Bug #78780 Crash on erroneous ndbassert
Submitted: 9 Oct 2015 11:17 Modified: 12 Oct 2015 15:10
Reporter: Mikael Ronström Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:7.4.9 OS:Any
Assigned to: CPU Architecture:Any

[9 Oct 2015 11:17] Mikael Ronström
Description:
NDB crashes in invalidateLcpInfoAfterSr due to lcpActive being set but not
activeStatus == Active.

This setting is a normal setting if the crash occurs while a node is starting
up, it has been away for a while, it is now participating in LCP to get up
and then the cluster crashes.

How to repeat:
testNodeRestart -n Bug25984 T1 occasionally gets this bug.

Suggested fix:
Remove ndbassert and add comment on why it can have more than one state.
[12 Oct 2015 15:10] Jon Stephens
Dcoumented fix in the NDB 7.2.23, 7.3.12, 7.4.9, and 7.5.0 changelogs as shown here:

  NDB failed during a node restart due to the LCP status being set but not
  as active, even though it could have other states under such conditions.
    
Closed.