Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Duplicate
-
10.4.13
-
None
-
(tested on 18.04)
Description
Hi,
I obtained a segfault when trying to alter a table in mariadb 10.13.
It's not reproduced 100% of the time but roughly 50% of the time, and looks similar to https://jira.mariadb.org/browse/MDEV-17177 .
- mysqld --version
/usr/sbin/mysqld Ver 10.4.13-MariaDB-1:10.4.13+maria~bionic for debian-linux-gnu on x86_64 (mariadb.org binary distribution)
When launching from gdb I obtained the following backtrace :
{{
Thread 43 "mysqld" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffff03e2700 (LWP 27295)]
0x0000555555eea267 in Item_func_in::cleanup() ()
(gdb) bt
#0 0x0000555555eea267 in Item_func_in::cleanup() ()
#1 0x0000555555b69c2d in Query_arena::free_items() ()
#2 0x0000555555b6bb4b in THD::cleanup_after_query() ()
#3 0x0000555555bbc02b in mysql_parse(THD*, char*, unsigned int, Parser_state*, bool, bool) ()
#4 0x0000555555bbe705 in dispatch_command(enum_server_command, THD*, char*, unsigned int, bool, bool) ()
#5 0x0000555555bbfbb4 in do_command(THD*) ()
#6 0x0000555555c9c23e in do_handle_one_connection(CONNECT*) ()
#7 0x0000555555c9c2fd in handle_one_connection ()
#8 0x00007ffff70d86db in start_thread (arg=0x7ffff03e2700) at pthread_create.c:463
#9 0x00007ffff5afaa3f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
}}
The errors happens when altering a table to add or remove a column, here my command was
ALTER TABLE nodegroup DROP COLUMN stack_id;
Attachments
Issue Links
- duplicates
-
-
- Closed
-
Activity
- no parallel client was writing to this table, not sure about reading but it's very unlikely
- show variables attached
- MariaDB [magnum]> SHOW CREATE TABLE nodegroup;
[...]nodegroup CREATE TABLE `nodegroup` (
`created_at` datetime DEFAULT NULL,
`updated_at` datetime DEFAULT NULL,
`id` int(11) NOT NULL AUTO_INCREMENT,
`uuid` varchar(36) NOT NULL,
`name` varchar(255) NOT NULL,
`cluster_id` varchar(255) NOT NULL,
`project_id` varchar(255) NOT NULL,
`docker_volume_size` int(11) DEFAULT NULL,
`labels` text DEFAULT NULL,
`flavor_id` varchar(255) DEFAULT NULL,
`image_id` varchar(255) DEFAULT NULL,
`node_addresses` text DEFAULT NULL,
`node_count` int(11) DEFAULT NULL,
`max_node_count` int(11) DEFAULT NULL,
`min_node_count` int(11) DEFAULT NULL,
`role` varchar(255) DEFAULT NULL,
`is_default` tinyint(1) DEFAULT NULL,
`stack_id` varchar(255) DEFAULT NULL,
`status` varchar(20) DEFAULT NULL,
`status_reason` text DEFAULT NULL,
PRIMARY KEY (`id`),
UNIQUE KEY `uniq_nodegroup0uuid` (`uuid`),
UNIQUE KEY `uniq_nodegroup0cluster_id0name` (`cluster_id`,`name`),
CONSTRAINT `CONSTRAINT_1` CHECK (`is_default` in (0,1))
) ENGINE=InnoDB AUTO_INCREMENT=13 DEFAULT CHARSET=utf8[...]
1 row in set (0.000 sec)
> It could be MDEV-22563 . 10.4.14 hasn't been released yet, but if you want to try the new version to see if the problem goes away, you can pull and build from github, or I can point you at our intermediate builds.
Ah, I missed that one in my bug search. I'm interested in dev builds, but if it's easier to build from code, it should be possible too.
You can find recent Bionic 64-bit packages here: http://hasky.askmonty.org/archive/10.4/build-33954/kvm-deb-bionic-amd64/
Thanks! Can't reproduce the segfault with those packages installed (*-common *-client-core *-client *-server-core *-server )
Is there a scheduled time for 10.4.14 ? 
> Is there a scheduled time for 10.4.14 ?
2020-07-24 according to the current schedule. It's not precise though, can be a few days off.
It could be
MDEV-22563. 10.4.14 hasn't been released yet, but if you want to try the new version to see if the problem goes away, you can pull and build from github, or I can point you at our intermediate builds.