[WFTO] Weekly Falcon Test Overview 2009-02-06

[Hakan Kuecuekyilmaz]

Dear Falcon fellows,


Yes! This was again a bug fixing week. Vladislav Vaintroub fixed a lot
of recovery related bugs and Kevin Lewis and Chris Powers took care
about the new scavenger and backlogging.

The full test overview is located at:
https://intranet.mysql.com/~hakank/weekly_falcon_test_overview.php

For our external guests, I have a blog entry here:
http://esslingen.homeunix.net/~hakan/blog/2009/02/06/wfto-weekly-falcon-test-overview-2009-02-06/

Good news of the week:
   - Vlad fixed a lot recovery bugs,
   - Chris and Kevin polished our new scavenger code.

Concerns of the week.
   - This week we saw a quite high new bug income rate (again).
   - object.objects_falcon failed to run (analyzing)
   - Windows debug run has strange results (analyzing)
   - Regression:
    falcon.falcon_bug_22169-big
    mysqltest: At line 49: query 'INSERT INTO t2 (id, grp, id_rev)
    SELECT id, grp, id_rev FROM t1' failed: 1296: Got error 305
    'record memory is exhausted' from Falcon


News for this week:
* Christopher Powers fixed
  o Bug#36442 Falcon: crash during optimize table
  o Bug#40937 ASSERT(transaction->thawedRecords > 0) fails
  o Bug#42424 Serious performance degradation after new scavenger
  o Bug#42505 Falcon: Record backlogging not enabled by the Scavenger
  o Bug#42510 Falcon: Default memory parameters must be changed

* Vladislav Vaintroub fixed
  o Bug#40946 assertion in PageInventoryPage::allocPage in recovery
  o Bug#42542 falcon_bug_28095 crashes in StorageHandler::dropTempTables
  o Bug#41227 Falcon: potential corruption of RecordLocatorPage on recovery
  o Bug#42227 Test falcon.falcon_bug_28095 fails sporadically on pushbuild


Changes in Weekly Falcon Test Overview.
   - None.

This week's results:
 walldorf debug:   Failed 3/242, 98.76% were successful.
 walldorf opt:     Failed 3/241, 98.76% were successful.
 ubuntu32x2 debug: Failed 1/230, 99.57% were successful.
 ubuntu32x2 opt:   Failed 1/229, 99.56% were successful.
 win64 debug:      Failed 6/6, 0.00% were successful.
 win64 opt:        Failed 1/229, 99.56% were successful.
 esslingen debug:  Failed 3/230, 98.70% were successful.
 esslingen opt:    Failed 1/229, 99.56% were successful.

Test observations:
   - New tests: 4 new tests.
   - falcon_bug_22150 regressed on Mac/PPC with
   # Pull the results of the preceeding call p1()
   SELECT count(*) FROM t1;
   count(*)
  -0
  +29
   # Disconnect conn1
   SET @@autocommit = DEFAULT;
   DROP TABLE t1;

Revision used:
revno: 3003
revision-id: hky@stripped
parent: hky@stripped
committer: Hakan Kuecuekyilmaz 
branch nick: mysql-6.0-falcon-to-merge
timestamp: Thu 2009-02-05 10:37:34 +0100

Notes:
   - The test suite was run like:
   ./mysql-test-run.pl --enable-disabled ---force --skip-ndb 
--suite=falcon,falcon_team
   walldorf additionally runs
   ./mysql-test-run.pl --enable-disabled ---force --skip-ndb
--suite=falcon,falcon_team \
   --big-test --suite-timeout=6360 --testcase-timeout=30 --do-test=.*-big

Future plans:
  - As new platforms for WFTO I plan Solaris 10/Intel 64-bit,
  and Solaris 9/SPARC (not in particular order).

Notes on the machines used:
   - Real machines have city names.
   - Virtual machines (VMware) have this naming scheme:
   <distribution|operating system><bits>x<CPUs>

   - walldorf: main machine running also VMware. Dell two-way dual-core,
   8GB RAM, 15k SAS RAID 10 storage. RHEL5.
   - esslingen: Mac Cube G4, PPC 450MHz., 1 GB RAM. Mac OS X 10.4.


Best regards,

Hakan

--
Hakan Küçükyılmaz, Senior Software Engineer DBTG/MySQL +49 160
98953296
Sun Microsystems GmbH     Sonnenallee 1, DE-85551 Kirchheim-Heimstetten
Geschaeftsfuehrer:  Thomas Schroeder, Wolfang Engels, Dr. Roland Boemer
Vorsitz d. Aufs.rat.: Martin Haering   HRB MUC 161028     49.011, 8.376