Merge lp:~laurynas-biveinis/percona-server/bug1204871 into lp:percona-server/5.6
Status: | Merged |
---|---|
Approved by: | Laurynas Biveinis |
Approved revision: | no longer in the source branch. |
Merged at revision: | 527 |
Proposed branch: | lp:~laurynas-biveinis/percona-server/bug1204871 |
Merge into: | lp:percona-server/5.6 |
Diff against target: |
87 lines (+35/-2) 5 files modified
Percona-Server/mysql-test/suite/parts/r/partition_bug71270.result (+8/-0) Percona-Server/mysql-test/suite/parts/t/partition_bug71270.test (+21/-0) Percona-Server/sql/ha_partition.cc (+1/-0) Percona-Server/sql/sql_class.h (+1/-1) Percona-Server/sql/sql_insert.cc (+4/-1) |
To merge this branch: | bzr merge lp:~laurynas-biveinis/percona-server/bug1204871 |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Vlad Lesin (community) | g2 | Approve | |
Review via email: mp+200529@code.launchpad.net |
Description of the change
No BT or ST but part of TokuDB QA.
Fix bug 1204871 (abort in ha_partition:
sql/ha_
Assertion `0' failed. (sig6)) /
http://
A debug server build will crash if, while performing a bulk insert to
a partitioned table, one of the partitions will return a failure for
end_bulk_insert handler call.
The crash happens because the failing ha_partition:
call will clear the m_bulk_
failure to the caller. The caller will handle the failure, and at one
point it will call select_
attempt end_bulk_insert again, which will hit DBUG_ASSERT(0) because
m_bulk_
Fix by resetting THD::bulk_
handler calls. Add an error-injecting testcase
parts/partition
http://