Bug #37628 mysqld can crash on node-failure (with replication)
Submitted: 25 Jun 2008 14:47 Modified: 2 Jul 2008 19:39
Reporter: Jonas Oreland
Status: Closed
Category:Server: Cluster Severity:S3 (Non-critical)
Version:>6.2.5 >6.3.2 OS:Any
Assigned to: Jonas Oreland Target Version:

[25 Jun 2008 14:47] Jonas Oreland 
Description:
If ndbd resend algorithm completed more than 1 bucket,
event subscriber could crash with segmentation fault
as it was looping too much in complete_out_of_order_gci
due to incorrect variable scoping

How to repeat:
.

Suggested fix:
.
[25 Jun 2008 14:49] 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/48480

2630 jonas@mysql.com	2008-06-25
      ndb - bug#37628
        Fix variable scoping (remove duplicate variable start_gci)
        in complete_out_of_order_buckets
        also correct a printout in same function
[26 Jun 2008 12:46] 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/48560
[1 Jul 2008 17:03] 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/48816
[2 Jul 2008 9:00] Jonas Oreland 
pushed to 6.3.16 and 6.2.16
[2 Jul 2008 19:39] Jon Stephens 
Documented in the NDB 6.2.16 and 6.3.16 changelogs as follows:

      Failure of a data node could sometimes cause mysqld to crash.
[13 Dec 2008 0:26] Bugs System 
Pushed into 6.0.6-alpha  (revid:jonas@mysql.com-20080625124913-ek8jhyn3drl6067i) (version
source revid:jonas@mysql.com-20080812185642-1nevjb94zj621dqx) (pib:5)