Recovery log stops processing and tries to enable database if there is an
unparsable statement in the recovery log
------------------------------------------------------------------------------------------------------------------
Key: SEQUOIA-969
URL: https://forge.continuent.org/jira/browse/SEQUOIA-969
Project: Sequoia
Type: Bug
Components: Recovery Log
Versions: Sequoia 2.10.9
Reporter: Robert Hodges
Priority: Critical
Fix For: Sequoia 2.9
The recovery log incorrectly signals that recovery has completed when it hits a
bad SQL command. This problem can be duplicated as follows:
1.) Initialize a cluster and disable one of the backends.
2.) Enter a command containinng garbage SQL such as "seleselect from *". This
command fails but is recorded in the recovery log.
3.) Enter additional updates that succeed, for example loading data into a
table.
4.) Enable the backend again. The backend will be enabled but in fact the
recovery log will only be read up through the bad SQL statement. There is no
error message indicating that there was problem.
After this error the backends are out of sync. The usual visible symptom is
that you start to lose backends because updates give different results, causing
backends to go off-line.
--
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
