MySQL Service stopping frequently

hthiers

Hi team. I just found out that the MySQL service is getting down pretty frequently. A few minutes ago, I tried to access the aaPanel without luck. Nothing loads. VPS resources were normal, with memory usage about 55%.
I had to restart the panel through ssh (bt 1) to get back on it. The SQL service also came back at the same time.
Here's a view of all the system messages regarding the MySQL service issue. Does someone know why this could happen with no memory issues?

aaPanel_Kern

Hi,
Please check whether it has been shut down by the system?
Also check whether the mysql log is running normally

dmesg|grep -iE "kill|httpd|apache|http|mysql|oom"

hthiers

I can see the "Out of memory" process kill events. However, the VPS hasn't even reached 55% usage:
[1938585.387639] [1525206] 0 1525206 2001 608 96 512 0 57344 0 0 mysqld_safe [1938585.388351] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/user.slice/user-1004.slice/user@1004.service/init.scope,task=systemd,pid=1524699,uid=1004 [1938585.388393] Out of memory: Killed process 1524699 (systemd) total-vm:20288kB, anon-rss:2048kB, file-rss:2944kB, shmem-rss:0kB, UID:1004 pgtables:84kB oom_score_adj:100 [1938587.052877] monarx-agent invoked oom-killer: gfp_mask=0x140cca(GFP_HIGHUSER_MOVABLE|__GFP_COMP), order=0, oom_score_adj=0 [1938587.052923] oom_kill_process+0x118/0x280 [1938587.052932] ? oom_evaluate_task+0x143/0x1e0 [1938587.052940] __alloc_pages_may_oom+0x10c/0x1d0 [1938587.053248] [ pid ] uid tgid total_vm rss rss_anon rss_file rss_shmem pgtables_bytes swapents oom_score_adj name [1938587.053997] [1525206] 0 1525206 2001 608 96 512 0 57344 0 0 mysqld_safe [1938587.054633] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/system.slice/mysqld.service,task=mariadbd,pid=1525403,uid=1002 [1938587.054825] Out of memory: Killed process 1525403 (mariadbd) total-vm:2135068kB, anon-rss:217568kB, file-rss:2816kB, shmem-rss:0kB, UID:1002 pgtables:1032kB oom_score_adj:0 [1938603.998079] BT-Task invoked oom-killer: gfp_mask=0x140cca(GFP_HIGHUSER_MOVABLE|__GFP_COMP), order=0, oom_score_adj=0 [1938603.998276] oom_kill_process+0x118/0x280 [1938603.998290] ? oom_evaluate_task+0x143/0x1e0 [1938603.998301] __alloc_pages_may_oom+0x10c/0x1d0 [1938603.998879] [ pid ] uid tgid total_vm rss rss_anon rss_file rss_shmem pgtables_bytes swapents oom_score_adj name [1938603.999605] [1525206] 0 1525206 2001 576 96 480 0 57344 0 0 mysqld_safe [1938604.000232] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/system.slice/mysqld.service,task=mariadbd,pid=1528871,uid=1002 [1938604.000433] Out of memory: Killed process 1528871 (mariadbd) total-vm:2121936kB, anon-rss:177352kB, file-rss:2816kB, shmem-rss:0kB, UID:1002 pgtables:900kB oom_score_adj:0

Here you can see the peaks and troughs on the VPS graph:

aaPanel_Kern

Hello, how much physical memory do you have?
Does the chart show the time before mysql was shut down?
You can also check the Monitor interface

hthiers

VPS has 6 of RAM. And yes, that chart shows you two downs for when the MySQL service went down.
aaPanel monitor was disabled... I have enabled it again.

But again, I know that linux will kill a sql process when there's no free memory.
Also, it was kind of weird that I had to restart the aaPanel when the service was down. Like both things at the same time.

aaPanel_Kern

It may also be closed by the system

hthiers

So, is there any option to change that threshold? I mean, I wouldn't be that worried with < 60% of memory usage...

aaPanel_Kern

Check the slow query of mysql and check the running status of mysql

hthiers