Complete sequoia deadlock when trying to perfom a backup via the sequoia
console. ... still waiting for 0 pending transactions to complete ...
-----------------------------------------------------------------------------------------------------------------------------------------------
Key: SEQUOIA-864
URL: https://forge.continuent.org/jira/browse/SEQUOIA-864
Project: Sequoia
Type: Bug
Components: Backup System
Versions: Sequoia 2.10.2
Environment: CentOS 4.4 Kernel 2.6.9-42.0.3.ELsmp x86_64 x86_64 x86_64
GNU/Linux,
MySQL-server-standard-5.0.24a-0.rhel4 (rpm from MySQL),
Reporter: Stefan Schank
We have a serious problem when trying to perform a backup via the sequoia
console Application. Our configuration is a RAIDb-1 with two separate single
MySQL instances.
We have had the situation, that a malformed statement was issued from an SQL
Query Tool which caused the following exception:
java.sql.SQLException: write request xxxxx failed: Backend xxxxx -
BackendWorkerThread for backend 'node1' with RAIDb level:1 failed (Unknown
column 'name' in 'field list')
Some hours later we tried to make a backup with the MyQSLBackuper and the
full_cluster Logfile was filled with the following entries:
2006-12-01 20:12:54,387 WARN sequoia.controller.scheduler Timeout in begin,
still waiting for 0 pending transactions to complete
2006-12-01 20:12:54,387 WARN controller.virtualdatabase.xxxxx Begin failed
(Timeout in begin, still waiting for 0 pending transactions to complete)
2006-12-01 20:12:54,387 WARN virtualdatabase.VirtualDatabaseWorkerThread.xxxxx
Error during command execution (Timeout in begin, still waiting for 0 pending
transaction
s to complete)
The console app began to hang and the databases couldn't be contacted any
further through sequoia. We didn't manage it to disable the backend nor to stop
the backup command. It also was not possible to kill the pending transaction,
so that the backup could proceed. After all we had to kill the sequoia
processes and had to recover the complete configuration including the databases
on both nodes.
We need a reliable solution even in the case, that malformed transactions or
statements are issued to sequoia. Any suggestions?
Configuraion files et. al. can be provided if necessary.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://forge.continuent.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
