bug891496.sh fails in galera55 builds

Bug #1203668 reported by Alexey Kopytov
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Percona XtraBackup moved to https://jira.percona.com/projects/PXB
Invalid
Undecided
Unassigned

Bug Description

After rebasing XB on Percona-Server-5.5.31-rel30, the bug891496.sh test fails as follows:

http://jenkins.percona.com/view/XtraBackup/job/percona-xtrabackup-2.1-param/382/BUILD_TYPE=release,Host=debian6-64,xtrabackuptarget=galera55/testReport/junit/%28root%29/t_bug891496/sh/

The test does a fairly basic thing:

1. It starts the server with innodb_data_file_path=ibdata1:3M;ibdata2:10M:autoextend
2. Takes a streaming backup
3. Prepares and restores.
4. Then starts the server. At this point Percona-Server-5.5.30-rel30.1 starts up just fine, but PXC 5.5.31-23.7.5 crashes on startup.

I'll see if PS 5.5.31 crashes in the same way. If not, it looks like a PXC bug. Assigning to Raghu for now to check if it looks familiar.

Changed in percona-xtrabackup:
milestone: none → 2.1.4
assignee: nobody → Raghavendra D Prabhu (raghavendra-prabhu)
Revision history for this message
Raghavendra D Prabhu (raghavendra-prabhu) wrote :

Nothing related I can think of, I don't see any debug symbols either, so I need to reproduce this locally.

I have fixed one another bug in another lp issue, lets see if this fixes that.

Changed in percona-xtrabackup:
status: New → Triaged
Revision history for this message
Raghavendra D Prabhu (raghavendra-prabhu) wrote :

So, I was able to get following backtrace on trunk build of PXC:

0x7e2dbc my_print_stacktrace + 44
0x6ae3b1 handle_fatal_signal + 1185
0x7fa51c38f0e0 _end + 456080384
0x7fa51bf6b66c _end + 451740044
0x8998d9 buf_page_is_corrupted + 1209
0x8dd353 fil_check_first_page + 211
0x8e21dd fil_read_first_page + 381
0x84c51a innobase_start_or_create_for_mysql + 3978
0x7fb395 innobase_init(void*) + 2853
0x6b0608 ha_initialize_handlerton(st_plugin_int*) + 72
0x5b4460 plugin_initialize(st_plugin_int*) + 96
0x5b8239 plugin_init(int*, char**, int) + 2073
0x5314b1 init_server_components() + 657
0x5365f1 mysqld_main(int, char**) + 2065
0x7fa51a914a15 _end + 428315957
0x52bc9d _start + 41

However, this is same as the multi ibdata1 regression's backtrace.

Next, I tried a build from lp:~percona-core/percona-xtradb-cluster/percona-xtradb-cluster-for-xtrabackup-tests/

and it passed.

=====================

./run.sh -d /pxc -t t/bug891496.sh
Detecting server version...
Running against Percona Server 5.5.31-23.7.5-log (XtraDB 5.5.31-rel30.3)
Using 'xtrabackup_55' as xtrabackup binary
Using 8 parallel workers

==============================================================================
TEST WORKER RESULT TIME(s) or COMMENT
------------------------------------------------------------------------------
bug891496 w1 [passed] 25
==============================================================================
Spent 25 of 30 seconds executing testcases

SUMMARY: 1 run, 1 successful, 0 skipped, 0 failed

See results/ for detailed output

==============

(I used binaries from http://jenkins.percona.com/job/percona-xtrabackup-2.1-binaries/label_exp=centos6-64/ for this)

Hence, I suspect an old 5.5.32 build (and not lp:~percona-core/percona-xtradb-cluster/percona-xtradb-cluster-for-xtrabackup-tests/
) may have been used/caused for this crash.

Revision history for this message
Alexey Kopytov (akopytov) wrote :

Raghu,

Thanks, I will check what's actually being used there.

Changed in percona-xtrabackup:
status: Triaged → Incomplete
assignee: Raghavendra D Prabhu (raghavendra-prabhu) → Alexey Kopytov (akopytov)
Revision history for this message
Raghavendra D Prabhu (raghavendra-prabhu) wrote :

I don't think this is failing anymore.

Revision history for this message
Raghavendra D Prabhu (raghavendra-prabhu) wrote :

Yes, it is not failing here http://jenkins.percona.com/view/XtraBackup/job/percona-xtrabackup-2.1-param/BUILD_TYPE=release,Host=debian6-64,xtrabackuptarget=galera55/393/

(the xb_galera_sst socat issue should be fixed in puppet by now).

Revision history for this message
Alexey Kopytov (akopytov) wrote :

Great, closing.

Changed in percona-xtrabackup:
status: Incomplete → Invalid
assignee: Alexey Kopytov (akopytov) → nobody
Changed in percona-xtrabackup:
milestone: 2.1.4 → none
Revision history for this message
Shahriyar Rzayev (rzayev-sehriyar) wrote :

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXB-1240

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.