linux-traipu team mailing list archive

[Patrick Crews <796729@xxxxxxxxxxxxxxxxxx>]

On 06/22/2011 07:52 AM, Olaf van der Spek wrote:
> I think this one is fixed in trunk.
>
Kind of - it is failing in a different manner and takes far more 
iterations to fail:
./dbqp --suite=pbms --repeat=500
It tends to fail around the 300-ish mark
Of course, this is most likely a different bug, but noting the behavior

20110622-200447  pbms.basic                                 [ fail ]     
1784
20110622-200447  drizzletest: At line 22: query 'insert into 
pbms_metadata (select id, offset, "Name", "Test1" from my_pbms_blob_ids 
ref, t1  where t1.a = 1 and ref.url =  t1.b)' failed: 1030: Got error 1 
from storage engine
20110622-200447
20110622-200447  The result from queries just before the failure was:
20110622-200447  CREATE TABLE t1 (a int primary key, b blob);
20110622-200447  insert into t1 values (1, "Hello World");
20110622-200447  insert into t1 values (2, "Hello PBMS");
20110622-200447  insert into t1 values (3, "Hello BLOB World");
20110622-200447  SELECT COUNT(*) FROM t1;
20110622-200447  COUNT(*)
20110622-200447  3
20110622-200447  select a, Table_name, Blob_size from pbms_reference 
ref, t1  where ref.Blob_url =  t1.b order by a;
20110622-200447  a    Table_name    Blob_size
20110622-200447  1    t1    11
20110622-200447  2    t1    10
20110622-200447  3    t1    16
20110622-200447  select a, MD5_Checksum from pbms_repository rep, 
pbms_reference ref, t1 where t1.b = ref.Blob_url and ref.Repository_id = 
rep.Repository_id and ref.Repo_blob_offset = rep.Repo_blob_offset order 
by a;
20110622-200447  a    MD5_Checksum
20110622-200447  1    B10A8DB164E0754105B7A99BE72E3FE5
20110622-200447  2    90D9BF084B48FBC92710E6807B8875B0
20110622-200447  3    6B230DE148EE7EA4A4454090784119E6
20110622-200447  CREATE Temporary TABLE my_pbms_blob_ids (id int, offset 
int, url varchar(120));
20110622-200447  insert into my_pbms_blob_ids (select Repository_id, 
Repo_blob_offset, Blob_url from pbms_reference);
20110622-200447  insert into pbms_metadata (select id, offset, "Name", 
"Test1" from my_pbms_blob_ids ref, t1  where t1.a = 1 and ref.url =  t1.b);

cat workdir/bot0/s0/var/log/s0.err
InnoDB: Doublewrite buffer not found: creating new
InnoDB: Doublewrite buffer created
InnoDB: 127 rollback segment(s) active.
InnoDB: Creating foreign key constraint system tables
InnoDB: Foreign key constraint system tables created
110622 20:01:35 [Note] PrimeBase Media Stream (PBMS) Daemon 0.5.14-beta 
loaded...
110622 20:01:35 [Note] Barry Leslie, PrimeBase Technologies GmbH, 
http://www.primebase.org
110622 20:01:35 [Note] startup: Media Stream Daemon  listening on port 9309
110622 20:04:48  InnoDB: Starting shutdown...

-- 
You received this bug notification because you are a member of UBUNTU -
AL - BR, which is subscribed to Drizzle.
https://bugs.launchpad.net/bugs/796729

Title:
  pbms.basic test failing randomly when run in rapid succession

Status in A Lightweight SQL Database for Cloud Infrastructure and Web Applications:
  Confirmed

Bug description:
  When you run the pbms.basic test in rapid succession (at least on my Natty laptop), we are seeing random failures - looks like extra (not cleaned up?) rows in the result set
  Found this when attempting to repeat that random engine error failure we are seeing in Jenkins slaves.

  This only occurs when we're using shared mem (linking the test-runner
  workdir to /dev/shm/...).  Running with the --no-shm option is fine.

  To repeat:
  From the drizzle/tests directory:
  ./dbqp --suite=pbms --repeat=20

  When failures occur, the output looks like this:
  20110613-120255  ===============================================================
  20110613-120255  TEST NAME                                  [ RESULT ] TIME (ms)
  20110613-120255  ===============================================================
  20110613-120255  pbms.basic                                 [ pass ]     1088
  20110613-120255  pbms.basic                                 [ pass ]       67
  20110613-120255  pbms.basic                                 [ pass ]       90
  20110613-120255  pbms.basic                                 [ pass ]       86
  20110613-120255  pbms.basic                                 [ pass ]       78
  20110613-120256  pbms.basic                                 [ pass ]      573
  20110613-120256  pbms.basic                                 [ pass ]       69
  20110613-120257  pbms.basic                                 [ fail ]      598
  20110613-120257  --- drizzle/plugin/pbms/tests/r/basic.result	2011-06-10 14:42:15.005786000 -0400
  20110613-120257  +++ drizzle/plugin/pbms/tests/r/basic.reject	2011-06-13 12:02:57.131554245 -0400
  20110613-120257  @@ -31,6 +31,7 @@
  20110613-120257  select a, name, value from pbms_metadata meta, pbms_reference ref, t1 where t1.b = ref.Blob_url and ref.Repository_id = meta.Repository_id and ref.Repo_blob_offset = meta.Repo_blob_offset order by a, name;
  20110613-120257  a	name	value
  20110613-120257  1	Name	Test1
  20110613-120257  +1	Name	Test1
  20110613-120257  2	Name	Test2
  20110613-120257  2	Name2	Test2.2
  20110613-120257  3	Name	Test3
  20110613-120257
  20110613-120257  drizzletest: Result content mismatch
  20110613-120257
  20110613-120257
  20110613-120257 ERROR Failed test.  Use --force to execute beyond the first test failure
  20110613-120257  ===============================================================
  20110613-120257 INFO Test execution complete in 3 seconds
  20110613-120257 INFO Summary report:
  20110613-120257 INFO Executed 8/1 test cases, 800.00 percent
  20110613-120257 INFO STATUS: FAIL, 1/8 test cases, 12.50 percent executed
  20110613-120257 INFO STATUS: PASS, 7/8 test cases, 87.50 percent executed
  20110613-120257 INFO FAIL tests: pbms.basic
  20110613-120257 INFO Spent 0 / 3 seconds on: TEST(s)

To manage notifications about this bug go to:
https://bugs.launchpad.net/drizzle/+bug/796729/+subscriptions