Higher memory usage after updating to Vesta 0.9.8-17

[prxbl]

Hi,
I am running Ubuntu 14.04 and recently updated from Vesta 0.9.8.16 to 0.9.8-17 and noticed an increase in memory usage. Here is graph, around 20:00 is when the update was installed and through the night memory usage was higher than the usual. Around 14:00 there is a drop, I did a reboot on the server thinking this would solve the problem, but it didn't as memory usage seems to be slowly going up again.

I also have htop installed on the server and it seems to be showing a different story:

A quick check with simple top gives a figure of around 700MB of memory used which is similar to what vestacp graph is showing:

I also checked with free -m command, the usage seems to be real:

How can I find out what went wrong after the update and fix this memory usage problem?

Thank you for your help!

[drsdre]

You should have a look which process take most of the memory. For example use the following command to see how much MB each process uses:

Code: Select all

ps -eo size,pid,user,command --sort -size | awk '{ hr=$1/1024 ; printf("%13.2f Mb ",hr) } { for ( x=4 ; x<=NF ; x++ ) { printf("%s ",$x) } print "" }'

[prxbl]

It looks like top 3 processes are these:
841.20 Mb /usr/sbin/mysqld
615.01 Mb /usr/bin/python /usr/bin/fail2ban-server -b -s /var/run/fail2ban
216.85 Mb rsyslogd
I also did some reading on ubuntu memory usage in general and the advice was to look at the -/+ buffers/cache: section from free -m command:
-/+ buffers/cache: used 384 free 609

Because the system is using ram to cache files so the memory appears as used but is in fact available for processes. That might explain why htop shows different numbers. It doesn't explain why there is a sudden change in the graphs though..
The system has a total of 1GB ram so it's not clear to me how the top 3 processes total over 1GB of memory usage.

[rlasmar]

Same problem here, but I saw another topic that its a bug from latest vesta update and should ignore until they fixed. The real memory usage is like you said: -/+ buffers/cache: used 384 free 609

"free -h" will show the real usage.

But until they fixed, It will be nice if someone knows a solution to fix this. In what file we must edit to fix this

[skurudo]

Unfortunately there is a bug in current version. New memory function ignores available cached/buffered memory. I can assure you that the real memory consumption on your server hasn't changed since release. Bugfix will be available soon.

Old method:

Code: Select all

[root@r6 ~]# free -m |awk '{print $4}'|head -n3 |tail -n1
404

New method:

Code: Select all

[root@r6 ~]# free -m |grep Mem |awk '{print $4}'
98

viewtopic.php?f=10&t=4590&start=10#p53313

[rlasmar]

Its just paste this code?

Or it must edited in some file?

Unfortunately there is a bug in current version. New memory function ignores available cached/buffered memory. I can assure you that the real memory consumption on your server hasn't changed since release. Bugfix will be available soon.

Old method:

Code: Select all

[root@r6 ~]# free -m |awk '{print $4}'|head -n3 |tail -n1
404

New method:

Code: Select all

[root@r6 ~]# free -m |grep Mem |awk '{print $4}'
98

viewtopic.php?f=10&t=4590&start=10#p53313

[skurudo]

Its just paste this code?
Or it must edited in some file?

It's not a fix, it's explanation why is that.
It'll be fixed with new version.