[Mariadb] Since updating to 10.4.17 getting system load notifications

B

burekas

Verified User
Joined
Feb 9, 2020
Messages
47
there are no hi-load processes at this TOP snapshot, usually system sends few emails, so you can see who caused overload only in first, while it still in TOP, then another emails shows how load decreased with time.
also mysql requests at bottom don't looks like problematic.
you can modify load-monitor sesitivity, for example if you have 2 CPU cores - it's better to change it from default LA 10 to 5 or less.

Version 1.36.0 | Directadmin Docs

DirectAdmin Knowledge Base
www.directadmin.com
 
This is an older topic.

But it looks like this server is out of ram.
The high IO wait could also mean it is heavily swapping.
 
Thanks for your comments.
This topic is still relevant, btw.

Currently I have 4GB RAM on the server.

The 2 main disks are RAID configured.

And according to another thread I have, it seems that I have some problematic parameters for mysql or parameters that not set yet:

Help with mysql/mariadb configuration

Hi, I would like to get your advice regarding my Mysql/Mariadb configuration. I ran this tool of MySQLTuner. That is their diagnostic: https://paste.ubuntu.com/p/h3v6FFWYg5 And that is my currently configuration: https://paste.ubuntu.com/p/CpDyjY5Y6D * Basically the majority of my tables in...
forum.directadmin.com forum.directadmin.com

I mentioned the mariddb upgrade because this start happen since 1-2 monthes ago, I think since the upgrade of mariadb so I thoughtthere is might be an impact.
 
You can check what process is using swap memory with this commando:
Code: 
for file in /proc/*/status ; do awk '/VmSwap|Name/{printf $2 " " $3}END{ print ""}' $file; done | sort -k 2 -n -r | less

If mariadb is using much swap.
Try to (temporary) undo the mariadb configuration changes. And restart mariadb afterwards.

Or just add more ram.
 
Driesp said:
You can check what process is using swap memory with this commando:
Code: 
for file in /proc/*/status ; do awk '/VmSwap|Name/{printf $2 " " $3}END{ print ""}' $file; done | sort -k 2 -n -r | less

If mariadb is using much swap.
Try to (temporary) undo the mariadb configuration changes. And restart mariadb afterwards.

Or just add more ram.
Click to expand...
Those are on the top of the list:
gnome-shell 104052 kB
httpd 78660 kB
httpd 77404 kB
named 69048 kB

What is "gnome-shell"?
 
ikkeben said:
Gnome shell is for desktop enviroment.
Do you need that on a DA server?
Click to expand...
Oh yes, now I know what is it.
Usually I don't use it, just on some particular situation when I have a problem with the server, but it's rare.
But why does it take so much? Might this trigger the system overload issue?
Because it was always enabled, and there wasn't any problem with this until a month ago when the "system load" warning started to show up.
 
burekas said:
Oh yes, now I know what is it.
Usually I don't use it, just on some particular situation when I have a problem with the server, but it's rare.
But why does it take so much? Might this trigger the system overload issue?
Because it was always enabled, and there wasn't any problem with this until a month ago when the "system load" warning started to show up.
Click to expand...
Don't know when what, but many many sites are now in Corona times and Shopping online and more, so more visits.

Then a problem with loads when not much "spare" could hapen... So you can take that part (gnomeshell) off and only load it when you need to use it, to check if it helps.

Still check disks!
 
I see this is older, but going to throw it out there as I just went through this similar - loads were high and system response sucked (websites) - using php (NOT wordpress) and DB backends. Running MariaDB 10.x and php 5.6 crushed the system, even with SSD raid sets. We downgraded back to mySQL 5.7 and same php 5.6 and the system went from 15 second TTFB on the web pages to sub 1 second response time! No real time to do lots of troubleshooting - we did try tuning the hell out of php and MariaDB, but the fact that switching back to mySQL 5.7 solved everything - well, there you go. We could not go to mySQL 8 because of changes in schema.
 
You must log in or register to reply here.
Back
Top