Bug #47215 Incorrectly computed pool-size can cause node crash during create table
Submitted: 9 Sep 2009 10:12 Modified: 9 Sep 2009 16:44
Reporter: Jonas Oreland Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:mysql-5.1-telco-7.0 OS:Any
Assigned to: Jonas Oreland CPU Architecture:Any

[9 Sep 2009 10:12] Jonas Oreland
Description:
While testing fix for bug#44908 i noticed that

1) node could crash when LQH sending CREATE_TAB_REF
   (this is a regression in 7.0)

2) the size of the tab-descriptor-pool used in TUP is super incorrect in 7.0
   (and somewhat incorrect in 6.3)

How to repeat:
Create 20320 tables with 128 attributes each

Suggested fix:
1) Even if calculation is incorrect, don't crash
2) Correct calculation
[9 Sep 2009 11:04] Bugs System
Pushed into 5.1.35-ndb-7.1.0 (revid:jonas@mysql.com-20090909110427-4vsmv05b2i6svff9) (version source revid:jonas@mysql.com-20090909110427-4vsmv05b2i6svff9) (merge vers: 5.1.35-ndb-7.1.0) (pib:11)
[9 Sep 2009 11:06] Bugs System
Pushed into 5.1.37-ndb-7.0.8 (revid:jonas@mysql.com-20090909110150-kdvot0410na89hk5) (version source revid:jonas@mysql.com-20090909105810-v3bws3srdbkdi3rx) (merge vers: 5.1.37-ndb-7.0.8) (pib:11)
[9 Sep 2009 13:12] 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/82824

2996 Jonas Oreland	2009-09-09
      ndb - bug#47215
        1) Fix so that CREATE_TAB_REF doesnt cause node-crash
        2) Fix calculation of TUP_TABLE_DESC by moving to TUP
           that knows how compute it
[9 Sep 2009 16:44] Jon Stephens
Documented bugfix in the NDB-7.0.8 changelog as follows:

        The size of the table descriptor pool used in the DBTUP kernel
        block was incorrect. This could lead to a data node crash when
        an LQH sent a CREATE_TAB_REF signal.

Closed.