Bug #68491 NDB Node fails with 2301 error. Also error 290 for specific table

• Submitted: 26 Feb 2013 5:49
• Modified: 26 Apr 2013 14:13
• Reporter: Ashwin Shah
• Status: Verified
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S2 (Serious)
• Version: 7.2.10
• OS: Linux (CentOS 6.2)
• CPU Architecture: Any

[26 Feb 2013 5:49] Ashwin Shah

Description:
NDB nodes crash with 2301 error. SQL API reports 290 error with a specific table.

This was a working setup with ndb 7.1.18. There were no issues being reported.

We wiped the entire cluster after taking a dump of the DBs with mysqldump, uninstalled the 7.1.18 binaries and installed the 7.2.10 binaries. started the cluster successfully. We imported the mysqldump files.

Now we get 290 errors on a specific table. Additionally node 4 crashed with a 2301 error. When we tried to bring up node 4 again it caused node 3 to crash as well again with a 2301 error. node 4 failed after that.

We have 128GB RAM on the data nodes

How to repeat:
Install 7.2.10 on CentOS 6.2

[26 Feb 2013 12:24] MySQL Verification Team

Hello Ashwin,

Thank you for the report.

This issue has been reported/tracked internally in Bug 16187976: NDBD NODE FAILS TO START WITH ILLEGAL SIGNAL RECEIVED (GSN 121 NOT ADDED)

Regards,
Umesh

[26 Feb 2013 13:24] Ashwin Shah

Thanks Umesh! 

Is there any timeline for a bugfix? Should I downgrade? If yes which version do you recommend?

[26 Apr 2013 13:45] Santo Leto

Issue has been fixed. You can upgrade to v. 7.2.12