[Case 103893] Clamdscan Process Hanging with File Manager

Hi,

I can confirm this here on my side as well. Processes hang and load spikes tremendously. The "solution" here is to pkill clamdscan and enable clamd service or uninstall the WHM ClamAV plugin.

I disagree with users having to enable clamd in conjunction with the WHM ClamAV plugin as this plugin alone still provides functionality to the user. I also disagree with uninstalling the WHM ClamAV plugin and re-installing via the package manager, as this is not optimal/efficient. The WHM ClamAV plugin allows for finer control, and users may simply "flip" the clamd service "switch" for more functions. If ClamAV is installed via the package manager and the user wishes to add more functionality, the user must re-install back to the plugin.

Switching between the two installations is not only inefficient, but would also imply there is no functional use of the WHM ClamAV plugin without clamd service enabled, so this should be auto-enabled upon plugin installation, which I disagree with.

Requested Solutions:
-----
1) Add clamd service status check prior to scan to avoid upload hangs
2) Add tweak setting to enable/disable auto-scanning of cPanel > File Manager uploads.

Thanks,

 

If you would this would be a great thing to push to the top of the priority list. We've got thousands of cPanel VPS's deployed and most of them have Clamd disabled in WHM so At the moment I'm constantly scanning and killing hung clamdscan processes to keep things in check.

The sooner you can get that bug patched the better.

 

Here's a one liner which resolves this issue:

Code:

ps axu | grep clamdscan | grep -v grep | awk '{print $2}' | xargs kill; rm -f /etc/clamddisable; /scripts/restartsrv_clamd; 

It wouldn't be so bad if it just broke the file manager but the load issues it's causing it a real headache. I hope it's fixed soon.

 

I'll second that. When managing tons of servers the resource usage across them is the worst part of this which is why we are hoping for a speedy fix.

 

Does this also affect clamdscan when the 'clamd' service is enabled? I have a slightly different although just as disturbing set of symptoms:

Code:

 486855 root      20   0 35884 1336 1088 R 100.0  0.0  18:03.54 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
 380692 root      20   0 35884 1340 1088 R 97.1  0.0  52:28.64 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
 513546 root      20   0 35884 1336 1088 R 92.2  0.0   8:27.00 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
 329637 root      20   0 35884 1336 1088 R 77.7  0.0  62:22.67 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd

Or, as it appears in ps axjf:

Code:

     1  329634  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 329634  329637  621170  502009 ?             -1 R        0  63:23  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
      1  380686  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 380686  380692  621170  502009 ?             -1 R        0  53:29  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
      1  486848  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 486848  486855  621170  502009 ?             -1 R        0  19:04  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
      1  513518  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 513518  513546  621170  502009 ?             -1 R        0   9:27  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd

We're not certain here what's causing this but if left unchecked, it starts eating CPU power... all of it. Linux is generally smart enough to task-switch around this, but unlike the clamav resident scanner, it doesn't seem to honor process nice levels.

Any idea what's going on here?

 

I've noticed something similar but not quite identical occurring on a few of our servers here.

First, a snippet of ps axjf:

Code:

     1  329634  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 329634  329637  621170  502009 ?             -1 R        0  63:23  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
      1  380686  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 380686  380692  621170  502009 ?             -1 R        0  53:29  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
      1  486848  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 486848  486855  621170  502009 ?             -1 R        0  19:04  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
      1  513518  621170  502009 ?             -1 S        0   0:00 /usr/local/cpanel/3rdparty/bin/perl /usr/local/cpanel/scripts/restartsrv_clamd --check
 513518  513546  621170  502009 ?             -1 R        0   9:27  \_ /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd

And then a view from top:

Code:

 380692 root      20   0 35884 1340 1088 R 99.3  0.0  60:17.85 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
 486855 root      20   0 35884 1336 1088 R 99.3  0.0  25:52.77 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
 329637 root      20   0 35884 1336 1088 R 97.4  0.0  70:12.59 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
 513546 root      20   0 35884 1336 1088 R 93.8  0.0  16:16.24 /usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd

This was captured from one of our servers with the clamav software installed from the 'manage plugins' section of WHM, and with the 'clamd' service enabled in our service manager.

It's using a very high amount of CPU, and we really can't tell what it's doing with it all, and that has us concerned. Is this related? Will it get fixed with that planned update or should this be a different thread?

 

Our server have similar issue with clamdscan. Hope to have permanent solution as it cause unnecessary load.

 

Do you have an ETA on when this issue will be resolved?

 

root spawning clamdscan which is consuming all resources

For the past week or so I have been getting High Load alerts via email from my WHM VPS. While investigating it appears root keeps on spawning clamdscans which consume a high amount of CPU.

This happens almost hourly, sometimes it is worst than others.

Has anyone experienced this before? How can I determine what is spawning these via root and how can I stop it from happening?

Code:

Pid	Owner	Priority	CPU %	Memory %	Command
9692 (Trace) (Kill)	root	-5	  53.9	0.0	/usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
3004 (Trace) (Kill)	root	-5	  43.1	0.0	/usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
11609 (Trace) (Kill)	root	-5	  39.1	0.0	/usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
30857 (Trace) (Kill)	root	-5	  36.2	0.0	/usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
5215 (Trace) (Kill)	root	-5	  32.6	0.0	/usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd
15655 (Trace) (Kill)	root	-5	  24.8	0.1	/usr/local/cpanel/3rdparty/bin/clamdscan --quiet --no-summary /etc/passwd

CENTOS 6.5 x86_64 xenpv – s1 WHM 11.44.0 (build 18)

 

Your output and symptoms look similar to the issue we are seeing, but you say that clamd is enabled in WHM > Manage Services? Are you sure that you killed the hung clamdscan processes as indicated above? You may try WiredChris' command above. You may also kill off those hung clamd restart processes as well.

Thanks,

 

I can confirm clamd is definitely enabled on our service manager in WHM.

That command from above does three things:
ps axu | grep clamdscan | grep -v grep | awk '{print $2}' | xargs kill; <-- does the killing of the hung processes
rm -f /etc/clamddisable; <-- tries to remove a file that isn't on our server.
/scripts/restartsrv_clamd; <-- spawns a new hung thread almost immediately.

At least for us, it's the actual attempt to restart clamd using that restartsrv_clamd or even the --check version that triggers another copy of this clamdscan thread to run, and it just sits there, using the full power of a CPU core for no apparent reason.

 

Problem with clamdscan

I have been getting a lot of heavy load reports from my servers , and by checking i found out that clamav cpanel plugin is consuming a LOT of resources scanning /etc/password , spawning a lot of processes for the same purpose which is scanning /etc/password .. what seems to be the problem ?

 

I think the problem is with the specific version om clamav.

Re: [clamav-users] Version 0.98.3 hard loops on "clamdscan -V"

Just saw this problem this day (autoupdates)
# rpm -qa |grep clamav
cpanel-clamav-virusdefs-0.98.3-1.cp1140.x86_64
cpanel-clamav-0.98.3-1.cp1140.x86_64

Cheers,
Josef

 

Re: root spawning clamdscan which is consuming all resources

I'm having the exact same issue. Is there a fix in the works?

 

Is there any way to fix this, even temporarily??? Users can't upload, edit or delete website files at all!