Bug #54859 >2 replica: Hanging if 1 node can not be restored using REDO log during SR
Submitted: 28 Jun 2010 9:12 Modified: 28 Jun 2010 9:22
Reporter: Jonas Oreland Email Updates:
Status: Closed Impact on me:
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version: OS:Any
Assigned to: Jonas Oreland CPU Architecture:Any

[28 Jun 2010 9:12] Jonas Oreland
During a system restart, nodes can "use" redo-logs from other nodes
in same node group.

If 1 node does not have sufficient REDO to start, it will be started
by copying from other node in node-group.

When deciding which REDO log to run, an incorrect variable is checked
if node will start using REDO or by copying.

This could if having 3 or 4 replicas lead to a hanging system restart.

Found while reading the code

How to repeat:
Read the code

Suggested fix:
Use correct variable
[28 Jun 2010 9:15] 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:


3229 Jonas Oreland	2010-06-28
      ndb - bug#54859 - use correct state-variable when computing redo intervals
[28 Jun 2010 9:22] Jonas Oreland
pushed to 6.3.36, 7.0.17 and 7.1.6
hmm...probably no need to document, closing.