Bug #12594 Error while reading the REDO log
Submitted: 16 Aug 2005 9:50 Modified: 21 Sep 2005 13:35
Reporter: Johan Andersson Email Updates:
Status: To be fixed later
Category:Server: Cluster Severity:S2 (Serious)
Version:5.1 OS:Linux (RH9)
Assigned to: Jonas Oreland Target Version:

[16 Aug 2005 9:50] Johan Andersson 
Description:
After a reboot of the computers running a  2 node ndbd cluster, it was not possible to
restart them, because one ndbd node crashed with this symptom:

Date/Time: Friday 12 August 2005 - 19:42:57
Type of error: error
Message: Error while reading the REDO log
Fault ID: 2310
Problem data: Error while reading REDO log.
D=10, F=0 Mb=0 FP=5 W1=34 W2=5359
Object of reference: DBLQH (Line: 14988) 0x00000008
ProgramName: /usr/mysql/libexec/ndbd
ProcessID: 951
TraceFile: /zxxxx/ndb_4_trace.log.8
Version 5.1.0 

The problem is that the ndbd node does not recover up to latest "good" GCP, and then
recover the rest (if any) from the node in the same node group with the healthy redo
log.

How to repeat:
The file system is located on the associated issue and by using the supplied config.ini
and the filesystem it is possible to reproduce.

Suggested fix:
-
[8 Sep 2005 9:53] Bugs System 
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/internals/29489
[12 Sep 2005 11:01] Bugs System 
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/internals/29646
[21 Sep 2005 13:35] Jonas Oreland 
There is a patch that makes future problems like this eaiser to find.