Web Hosting Talk : Web Hosting Main Forums : Hosting Security and Technology : Server load problems after upgrading to centOS 3.3?

[Serhat]

Today my server is upgraded to CentOS 3.3 automatically and then I started to have problems.

Cpu load is very very high. This is the top output:

12929 root 0 98.9 0.0 /usr/lib/sa/sadc11/var/log/sa/sa21
10859 root 0 98.1 0.0 /usr/lib/sa/sadc11/var/log/sa/sa21

Top Process %CPU 99.9 /usr/lib/sa/sadc 1 1 /var/log/sa/sa21
Top Process %CPU 99.8 /usr/lib/sa/sadc 1 1 /var/log/sa/sa21

Server Load 19.41 (2 cpus )

I kill them manually and the load goes down again. But in a few minutes those "sa21" things start to appear and server load goes high again.

Any ideas how to fix it?

[andreyka]

It is System activity data collector.
May be it broken. I reccomed disable it, if you not need this informatoin.
Or rebuild it from source and test before use.

[Projektzwo]

Temporary workaround for ya..

# vi /etc/cron.d/sysstat

put a # to line 2 ...

# */10 * * * * root /usr/lib/sa/sa1 1 1

[BLazeD]

I have this problem as well

[Scotty_B]

Got the same problem here too.

[Steven]

We experienced this issue on RHE a frew weeks back. Spawned 500 processes of the software, we have disabled it for now.

[galantina]

I have a brend new server with centos 3.3 and it worked fine for about two weaks, till this morning. Thaen this problem arised. Somehow i have a feeling that this problem araised on a same day for all of us, except thelinuxguy. Is it possible that this is some kind of virus or something?

[Steven]

Well Redhat enterprise released alot of updates that centos did not release until this week.

[Neosurge]

heh, wasn't that the point of redhat ENTERPRISE.. To have solid pre-tested releases?

[Steven]

Umm no.. centos is just a copy of redhat enterprise, so you cannot rely on centos to be stable since they are just a copy of rhe that is free. RHE has released several unstable releases.

[Neosurge]

I wasn't talking about centos.. I was talking about redhat enterprise and the problems you had with the System activity data collector.

[galantina]

I contacted my hosting company and they have aplied some update for this. I don't know which update is it. They told me that the problem is now solved by this update.

[gounder]

emm I noticed the same problem as well.

[F5Hosting]

From the cPanel forum

Originally Posted by AviN

The sadc problem can be resolved by slightly downgrading the sysstat RPM package and then preventing yum from upgrading the package. The commands below will complete these steps automatically.

killall -9 sadc
wget \
http://ibiblio.org/pub/linux/distrib...tos.2.i386.rpm
rpm -Uvh --force sysstat-4.0.7-4.centos.2.i386.rpm
cp /etc/yum.conf /etc/yum.conf.backup
sed 's/exclude=*/exclude=sysstat /g' /etc/yum.conf.backup > /etc/yum.conf

You can then try running the command "/usr/lib/sa/sadc". If it exists immediately rather than hanging, the problem has been resolved.

Worked a treat!

[gounder]

Thanks dgbaker.