Server hang in 5.5.27-rel28
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Percona Server moved to https://jira.percona.com/projects/PS |
Fix Released
|
Critical
|
Unassigned | ||
5.1 |
New
|
Undecided
|
Unassigned | ||
5.5 |
Fix Released
|
Critical
|
Unassigned |
Bug Description
After upgrading one of my database servers to 5.5.27-rel28 (from 5.5.24-rel26) this morning , it ran for about an hour before hanging all connections due to semaphore problems.
This appears to be related to #1026926 or the fix for it. We never had the original problem, but there's something wrong in this release for sure.
I've rolled this server back to 5.5.25a-27.1 for now.
Here's the my.cnf and I've attached the error log.
Dan.
[mysqladmin]
socket=
[mysql]
socket=
[mysqld]
#skip-slave-start
datadir=/data/mysql
socket=
#performance_schema
slow-query-log = 1
slow-query-
# Log all queries, subject to the rate limit below
long_query_time = 0
# Log queries that are engaging in these kinds of anti-social behaviours
# log_slow_filter = full_scan,
# Enable microsecond level precision for query times
slow_query_
# Log as much detail as we can, including query plan, etc.
log_slow_verbosity = full
# Log every nth session (1 = all)
log_slow_rate_limit = 250
# This will log queries that have data that isn't covered by indexes
# log-queries-
server-id = 14
log-bin=
relay-log=
# gt disabled due to disk space issues on 3/4/10
#log_slave_updates
max_connections=768
max_user_
join_buffer_size=2M
wait_timeout=15
myisam_
table_cache=4000
table_definitio
thread_
connect_timeout=5
max_allowed_
max_connect_
key_buffer_size=16M
skip-name-resolve
innodb_
innodb_
read_buffer_
read_rnd_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_flush_method = ALL_O_DIRECT
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
innodb_
# Added 8/25/2009 by areitz.
default-
# Added 3/10/2010 by gtoubassi
character-
#query_cache_limit = 2M
#query_cache_size = 192M
query_cache_type = OFF
#bind-address=
max_heap_
tmp_table_size=32M
[mysql.server]
user=mysql
[mysqld_safe]
err-log=
pid-file=
malloc-
tags: | added: i24252 |
tags: | added: i25775 |
Thanks for your report.
There is a workaround of innodb_ lazy_drop_ table=1. Can you see if it helps for you?