Bug #45853 MySQL Cluster auto-discovery table disappears and cannot be dropped anymore
Submitted: 30 Jun 2009 12:41 Modified: 20 Sep 2009 10:06
Reporter: Oli Sennhauser Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S1 (Critical)
Version:mysql-5.1-telco-6.3 OS:Linux
Assigned to: CPU Architecture:Any
Tags: auto-discovery, Drop, MySQL Cluster, nbdb-6.3.24, table
Triage: Triaged: D2 (Serious) / R6 (Needs Assessment) / E6 (Needs Assessment)

[30 Jun 2009 12:41] Oli Sennhauser
Description:
Customer experiences a loss of a table. The table does still exist on the cluster but is not auto-discovered by the SQL node.

Further he cannot drop the table on the cluster.

How to repeat:
Customer did a restart of mysqld, then a SHOW TABLES followed by a SELECT * FROM table.

Then he got:

ERROR 1146 (42S02) : Table '...' doesn't exist.

Then he dropped the database on the SQL node but the database and the tables still exist on the cluster.

Then he tried to drop the table on the cluster:

$ ndb_drop_table ... -d ...
Dropping table ...
723: No such table existed

But also re-creating the table on the SQL node fails:

ERROR 1050 (42S01) at line 10 in file: '...': Table '...' already exists

So he cannot go forward nor backward.

Suggested fix:
Make auto-discovery work correctly OR make dropping table on cluster with force possible.
[20 Aug 2009 10:06] Sveta Smirnova
Thank you for the report.

Please check if this is same as bug #37419 or bug #40854
[20 Sep 2009 23:00] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".