Bug 1096787 - Tracker: Failing mariadb-10.0.x testsuite for different architectures
Summary: Tracker: Failing mariadb-10.0.x testsuite for different architectures
Keywords:
Status: CLOSED EOL
Alias: None
Product: Fedora
Classification: Fedora
Component: mariadb
Version: 23
Hardware: All
OS: Linux
unspecified
low
Target Milestone: ---
Assignee: Michal Schorm
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2014-05-12 12:30 UTC by Honza Horak
Modified: 2016-12-20 12:48 UTC (History)
5 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2016-12-20 12:48:02 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
log of the testsuite run (479.70 KB, text/x-log)
2014-05-12 12:30 UTC, Honza Horak
no flags Details
mysqlslap-fix-deadlock.patch (495 bytes, patch)
2014-08-26 08:21 UTC, Michal Toman
no flags Details | Diff

Description Honza Horak 2014-05-12 12:30:58 UTC
Created attachment 894674 [details]
log of the testsuite run

Description of problem:
Running the testsuite of mariadb-10.0.x on ppc64 results in many failing tests.

Version-Release number of selected component (if applicable):
mariadb-10.0.10-1.fc21

Check of testcase failed for:
rpl.rpl_extra_col_master_myisam
rpl.rpl_extra_col_master_innodb
rpl.rpl_deadlock_innodb

Failing test(s):
connect.alter
connect.bin
connect.dbf
connect.index
multi_source.skip_counter
rpl.rpl_auto_increment
rpl.rpl_mixed_binlog_max_cache_size
rpl.rpl_slave_skip
rpl.rpl_gtid_until
rpl.rpl_skip_replication
rpl.rpl_stm_max_relay_size
rpl.rpl_row_loaddata_concurrent
rpl.rpl_extra_col_slave_innodb
rpl.rpl_row_binlog_max_cache_size
rpl.rpl_row_tabledefs_3innodb
rpl.rpl_extra_col_slave_myisam
rpl.rpl_typeconv
rpl.rpl_row_max_relay_size
rpl.rpl_row_tabledefs_2myisam
rpl.rpl_stm_loaddata_concurrent
rpl.rpl_loaddata
rpl.rpl_stm_binlog_max_cache_size
rpl.rpl_filter_tables_not_exist
rpl.rpl_filter_tables_dynamic
rpl.rpl_stm_000001
rpl.rpl_gtid_basic
rpl.rpl_gtid_master_promote
rpl.rpl_gtid_stop_start
rpl.rpl_rotate_logs
main.statistics
main.gis-precise
sys_vars.max_relay_log_size_basic
sys_vars.sql_slave_skip_counter_basic

Comment 2 Dan Horák 2014-05-26 08:54:40 UTC
I think the same problems exist on s390(x)

from build.log
...
--------------------------------------------------------------------------
The servers were restarted 261 times
Spent 495.459 of 1925 seconds executing testcases
Check of testcase failed for: rpl.rpl_deadlock_innodb
Too many failed: Failed 10/119 tests, 91.60% were successful.
Failing test(s): multi_source.skip_counter rpl.rpl_auto_increment rpl.rpl_mixed_binlog_max_cache_size rpl.rpl_slave_skip rpl.rpl_gtid_until rpl.rpl_skip_replication
The log files in var/log may give you some hint of what went wrong.
If you want to report this error, please read first the documentation
at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html
412 tests were skipped, 143 by the test itself.
mysql-test-run: *** ERROR: there were failing test cases
error: Bad exit status from /var/tmp/rpm-tmp.ruicdB (%check)
RPM build errors:

Comment 3 Dan Horák 2014-05-29 12:22:09 UTC
with http://pkgs.fedoraproject.org/cgit/mariadb.git/tree/rh-skipped-tests-ppc64.list I was able to build mariadb-10.0.11 on s390(x) and ppc64

Comment 4 Michel Normand 2014-06-27 11:51:52 UTC
mariadb 10.0.1-5 failing on ppc64 as reported in http://ppc.koji.fedoraproject.org/koji/buildinfo?buildID=238271

I need to add new tests in skipped list files on top of release 10.0.12-2 as suggested below to make build ppc64/ppc64le to complete in scratch build
http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=1905057
===
diff --git a/rh-skipped-tests-ppc-s390.list b/rh-skipped-tests-ppc-s390.list
index 8c15a79..4509693 100644
--- a/rh-skipped-tests-ppc-s390.list
+++ b/rh-skipped-tests-ppc-s390.list
@@ -3,6 +3,7 @@ connect.bin                                  : rhbz#1096787
 connect.dbf                                  : rhbz#1096787
 connect.index                                : rhbz#1096787
 main.gis-precise                             : rhbz#1096787
+main.openssl_1                               : rhbz#1096787
 main.statistics                              : rhbz#1096787
 multi_source.skip_counter                    : rhbz#1096787
 rpl.rpl_auto_increment                       : rhbz#1096787
diff --git a/rh-skipped-tests-ppc64le.list b/rh-skipped-tests-ppc64le.list
index 716f7b3..1fc4a2b 100644
--- a/rh-skipped-tests-ppc64le.list
+++ b/rh-skipped-tests-ppc64le.list
@@ -1,3 +1,5 @@
 main.gis-precise                             : rhbz#1096787
 main.mysqlslap                               : rhbz#1096787
+main.openssl_1                               : rhbz#1096787
 rpl.rpl_insert                               : rhbz#1096787
+vcol.vcol_supported_sql_funcs_myisam         : rhbz#1096787
===

Comment 5 Michel Normand 2014-07-23 14:39:41 UTC
Could ignore the previous patches suggested in comment #4
with new version mariadb-10.0.12-4.fc21 where the ppc64le archi build is failing
for 3 tests as reported in Task Id http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=1960557
=== extract:
...                
Spent 2658.015 of 8880 seconds executing testcases                                                   
Completed: Failed 3/2890 tests, 99.90% were successful.                                              
Failing test(s): rpl.rpl_insert_delayed main.openssl_1 vcol.vcol_supported_sql_funcs_myisam          
===  

Able to bypass related tests on ppc64le with following git patch as reported by scratch build Id http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=1960681
===
$git diff        
diff --git a/rh-skipped-tests-ppc64le.list b/rh-skipped-tests-ppc64le.list                           
index 716f7b3..9cf211b 100644
--- a/rh-skipped-tests-ppc64le.list
+++ b/rh-skipped-tests-ppc64le.list
@@ -1,3 +1,6 @@
 main.gis-precise                             : rhbz#1096787
 main.mysqlslap                               : rhbz#1096787
+main.openssl_1                               : rhbz#1096787
 rpl.rpl_insert                               : rhbz#1096787
+rpl.rpl_insert_delayed                       : rhbz#1096787
+vcol.vcol_supported_sql_funcs_myisam         : rhbz#1096787
===

Comment 6 Michel Normand 2014-07-23 14:45:13 UTC
Oups comment #5 was sent too early :( sorry the Task Id 1960681 is still running.

Comment 7 Honza Horak 2014-08-20 07:15:05 UTC
With new version 10.0.13 two tests newly fail on armv7hl:
innodb.innodb_simulate_comp_failures_small
main.key_cache
Disabling for now, since there is not enough time to debug.

Comment 8 Michal Toman 2014-08-26 08:20:53 UTC
(In reply to Michel Normand from comment #5)
> +rpl.rpl_insert_delayed                       : rhbz#1096787

The failure of rpl_insert_delayed is caused by mysqlslap deadlocking. The issue is fixed in upstream mariadb 10.0.13 - https://mariadb.atlassian.net/browse/MDEV-6489. I can confirm it passes with mariadb-10.0.13-4.fc22. We may need to backport the patch for F21 - attaching.

openssl_1 and vcol_supported_sql_funcs_myisam are still failing.

Comment 9 Michal Toman 2014-08-26 08:21:32 UTC
Created attachment 930782 [details]
mysqlslap-fix-deadlock.patch

Comment 10 Dan Horák 2014-08-26 08:28:33 UTC
I think the openssl_1 started failing in mariadb-10.0.12-3.fc21, mariadb-10.0.12-2.fc21 was OK (in both ppc and s390 koji), so either the test used to be disabled, or some other change made it to fail

Comment 11 Honza Horak 2014-08-26 10:38:34 UTC
(In reply to Dan Horák from comment #10)
> I think the openssl_1 started failing in mariadb-10.0.12-3.fc21,
> mariadb-10.0.12-2.fc21 was OK (in both ppc and s390 koji), so either the
> test used to be disabled, or some other change made it to fail

This test might fail because of changes in openssl:
* Mon Jun 30 2014 Tomáš Mráz <tmraz> 1.0.1h-5 - disable SSLv2 and SSLv3 protocols by default (can be enabled via appropriate SSL_CTX_clear_options() call)
http://koji.fedoraproject.org/koji/buildinfo?buildID=540893

(In reply to Michal Toman from comment #9)
> Created attachment 930782 [details]
> mysqlslap-fix-deadlock.patch

Great, thanks! I'm planning to sync rawhide and F21 after alpha is out, so it should get to F21 in few weeks.

Also, I've just find out that some fixes for main.gis-precise test are already part of 10.0.13 -- only test is fixed, no changes necessary in the code; plus some other tests have been fixed to work properly on power8, which could cover all ppc arches.

We may re-run the suite on all arches without disabling the tests (no --skip-test-list=) and see which are already fixed (not sure if I manage to do it today and I'll be out for the rest of the week, so if nobody else does it, I'll try that the next week hopefully).

Comment 12 Michal Toman 2014-08-26 12:02:22 UTC
(In reply to Honza Horak from comment #11)
> (In reply to Dan Horák from comment #10)
> > I think the openssl_1 started failing in mariadb-10.0.12-3.fc21,
> > mariadb-10.0.12-2.fc21 was OK (in both ppc and s390 koji), so either the
> > test used to be disabled, or some other change made it to fail
> 
> This test might fail because of changes in openssl:
> * Mon Jun 30 2014 Tomáš Mráz <tmraz> 1.0.1h-5 - disable SSLv2 and
> SSLv3 protocols by default (can be enabled via appropriate
> SSL_CTX_clear_options() call)
> http://koji.fedoraproject.org/koji/buildinfo?buildID=540893

This is indeed true, the test passes with openssl-1.0.1h-2.fc21 and openssl-1.0.1h-3.fc21 (the -3 is not build for power) and does not with openssl-1.0.1h-4.fc21. Tested on ppc and s390.

Comment 13 Michal Toman 2014-08-27 12:34:55 UTC
So I found the openssl problem. Long story short, openssl unintentionally disabled the RC4 cipher which is used in mariadb's testsuite - #1134379.

crypto-policies-20140827-1.git4e06f1d.fc21* should fix the problem.

* http://koji.fedoraproject.org/koji/taskinfo?taskID=7466056

Comment 15 Jaroslav Reznik 2015-03-03 17:01:19 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 22 development cycle.
Changing version to '22'.

More information and reason for this action is here:
https://fedoraproject.org/wiki/Fedora_Program_Management/HouseKeeping/Fedora22

Comment 16 Honza Horak 2015-04-10 10:11:21 UTC
New bug reports connected to failing tests:
https://mariadb.atlassian.net/browse/MDEV-7973
http://bugs.mysql.com/bug.php?id=76654

Comment 17 Jan Kurik 2015-07-15 14:40:42 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 23 development cycle.
Changing version to '23'.

(As we did not run this process for some time, it could affect also pre-Fedora 23 development
cycle bugs. We are very sorry. It will help us with cleanup during Fedora 23 End Of Life. Thank you.)

More information and reason for this action is here:
https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora23

Comment 18 Fedora Admin XMLRPC Client 2015-08-11 10:03:37 UTC
This package has changed ownership in the Fedora Package Database.  Reassigning to the new owner of this component.

Comment 19 Honza Horak 2016-06-27 11:00:47 UTC
Since 10.0.24, mariadb started to fail also in main.ssl_cert_verify. So adding to the devil's list.

main.ssl_cert_verify                     [ fail ]
        Test ended at 2016-06-24 15:28:03
CURRENT_TEST: main.ssl_cert_verify
mysqltest: In included file "/builddir/build/BUILD/mariadb-10.0.25/mysql-test/include/wait_until_connected_again.inc": 
included from /builddir/build/BUILD/mariadb-10.0.25/mysql-test/t/ssl_cert_verify.test at line 20:
At line 26: query 'show status' failed with wrong errno 2026: 'SSL connection error: error:14090086:SSL routines:ssl3_get_server_certificate:certificate verify failed', instead of 0...
The result from queries just before the failure was:
#T1: Host name (/CN=localhost/) as OU name in the server certificate, server certificate verification should fail.
 - saving '/builddir/build/BUILD/mariadb-10.0.25/mysql-test/var/log/main.ssl_cert_verify/' to '/builddir/build/BUILD/mariadb-10.0.25/mysql-test/var/log/main.ssl_cert_verify/'

Comment 20 Menanteau Guy 2016-08-25 12:59:53 UTC
Got some more tests failing with mariadb-10.1.16-1 on ppc64 and ppc64le.
https://ppc.koji.fedoraproject.org/koji/packageinfo?packageID=15127
here after is (a non exhaustive I guess) list of fail tests:
encryption.innodb_onlinealter_encryption
encryption.innodb_encryption_filekeys
encryption.innodb_encryption_discard_import
innodb_fts.innodb_fts_transaction
rpl.rpl_innodb_bug30888 
rpl.rpl_mdev6020 
innodb.innodb_mutexes                       
rpl.rpl_gtid_master_promote

encryption.innodb_onlinealter_encryption test fails everytime but other tests passed ok sometimes.
main.failed_auth_unixsocket fails when user name is "test"

Not sure if the problem is a thread problem or a timing problem but
I didn't have fail tests when:
Tests are running on a VM with qemu parameters: "-smp cpus=1,threads=1"

Most of the time tests are successfull when I use command:
"numactl --physcpubind=+1"
ex: numactl --physcpubind=+1 ./mysql-test-run.pl...
but sometime we can have one test fail randomly from the above list.

Comment 21 Fedora Admin XMLRPC Client 2016-11-15 10:32:27 UTC
This package has changed ownership in the Fedora Package Database.  Reassigning to the new owner of this component.

Comment 22 Fedora End Of Life 2016-11-24 11:09:57 UTC
This message is a reminder that Fedora 23 is nearing its end of life.
Approximately 4 (four) weeks from now Fedora will stop maintaining
and issuing updates for Fedora 23. It is Fedora's policy to close all
bug reports from releases that are no longer maintained. At that time
this bug will be closed as EOL if it remains open with a Fedora  'version'
of '23'.

Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version' 
to a later Fedora version.

Thank you for reporting this issue and we are sorry that we were not 
able to fix it before Fedora 23 is end of life. If you would still like 
to see this bug fixed and are able to reproduce it against a later version 
of Fedora, you are encouraged  change the 'version' to a later Fedora 
version prior this bug is closed as described in the policy above.

Although we aim to fix as many bugs as possible during every release's 
lifetime, sometimes those efforts are overtaken by events. Often a 
more recent Fedora release includes newer upstream software that fixes 
bugs or makes them obsolete.

Comment 23 Michal Schorm 2016-11-28 19:55:20 UTC
Issues with tests:
 * sys_vars.wsrep_provider_options_basic
 * main.ssl_8k_key
solved in Rawhide.


Recent list of remaining skipped tests (rh-skipped-tests-base.list) in Rawhide:
 * main.userstat                                   : rhbz#1096787
 * main.multi_update                               : rhbz#1096787
 * main.set_statement_notembedded_binlog           : rhbz#1096787
 * main.ssl_7937                                   : rhbz#1096787
 * main.ssl_crl_clients                            : rhbz#1096787
 * main.ssl_cert_verify                            : rhbz#1096787
 * perfschema.nesting                              : rhbz#1096787
 * perfschema.socket_summary_by_event_name_func    : rhbz#1096787
 * perfschema.socket_summary_by_instance_func      : rhbz#1096787

Comment 24 Michal Schorm 2016-12-01 00:32:00 UTC
Directing back to F23, because it is the last Fedora version, that use mariadb-10.0.x

For F>=24, which use mariadb 10.1.x, was created appropriate tracker
  └── https://bugzilla.redhat.com/show_bug.cgi?id=1399847

______________________________________________________________________________________

In version F23 - mariadb-10.0.28 - only following tests are currently failing
  ├── main.ssl_7937
  ├── main.ssl_crl_clients
  ├── main.ssl_cert_verify
  ├── perfschema.nesting
  ├── perfschema.socket_summary_by_event_name_func
  └── perfschema.socket_summary_by_instance_func

______________________________________________________________________________________

Changes were pushed, but waiting to be built (aggregating patches for a last build before F23 EOL )

Comment 25 Fedora End Of Life 2016-12-20 12:48:02 UTC
Fedora 23 changed to end-of-life (EOL) status on 2016-12-20. Fedora 23 is
no longer maintained, which means that it will not receive any further
security or bug fix updates. As a result we are closing this bug.

If you can reproduce this bug against a currently maintained version of
Fedora please feel free to reopen this bug against that version. If you
are unable to reopen this bug, please file a new report against the
current release. If you experience problems, please add a comment to this
bug.

Thank you for reporting this bug and we are sorry it could not be fixed.


Note You need to log in before you can comment on or make changes to this bug.