Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
10.1(EOL)
-
None
Description
For example, see
- https://mariadb.atlassian.net/browse/MDEV-7413
- https://mariadb.atlassian.net/browse/MDEV-5308
one can find many more examples
The first example has:
Query (0x7f97ab53b020): is an invalid pointer
|
Connection ID (thread ID): 591
|
Status: NOT_KILLED
|
The second one has:
Nov 18 16:22:21 thorstenknbl1 mysqld: Query (0x7f4ecd82d020): is an invalid pointer
|
Nov 18 16:22:21 thorstenknbl1 mysqld: Connection ID (thread ID): 6
|
Nov 18 16:22:21 thorstenknbl1 mysqld: Status: NOT_KILLED
|
Note that the second one is just an assertion failure. Stack or THD structure
are not corrupted. However, the query pointer is still not printed.
Attachments
Activity
Hi Sergei,
Can you review a patch for this problem? The fix is targeted for 10.0. But I think it should work for any version as we're not introducing anything that is "unstable".
http://lists.askmonty.org/pipermail/commits/2016-May/009405.html
Thanks!
CC: psergey
I don't see any improvement, we are still getting the same "invalid pointer" in reports from the real-life world:
Server version: 10.1.16-MariaDB-1~trusty
|
key_buffer_size=16777216
|
read_buffer_size=131072
|
max_used_connections=23
|
max_threads=502
|
thread_count=13
|
It is possible that mysqld could use up to
|
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 1118998 K bytes of memory
|
Hope that's ok; if not, decrease some variables in the equation.
|
|
Thread pointer: 0x0x7f57a83d6008
|
Attempting backtrace. You can use the following information to find out
|
where mysqld died. If you see no messages after this, something went
|
terribly wrong...
|
stack_bottom = 0x7f61375fedf0 thread_stack 0x80000
|
/usr/sbin/mysqld(my_print_stacktrace+0x2e)[0x7f65391b28fe]
|
/usr/sbin/mysqld(handle_fatal_signal+0x2d5)[0x7f6538cd9e75]
|
/lib/x86_64-linux-gnu/libpthread.so.0(+0x10330)[0x7f653722f330]
|
/usr/sbin/mysqld(+0x4812d4)[0x7f6538baa2d4]
|
/usr/sbin/mysqld(_ZN13st_select_lex5printEP3THDP6String15enum_query_type+0x4b2)[0x7f6538baac02]
|
/usr/sbin/mysqld(_ZN14Item_subselect5printEP6String15enum_query_type+0x5f)[0x7f6538d69faf]
|
/usr/sbin/mysqld(_ZN22Item_func_conv_charset5printEP6String15enum_query_type+0x44)[0x7f6538d57e64]
|
/usr/sbin/mysqld(_ZN9Item_func10print_argsEP6Stringj15enum_query_type+0x4f)[0x7f6538d3a9df]
|
/usr/sbin/mysqld(_ZN9Item_func5printEP6String15enum_query_type+0x4d)[0x7f6538d3aabd]
|
/usr/sbin/mysqld(_ZN9Item_func10print_argsEP6Stringj15enum_query_type+0x4f)[0x7f6538d3a9df]
|
/usr/sbin/mysqld(_ZN9Item_func5printEP6String15enum_query_type+0x4d)[0x7f6538d3aabd]
|
/usr/sbin/mysqld(_ZN4Item17print_item_w_nameEP6String15enum_query_type+0x1c)[0x7f6538cf0fdc]
|
/usr/sbin/mysqld(_ZN13st_select_lex5printEP3THDP6String15enum_query_type+0x226)[0x7f6538baa976]
|
/usr/sbin/mysqld(_ZN18st_select_lex_unit5printEP6String15enum_query_type+0x57)[0x7f6538b4c267]
|
/usr/sbin/mysqld(_Z29mysqld_show_create_get_fieldsP3THDP10TABLE_LISTP4ListI4ItemEP6String+0x302)[0x7f6538bb9eb2]
|
/usr/sbin/mysqld(_Z18mysqld_show_createP3THDP10TABLE_LIST+0xc0)[0x7f6538bba8e0]
|
/usr/sbin/mysqld(_Z21mysql_execute_commandP3THD+0x3fbc)[0x7f6538b5aecc]
|
/usr/sbin/mysqld(_Z11mysql_parseP3THDPcjP12Parser_state+0x26d)[0x7f6538b6073d]
|
/usr/sbin/mysqld(_Z16dispatch_command19enum_server_commandP3THDPcj+0x2460)[0x7f6538b63a80]
|
/usr/sbin/mysqld(_Z10do_commandP3THD+0x169)[0x7f6538b64239]
|
/usr/sbin/mysqld(_Z24do_handle_one_connectionP3THD+0x18a)[0x7f6538c2889a]
|
/usr/sbin/mysqld(handle_one_connection+0x40)[0x7f6538c28a70]
|
/lib/x86_64-linux-gnu/libpthread.so.0(+0x8184)[0x7f6537227184]
|
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d)[0x7f653674637d]
|
|
Trying to get some variables.
|
Some pointers may be invalid and cause the dump to abort.
|
Query (0x7f57a70fe020): is an invalid pointer
|
Connection ID (thread ID): 478
|
Hi elenst,
Can you paste the full crash message? The added string printing is at the end. No change in behaviour is done at the Query (address): <string>.
About debug info not being available for some packages: see
MDEV-4646(for RPMs) andMDEV-658(for debs).The issue with debug info is, however, orthogonal to the problem of why we keep getting Query (0x...): is an invalid pointer so often.