Details
-
Bug
-
Status: Closed (View Workflow)
-
Blocker
-
Resolution: Fixed
-
10.1.23, 10.2.6
Description
When a tablespace is being removed from the InnoDB tablespace cache, we occasionally have an assertion failing in 10.2:
2017-04-27 10:38:42 0x4538 InnoDB: Assertion failure in file D:\win32-debug\build\src\storage\innobase\fil\fil0fil.cc line 2875
|
InnoDB: Failing assertion: space->n_pending_ops == 0
|
This has failed at least twice in 10.2:
non-debug x86, Debian Stretch
debug x86, Microsoft Windows
The likely cause is the MDEV-12428 fix, which allows I/O threads to increment the fil_space_t::n_pending_ops count even after the tablespace has been marked for dropping.
We must review the code also in 10.1 before the 10.1.23 release.
Attachments
Issue Links
- relates to
-
MDEV-11581 Mariadb starts innodb encryption threads when key has not changed or data scrubbing turned off
-
- Closed
-
-
MDEV-11738 Mariadb uses 100% of several of my 8 cpus doing nothing
-
- Closed
-
-
MDEV-11929 During delete: InnoDB: Assertion failure in thread 139757855127296 in file buf0buf.cc line 4803
-
- Closed
-
-
MDEV-12253 Buffer pool blocks are accessed after they have been freed
-
- Closed
-
-
MDEV-12396 MariaDB Server SEGV crash when importing tablespace for a large table (4GB)
-
- Closed
-
-
MDEV-12428 SIGSEGV in buf_page_decrypt_after_read() during DDL
-
- Closed
-
-
MDEV-12637 main.range_vs_index_merge_innodb fails in buildbot with timeout
-
- Closed
-
-
MDEV-12694 test failure: encryption.create_or_replace hangs during DROP for CREATE OR REPLACE TABLE
-
- Closed
-
-
MDEV-12898 encryption.innodb-redo-badkey failed in buildbot, server hung on startup
-
- Closed
-
Activity
Field | Original Value | New Value |
---|---|---|
Link |
This issue relates to |
Link |
This issue relates to |
Link |
This issue relates to |
Link |
This issue relates to |
Status | Open [ 1 ] | In Progress [ 3 ] |
Assignee | Marko Mäkelä [ marko ] | Jan Lindström [ jplindst ] |
Status | In Progress [ 3 ] | In Review [ 10002 ] |
Assignee | Jan Lindström [ jplindst ] | Marko Mäkelä [ marko ] |
Status | In Review [ 10002 ] | Stalled [ 10000 ] |
Assignee | Marko Mäkelä [ marko ] | Jan Lindström [ jplindst ] |
Status | Stalled [ 10000 ] | In Review [ 10002 ] |
Status | In Review [ 10002 ] | Stalled [ 10000 ] |
Assignee | Jan Lindström [ jplindst ] | Marko Mäkelä [ marko ] |
issue.field.resolutiondate | 2017-04-28 14:16:22.0 | 2017-04-28 14:16:22.069 |
Resolution | Fixed [ 1 ] | |
Status | Stalled [ 10000 ] | Closed [ 6 ] |
Link |
This issue relates to |
Link |
This issue relates to |
Link |
This issue relates to |
Link |
This issue relates to |
Link |
This issue relates to |
Workflow | MariaDB v3 [ 80489 ] | MariaDB v4 [ 152024 ] |
bb-10.2-marko
I think that we need this patch also in the 10.1.23 release. For that, we must backport the two-phase freeing of fil_space_t (fil_space_detach(), fil_space_free_low()) from 10.2.