Aria recovery crash, or assertion `!new_page' failure in _ma_apply_redo_insert_row_head_or_tail, or assertion `page_offset >= keypage_header && page_offset <= page_length' failure in _ma_apply_redo_index
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
MariaDB |
New
|
Undecided
|
Michael Widenius |
Bug Description
One of our servers that was recently upgraded to MariaDB 5.3 crashed at startup:
====
Apr 16 10:12:04 server1 mysqld_safe: Starting mysqld daemon with databases from /var/lib/mysql
Apr 16 10:12:04 server1 mysqld: 120416 10:12:04 [Warning] 'master-host' is deprecated and will be removed in a future release. Please use ''CHANGE MASTER'' instead.
Apr 16 10:12:04 server1 mysqld: 120416 10:12:04 [Note] Plugin 'BLACKHOLE' is disabled.
Apr 16 10:12:04 server1 mysqld: 120416 10:12:04 [Note] Plugin 'FEDERATED' is disabled.
Apr 16 10:12:04 server1 mysqld: 120416 10:12:04 [Note] mysqld: Aria engine: starting recovery
Apr 16 10:12:09 server1 mysqld: recovered pages: 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%2d2d2d d:2d:2d [ERROR] mysqld got signal 11 ;
Apr 16 10:12:09 server1 mysqld: This could be because you hit a bug. It is also possible that this binary
Apr 16 10:12:09 server1 mysqld: or one of the libraries it was linked against is corrupt, improperly built,
Apr 16 10:12:09 server1 mysqld: or misconfigured. This error can also be caused by malfunctioning hardware.
Apr 16 10:12:09 server1 mysqld:
Apr 16 10:12:09 server1 mysqld: To report this bug, see http://
Apr 16 10:12:09 server1 mysqld:
Apr 16 10:12:09 server1 mysqld: We will try our best to scrape up some info that will hopefully help
Apr 16 10:12:09 server1 mysqld: diagnose the problem, but since we have already crashed,
Apr 16 10:12:09 server1 mysqld: something is definitely wrong and this may fail.
Apr 16 10:12:09 server1 mysqld:
Apr 16 10:12:09 server1 mysqld: Server version: 5.3.5-MariaDB-
Apr 16 10:12:09 server1 mysqld: key_buffer_
Apr 16 10:12:09 server1 mysqld: read_buffer_
Apr 16 10:12:09 server1 mysqld: max_used_
Apr 16 10:12:09 server1 mysqld: max_threads=251
Apr 16 10:12:09 server1 mysqld: thread_count=0
Apr 16 10:12:09 server1 mysqld: connection_count=0
Apr 16 10:12:09 server1 mysqld: It is possible that mysqld could use up to
Apr 16 10:12:09 server1 mysqld: key_buffer_size + (read_buffer_size + sort_buffer_
Apr 16 10:12:09 server1 mysqld: Hope that's ok; if not, decrease some variables in the equation.
Apr 16 10:12:09 server1 mysqld:
Apr 16 10:12:09 server1 mysqld: Thread pointer: 0x0
Apr 16 10:12:09 server1 mysqld: Attempting backtrace. You can use the following information to find out
Apr 16 10:12:09 server1 mysqld: where mysqld died. If you see no messages after this, something went
Apr 16 10:12:09 server1 mysqld: terribly wrong...
Apr 16 10:12:09 server1 mysqld: stack_bottom = 0 thread_stack 0x40000
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /lib/libpthread
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: /lib/libc.
Apr 16 10:12:09 server1 mysqld: /usr/sbin/
Apr 16 10:12:09 server1 mysqld: The manual page at http://
Apr 16 10:12:09 server1 mysqld: information that should help you find out what is causing the crash.
Apr 16 10:12:09 server1 mysqld: Writing a core file
Apr 16 10:12:09 server1 mysqld_safe: mysqld from pid file /var/run/
Apr 16 10:12:18 server1 /etc/init.
Apr 16 10:12:18 server1 /etc/init.
Apr 16 10:12:18 server1 /etc/init.
Apr 16 10:12:18 server1 /etc/init.
Apr 16 10:12:18 server1 /etc/init.
====
Removing /var/lib/
summary: |
- recovered pages: 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%2d2d2d d:2d:2d - [ERROR] mysqld got signal 11 ; + Aria recovery crash, or assertion `!new_page' failure in + _ma_apply_redo_insert_row_head_or_tail, or assertion `page_offset >= + keypage_header && page_offset <= page_length' failure in + _ma_apply_redo_index |
tags: | added: aria assertion crash |
Hi Dreas,
I've tried to "recover" with 5.3.5 and your Aria logs on top of an empty database, and didn't get the crash so far. Could you please provide the schema (for starters, at least DDL, if not the data), or, if you already did it earlier in previous bug reports, point at one that we can use?
Thanks.