Bug #48531 Total length of table names >= 1024 bytes kills ndbd with backup

• Submitted: 4 Nov 2009 14:35
• Modified: 5 Nov 2009 6:16
• Reporter: Andrew Hutchings
• Status: Closed
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S1 (Critical)
• Version: 6.3.28
• OS: Any
• Assigned to: Jonas Oreland
• CPU Architecture: Any

[4 Nov 2009 14:35] Andrew Hutchings

Description:
When the total length of all cluster table names >= 1024 bytes the cluster will crash.

Time: Wednesday 4 November 2009 - 15:26:45
Status: Temporary error, restart node
Message: Internal program error (failed ndbrequire) (Internal error, programming error or missing error message, please report a bug)
Error: 2341
Error data: SimulatedBlock.cpp
Error object: DBDICT (Line: 1424) 0x0000000a
Program: /data1/mysql/ndb-6.3.28/libexec/ndbd
Pid: 19023
Trace: /data2/csc/41989/data/ndb_2_trace.log.1
Version: mysql-5.1.39 ndb-6.3.28-GA
***EOM***

How to repeat:
Insert SQL (to be attached shortly) into a single node cluster
ndb_mgm> START BACKUP

[4 Nov 2009 14:37] Andrew Hutchings

SQL to reproduce problem

Attachment: bug-48531.sql.gz (application/x-gzip, text), 1.05 KiB.

[4 Nov 2009 15:10] 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/commits/89322

3146 Jonas Oreland	2009-11-04
      ndb - bug#48531

[4 Nov 2009 15:16] 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/commits/89325

3177 Jonas Oreland	2009-11-04
      ndb - bug#48531 (7.0.9)

[4 Nov 2009 16:50] Bugs System

Pushed into 5.1.39-ndb-7.0.10 (revid:[e-mail]-20091104164902-cs02vkzkavxw5ywo) (version source revid:[e-mail]-20091104164902-cs02vkzkavxw5ywo) (merge vers: 5.1.39-ndb-7.0.10) (pib:13)

[4 Nov 2009 20:04] Jonas Oreland

pushed to 6.3.28a and 7.0.9a (new releases)

[4 Nov 2009 20:10] Jonas Oreland

Bug is not a regression per se, since it has been around since 2008-04,
however, due to changes in internal timing, the problem is now very very likely 
to occur if $subject is true, whereas earlier it was very very unlikely.

Therefor one can argue that it is a regression.

[4 Nov 2009 20:58] Jon Stephens

Documented bugfix in the NDB-6.3.28a and 7.0.9a changelogs as follows:

        When the combined length of all names of tables using the NDB
        storage engine was greater than or equal to 1024 bytes, issuing
        the START BACKUP command in the ndb_mgm client caused the
        cluster to crash.

Updated the 6.3.28 and 7.0.9 changelogs as follows:
  ·Removed download links for the releases
  ·Added notices that releases were pulled due to this bug and users of these releases should upgrade to 6.3.28a/7.0.9a ASAP.

Closed.

Docs commit is here: http://lists.mysql.com/commits/89361

[5 Nov 2009 5:55] Bugs System

Pushed into 5.1.39-ndb-7.1.0 (revid:[e-mail]-20091105055409-9f82ewita0a4ptlh) (version source revid:[e-mail]-20091105055409-9f82ewita0a4ptlh) (merge vers: 5.1.39-ndb-7.1.0) (pib:13)

[5 Nov 2009 6:16] Jon Stephens

Thank you for your bug report. This issue has already been fixed in the latest released version of that product, which you can download at

  http://www.mysql.com/downloads/