Bug #75916 Test case for Bug34702 fails
Submitted: 16 Feb 2015 11:10 Modified: 11 May 2015 20:27
Reporter: Mikael Ronström Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S2 (Serious)
Version:7.4.4 OS:Any
Assigned to: CPU Architecture:Any

[16 Feb 2015 11:10] Mikael Ronström 
Description:
When attempting an initial node restart that fails followed by an immediate new restart of the same node
we could get blocked forever in the second restart if a non-master node is still not completed with
invalidating node LCPs when the master is done. This can happen in test case developed for Bug#34702.

How to repeat:
Run testNodeRestart -n Bug34702 T1

Suggested fix:
Fix reset of state in master and ensure that non-masters can accept receiving multiple START_INFOREQ for same restart.
[11 May 2015 20:27] Jon Stephens 
Documented fix in the NDB 7.4.5 changelog, as follows:

    In the event of a node failure during an initial node restart
    followed by another node start, the restart of the the affected
    node could hang with a START_INFOREQ that occurred while
    invalidation of local checkpoints was still ongoing. 

Closed.