Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Not a Bug
-
10.5.12
-
None
Description
Shutdown hangs after upgrade to 10.5.12
Shutdown hangs was reported on some sites after upgrade to 10.5.12 (from 10.5.10).
This is semisync replication. Master was started and running while 2 slaves hang on shutdown (done with innodb_fast_shutdown=0)
Configuration parameters are same between those site that work and does not.
Backtraces are attached here.
I am trying to get more details, but according to initial report some nodes work without problems, while other hangs (for days).
Some of parameters are:
MemTotal: 32471016 kB(30.97G)
innodb_buffer_pool_size 4294967296 (4GB)
tmp_table_size 16777216 (16MB) <--I asked bump up to 134217728 (128MB). But it was rejected since "worked before upgrade"
Max_used_connections 32
innodb_autoinc_lock_mode 1
innodb_flush_log_at_timeout 1
Attachments
Issue Links
- relates to
-
MDEV-22340 Server hangs on ibuf_merge_or_delete_for_page during shutdown with innodb_fast_shutdown=0
- Closed
-
MDEV-23839 innodb_fast_shutdown=0 hang on change buffer merge
- Closed