[MDEV-11147] Crash during ALTER TABLE DROP PARTITION Created: 2016-10-26  Updated: 2016-11-10  Resolved: 2016-11-10

Status: Closed
Project: MariaDB Server
Component/s: Data Definition - Alter Table, wsrep
Affects Version/s: 10.1.14
Fix Version/s: 10.1.18

Type: Bug Priority: Major
Reporter: Rodney Antonio Ramos Assignee: Nirbhay Choubey (Inactive)
Resolution: Duplicate Votes: 0
Labels: crash
Environment:

Operating system: 2 nodes

  • Linux 3.10.0-327.18.2.el7.x86_64 #1 SMP Thu May 12 11:03:55 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
  • CentOS Linux release 7.2.1511 (Core)
  • 16 GB RAM
  • 4 vCPU

MariaDB:

  • Server version: 10.1.14-MariaDB MariaDB Server

Galera Cluster: 2 nodes

  • Galera 25.3.15(r3578)

Attachments: File messages.crash.2016-10-26     File my.cnf     File zbx_db_partitioning.pl    
Issue Links:
Duplicate
is duplicated by MDEV-9416 MariaDB galera got signal 11 when alt... Closed

 Description   

I have two hosts running MariaDB with Galera Cluster.

This DB is used by the Monitoring Application Zabbix.

The database is very big, more than 200GB of data.

I have a script to make partitioning in some tables of the database.

Sometimes, during the partitioning process, MariaDB crashes, with the messages in attachment.

The partitioning script and my.cnf are also in attachment.

Any help?

Thanks

 Comments   
Comment by Elena Stepanova [ 2016-11-02 ]

From the error log:

Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(handle_fatal_signal+0x38d) [0x7f9302c7ff0d]
 
Oct 26 10:30:05 npaa2629 mysqld: /lib64/libpthread.so.0(+0xf100) [0x7f930229e100]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(ha_abort_transaction(THD*, THD*, char)+0x9f) [0x7f9302c8a8bf]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(wsrep_abort_thd(void*, void*, char)+0x139) [0x7f9302c315f9]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(wsrep_grant_mdl_exception(MDL_context*, MDL_ticket*, MDL_key const*)+0x2a7) [0x7f9302c22687]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(MDL_lock::can_grant_lock(enum_mdl_type, MDL_context*, bool) const+0x11f) [0x7f9302bd649f]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(MDL_context::try_acquire_lock_impl(MDL_request*, MDL_ticket**)+0xf1) [0x7f9302bd7121]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(MDL_context::acquire_lock(MDL_request*, double)+0x2e) [0x7f9302bd768e]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(MDL_context::upgrade_shared_lock(MDL_ticket*, enum_mdl_type, double)+0xae) [0x7f9302bd837e]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(mysql_alter_table(THD*, char*, char*, HA_CREATE_INFO*, TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool)+0x12fd) [0x7f9302b8a22d]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(Sql_cmd_alter_table::execute(THD*)+0x61a) [0x7f9302bd135a]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(mysql_execute_command(THD*)+0x1206) [0x7f9302afeb16]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(mysql_parse(THD*, char*, unsigned int, Parser_state*)+0x28e) [0x7f9302b0717e]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(+0x438989) [0x7f9302b07989]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(dispatch_command(enum_server_command, THD*, char*, unsigned int)+0x1fb0) [0x7f9302b0a010]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(do_command(THD*)+0x169) [0x7f9302b0aeb9]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(do_handle_one_connection(THD*)+0x18a) [0x7f9302bce62a]
 
Oct 26 10:30:05 npaa2629 mysqld: /usr/sbin/mysqld(handle_one_connection+0x40) [0x7f9302bce800]
 
Oct 26 10:30:05 npaa2629 mysqld: /lib64/libpthread.so.0(+0x7dc5) [0x7f9302296dc5]
 
Oct 26 10:30:05 npaa2629 mysqld: /lib64/libc.so.6(clone+0x6d) [0x7f93006b8ced]
 
Oct 26 10:30:05 npaa2629 mysqld: Trying to get some variables.
 
Oct 26 10:30:05 npaa2629 mysqld: Some pointers may be invalid and cause the dump to abort.
 
Oct 26 10:30:05 npaa2629 mysqld: Query (0x7f8f41bf8020): ALTER TABLE zabbix.history DROP PARTITION p2016_10_17
 
Oct 26 10:30:05 npaa2629 mysqld: Connection ID (thread ID): 501884
 
Oct 26 10:30:05 npaa2629 mysqld: Status: NOT_KILLED

Comment by Rodney Antonio Ramos [ 2016-11-04 ]

I found a problem on my servers that can be related with this issue.

I have 16 GB of RAM memory on each server, but the swap memory was with only 2 GB.

I´ve increased de swap to 16 GB too and since then no crash was observed.

I still will wait for some days berfore close this problem.

Thanks.

Comment by Nirbhay Choubey (Inactive) [ 2016-11-10 ]

rodneyra This issue looks pretty similar to one fixed in 10.1.18. (MDEV-9416)

Generated at Thu Feb 08 07:47:40 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.