Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.5.6
-
None
-
None
-
None
-
Alpine Linux (LXC container), mariadb-10.5.6-r0
architecture: armhf (32bit)
hw specs: https://www.turris.com/en/omnia/specifications/
Description
I have a tiny 250MB database dedicated to a personal nextcloud instance and running inside LXC container. It's been running w/o for months but recently I noted increased temperatures and it got my attention.
It was a bit of a surprise to find out that even with a database totally idle (application is shutdown) one of two CPU cores is 100 % busy with mariadb process.
# mysqld --print-defaults
|
mysqld would have been started with the following arguments:
|
--query_cache_type=on --innodb_buffer_pool_size=10M --symbolic-links=0 --skip-networking
|
lldb is not available in repositories and neither are debugging symbols mariadb. This is output of strace on one (and only) busy thread, run for approx. 10 seconds:
% time seconds usecs/call calls errors syscall
|
------ ----------- ----------- --------- --------- ----------------
|
84.62 0.110000 873 126 66 futex
|
15.38 0.020000 645 31 poll
|
0.00 0.000000 0 1 close
|
0.00 0.000000 0 1 access
|
0.00 0.000000 0 1 fcntl64
|
0.00 0.000000 0 1 1 madvise
|
0.00 0.000000 0 1 gettid
|
0.00 0.000000 0 396 clock_gettime
|
0.00 0.000000 0 30 sendto
|
0.00 0.000000 0 90 30 recvfrom
|
0.00 0.000000 0 2 1 setsockopt
|
0.00 0.000000 0 10 9 accept4
|
0.00 0.000000 0 396 396 clock_gettime64
|
------ ----------- ----------- --------- --------- ----------------
|
100.00 0.130000 119 1086 503 total
|