mysql crash

201 Version Buffer Test: Failed (check test201/diff.txt)

> ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/usr/local/MariaDB/Columnstore/mysql/lib/mysql/mysql.sock' (2 "No such file or directory")

Latest nightly run log is better (no crash), though the test is still failing

201 Version Buffer Test: Failed (check test201/diff.txt)

hill Here is the analysis of the diff log

cat test201/diff.txt
97c97
< ERROR 122 (HY000) at line 23: CAL0002: Update Failed: IDB-2008: The version buffer overflowed. Increase VersionBufferFileSize or limit the rows to be processed.
—
> ERROR 1815 (HY000) at line 23: Internal error: CAL0002: Update Failed: IDB-2008: The version buffer overflowed. Increase VersionBufferFileSize or limit the rows to be processed.

The mis-match is due to ERROR 122 vs ERROR 1815 - is not true failure - It is due the fact that MariaDB ColumnStore that is based on MariaDB Server uses error number 1815, where as the reference log is showing old error number of 122 - Other wise the text of the error is essentially the same.

hill Please update line 97 in test201/diff.ref.log with "ERROR 1815 (HY000) at line 23: Internal error: CAL0002: Update Failed: IDB-2008: The version buffer overflowed. Increase VersionBufferFileSize or limit the rows to be processed."
After this ref log change is made, and if next regression run shows test script passing then we can close this issue.

change made to the ref log

hill test201/diff.ref.log gets copied from one of the following files during test run (depending upon dbroot)

test201/test201.sql.1.ref.log
test201/test201.sql.2.ref.log
test201/test201.sql.4.ref.log
test201/test201.sql.6.ref.log

For all of the 4 files above, please update line number 109 to "ERROR 1815 (HY000) at line 23: Internal error: CAL0002: Update Failed: IDB-2008: The version buffer overflowed. Increase VersionBufferFileSize or limit the rows to be processed."

After this the test script should pass

copied ref logs from old infinidb, but still failing.. This ALSO has the problem with the 0 rows affected BUG..

[root@srvregtest test201]# diff test201.sql.log test201.sql.ref.log
5c5
< Query OK, 0 rows affected (27.30 sec)
—
> Query OK, 25000 rows affected (32.77 sec)
17c17
< 1 row in set (0.45 sec)
—

Ok that is related to MCOL-72 - so linking them now

In the latest run this is passing