Bug #77502 ndb_join_pushdown tests unstable execute_count
Submitted: 26 Jun 2015 9:44 Modified: 12 Oct 2015 12:59
Reporter: Magnus Blåudd Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:7.5.0 OS:Any
Assigned to: CPU Architecture:Any

[26 Jun 2015 9:44] Magnus Blåudd
Description:
The ndb_join_pushdown_* tests are unstable due to the execute_count variable being captured for several hundred pushed join queries in the same test. This is not stable since the execute_count may vary due to scheduling and batch size in the data nodes.

How to repeat:
http://ndbmaster.se.oracle.com/~club/showpush.pl?dir=git_mysql-5.7-cluster-7.5&order=42

Suggested fix:
Remove the ndb_execute_count from ndb_join_pushdown.inc. Shouldd we use that counter it can only be done for individual queries and in such case it should probablya > check.
[12 Oct 2015 12:59] Jon Stephens
Fixed in NDB 7.5.0. 

Appears to relate to testing only, and not to affect any releases in any case. 

Closed.