jboss.com/jira/browse/JBMESSAGING-1359
Project: JBoss Messaging
Issue Type: Bug
Components: Messaging Core Persistence
Affects Versions: 1.4.0.SP3.CP02
Environment: SOA-P 4.2 CP02, RHEL5, MSSQL
Reporter: Martin Vecera
Assigned To: Tim Fox
Priority: Critical
Attachments: dbstat.csv.gz, jbm_msg.csv.gz, soaesb1_dbstat.sql.gz
Reproducible: 100%
Description: There is something that prevents MSSQL to clean deleted rows from JBM_MSG
table and the disk space occupied by that table grows permanently.
I had HelloWorld ESB service listening on JMS queue. I was constantly sending messages to
that queue. JBM_MSG table contains around 1000 in average - this is balanced between
sender and receiver. Consumer consumes old messages, new messages are still arriving.
Every message has 5kB.
The DB grows in size during this procedure until it takes the whole disk space and make
the DB unusable for JBM.
Evidence - attached files:
dbstat.csv - status report for all the tables in DB - see RowCount, DataSpace for
JBM_MSG
jbm_msg.csv - content of JBM_MSG table when status report was generated
soaesb1_dbstat.sql - SQL script for generating status report
Caution: MSSQL is required in SOA-P PRD. If this cannot be solved, the requirement should
be removed.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira