Installing ClamAV on CentOS 7 and Using Freshclam
Install and Configure ClamAV on CentOS 7
Including the usage of Freshclam
To get ClamAV on CentOS installed, we have to use the EPEL repository (Extra Packages for Enterprise Linux). Fortunately, the Fedora project provides this with an easy installation. Unfortunately the default configuration is not properly working. In this post we collect some of the issues and required changes.
Let’s start with installing the EPEL support.
yum install epel-release
Next step is installing all ClamAV components.
yum install clamav-server clamav-data clamav-update clamav-filesystem clamav clamav-scanner-systemd clamav-devel clamav-lib clamav-server-systemd
The output should be similar to:
Configure SELinux for ClamAV
If you are using ClamAV on CentOS, together with SELinux, we should configure it a little bit. This way ClamAV can access all files on disk, and update its data definition files.
Enable antivirus_can_scan_system:
setsebool -P antivirus_can_scan_system 1
If you don’t perform this step, Freshclam will log something like:
During database load : LibClamAV Warning: RWX mapping denied: Can't allocate RWX Memory: Permission denied
Configuration of Clam daemon
Copy a the clamd.conf template, in case you don’t have a configuration file yet.
cp /usr/share/clamav/template/clamd.conf /etc/clamd.d/clamd.conf
sed -i ‘/^Example/d’ /etc/clamd.d/clamd.conf
Change /etc/clamd.d/clamd.conf file and define if you want to run the scanner as root, or a specific user. Check your /etc/passwd file for the related Clam user.
Change the following two options:
User clamscan
LocalSocket /var/run/clamd.<SERVICE>/clamd.sock
Enable Freshclam
Freshclam helps with keeping the database of ClamAV up-to-date. First delete the related “Example” line from /etc/freshclam.conf.
cp /etc/freshclam.conf /etc/freshclam.conf.bak
sed -i ‘/^Example/d’ /etc/freshclam.conf
Check the other options in the file, and change it to your preferred settings.
Missing systemd service file
We didn’t get a systemd service file, so creating a quick file here. The process should be forking itself and start freshclam in daemon mode. In this case we configure it to check 4 times a day for new files.
Create a new file /usr/lib/systemd/system/clam-freshclam.service
# Run the freshclam as daemon [Unit] Description = freshclam scanner After = network.target [Service] Type = forking ExecStart = /usr/bin/freshclam -d -c 4 Restart = on-failure PrivateTmp = true [Install] WantedBy=multi-user.target
Now enable and start the service.
systemctl enable clam-freshclam.service
systemctl start clam-freshclam.service
Check the status.
[root@centos7 system]# systemctl status clam-freshclam.service
clam-freshclam.service - freshclam scanner
Loaded: loaded (/usr/lib/systemd/system/clam-freshclam.service; enabled)
Active: active (running) since Thu 2015-06-11 11:09:24 CEST; 1s ago
Process: 3158 ExecStart=/usr/bin/freshclam -d -c 4 (code=exited, status=0/SUCCESS)
Main PID: 3159 (freshclam)
CGroup: /system.slice/clam-freshclam.service
└─3159 /usr/bin/freshclam -d -c 4
Change service files
By default, the service files seem to be messy and not working.
These are the files bundled:
[root@centos7 system]# ls -l /usr/lib/systemd/system/clam* -rw-r--r--. 1 root root 136 Apr 29 20:38 /usr/lib/systemd/system/clamd@scan.service -rw-r--r--. 1 root root 231 Apr 29 20:38 /usr/lib/systemd/system/clamd@.service
When enabling the clamd service, we would see something like this:
[root@centos7 system]# systemctl enable /usr/lib/systemd/system/clamd@.service Failed to issue method call: Unit /usr/lib/systemd/system/clamd@.service does not exist.
So let’s fix it. First rename the /usr/lib/systemd/system/clamd@.service file.
Rename the clamd@ file.
mv /usr/lib/systemd/system/clamd@.service /usr/lib/systemd/system/clamd.service
Now we have to change the clamd@scan service as well, as it refers to a non-existing file now. Change this line in /usr/lib/systemd/system/clamd@scan.service and remove the @ sign.
.include /lib/systemd/system/clamd@.service
Next step is changing the clamd service file /usr/lib/systemd/system/clamd.service
[Unit] Description = clamd scanner daemon After = syslog.target nss-lookup.target network.target [Service] Type = simple ExecStart = /usr/sbin/clamd -c /etc/clamd.d/clamd.conf --foreground=yes Restart = on-failure PrivateTmp = true[Install] WantedBy=multi-user.target
Move into the directory.
cd /usr/lib/systemd/system
Start all services.
[root@centos7 system]# systemctl enable clamd.service
[root@centos7 system]# systemctl enable clamd@scan.service
[root@centos7 system]# systemctl start clamd.service
[root@centos7 system]# systemctl start clamd@scan.service
Checking the status
With all these changes, ClamAV on CentOS 7 should be running now. The easiest way to check, is using the ps command and see if freshclam and clamd are running.
Useful resources for debugging are the systemctl status command, followed by the service. Then there is logging in /var/log/messages, which usually will reveal when and why something is (not) running.
More tips? Leave them in the comments!
4:49 AM
Thx a lot for this article! I searched a lot to get clamd work on CentOS 7!
Nice greetings from Vienna!
1:05 PM
Great to hear it helped for you as well Rocco. Any other topics you like to read on our blog?
4:21 AM
Hi Mic,
thank you very much for your posting. i have do that sequence but i have get error
“Jul 28 09:13:10 localhost.localdomain freshclam[12408]: During database load : LibClamAV Warning: RWX mapping denied: Can’t allocate RWX Memory: Permission denied”
the error disappear after type command “setsebool -P clamd_use_jit on”
Best regards / Azzinar
8:37 PM
Hi
Every time I use this line sed -i ‘/^Example/d’ /etc/freshclam.conf the ‘ gets change to . and it fails and if I use the correct character I get the same error
Any idea?
11:59 AM
Copy the line manually in a text browser and replace the quotes with single ones. It might be due to WordPress.
3:13 AM
You can use vim editor and remove this line with dd command
10:59 PM
Hi there!
Very good tutorial! I’m begginer in linux but I could follow this almost complete…
My doubts are:
1. You said to change this to any user I want, but which should be better for security matters, root or clamscan?
2. sed -i ‘/^Example/d’ /etc/clamd.d/clamd.conf (What should happen after using that command?)
3. You said to change this: LocalSocket /var/run/clamd./clamd.sock. To point I did not understand is, change what and for what value in there?
Thank you!
5:24 PM
Best is using a non-privilged user, to reduce the chance of attacks succeeding and having full permissions. The sed command will remove the “Example” line.
The socket file should point to the right directory, which is determined by the service name in the related example.
Did that help?
10:18 PM
Hi, thanks for the reply!
To be honest, no XD
Let me clarify…
2. sed -i ‘/^Example/d’ /etc/clamd.d/clamd.conf (What should happen after using that command?)
This command didn’t delete the Exampe (I think…)
When I get this point “LocalSocket /var/run/clamd./clamd.sock”, the problem is that I didn’t understood what should be changed there and under /var/run the only folder about clam is clamd.scan which there is nothing inside. It means that those socket file doesn’t exist.
Until now that are my problems! :/
As you can see I’m newbie… trying to lean XD
10:12 PM
As per above, just add this line:
LocalSocket /var/run/clamd.scan/clamd.sock
It will create your clamd.sock
5:08 PM
Thanks for the directions. I was able to get it working, however, I end up with 2 clamd processes running:
# ps -ef|grep clam
clamupd+ 11639 1 0 Nov23 ? 00:00:10 /usr/bin/freshclam -d -c 4
clamscan 25171 1 1 09:42 ? 00:00:17 /usr/sbin/clamd -c /etc/clamd.dclamd.conf –nofork=yes
clamscan 26210 1 99 10:02 ? 00:00:04 /usr/sbin/clamd -c /etc/clamd.dclamd.conf –nofork=yes
stopping clamd.service will kill one and I can manually kill -9 the other process, but it always restarts another process.
Any idea why I’m getting 2 clamd processes? Thanks.
2:23 PM
I had the same problem, turns out you must not enable both services,
just enable clamd@scan.service:
systemctl enable clamd@scan.service
systemctl disable calmd.service
Cheers, thans for the tut
2:25 PM
systemctl disable clamd.service
then restart
11:50 PM
Hi All,
Thanks to Michael Boelen for this training
Reply to friends about the following problem:
#LocalSocket /var/run/clamd./clamd.sock
1) Please first run the command:
# ls ls -la /var/run/
2) Now find the following folder clamd.???
drwx–x—. 2 clamscan clamscan 80 Feb 01 10:27 clamd.scan
3) Open the file /etc/clamd.d/clamd.conf
4) Finde the line: #LocalSocket /var/run/clamd./clamd.sock
5) Now change to: LocalSocket /var/run/clamd.scan/clamd.sock
Thanks.
6:30 PM
I am not sure why you would need to change the name of /usr/lib/systemd/system/clamd@.service
I assume this is a subprocess that is called when /usr/lib/systemd/system/clamd@scan.service is started
You would run into no error if you enable and start /usr/lib/systemd/system/clamd@scan.service without renaming clamd@.service
Thank you for your article. It was helpful.
6:48 PM
if you do:
systemctl list-unit-files –type=service
you will see:
….
clamd@.service static
clamd@scan.service enabled
…..
In particular, “static” means “enabled because something else wants it”. Think by analogy to pacman’s package install reasons:
enabled :: explicitly installed
static :: installed as dependency
disabled :: not installed
So you should not need to change or enable the clamd@.service
The dependency will just work if you enable and start clamd@scan.service
11:13 PM
Nice work. Got me up and running. I wouldn’t mind a little more at the end with a few basic commands or to know if we are all set on a reboot to auto-start, how I will know if a virus is detected and such. But I can break out the manual. Their install instructions were all kinds of wrong so glad your post was here.
4:52 PM
Great to be of help here and good feedback. Let’s help others: can you share the commands you used (after you got things set up)? Then I will add them to the article!
9:29 PM
I am dealing with a constant error when the service starts. I have touched clamd.sock then chown’d the directory and the socket to clamupdate:clamupdate. Immediately after I do that it loads properly. Howewver, upon reboot when it fixes the stale socket it the service fails to start again until I re-touch and chown. Any suggestions would be great!
Feb 1 14:24:01 mail clamd[4906]: Log file size limited to 1048576 bytes.
Feb 1 14:24:01 mail clamd[4906]: Reading databases from /var/lib/clamav
Feb 1 14:24:01 mail clamd[4906]: Not loading PUA signatures.
Feb 1 14:24:01 mail clamd[4906]: Bytecode: Security mode set to “TrustSigned”.
Feb 1 14:24:14 mail clamd[4906]: Loaded 4244905 signatures.
Feb 1 14:24:16 mail clamd[4906]: LOCAL: Socket file /var/run/clamd.scan/clamd.sock could not be bound: Permission denied
Feb 1 14:24:16 mail clamd: ERROR: LOCAL: Socket file /var/run/clamd.scan/clamd.sock could not be bound: Permission denied
Feb 1 14:24:16 mail systemd: clamd.service: main process exited, code=exited, status=1/FAILURE
Feb 1 14:24:16 mail systemd: Unit clamd.service entered failed state.
8:01 AM
Check if your temporary directory is properly created via a file in /etc/tmpfiles.d/.
You could try to force it in your service file:
ExecStartPre=/bin/mkdir -p /var/run/clamd.scan
ExecStartPre=/bin/chown -R clamuser:clamgroup /var/run/clamd.scan
5:16 PM
Have the same problem. Unfortunately forcing directory creation and owner change with ExecStartPre didnt change much. Have noticed, when i manually launch /usr/sbin/clamd -c /etc/clamd.d/clamd.conf –nofork=yes – clamd starts without any errors and socket file is created.
/etc/clamd.d/clamd.conf has the following options active:
LogSyslog yes
LocalSocket /var/run/clamd/clamd.sock
User clamscan
AllowSupplementaryGroups yes
ls -la /var/run/clamd
total 0
drwxr-xr-x. 2 clamscan clamscan 40 Apr 13 11:30 .
drwxr-xr-x. 28 root root 800 Apr 11 16:18 ..
Any ideas?
1:39 PM
I have the same problem. It is annoying. Did you managed to fix this? Thanks.
6:02 AM
Hi, when i change line #LocalSocket /var/run/clamd./clamd.sock to:
LocalSocket /var/run/clamd.scan/clamd.sock
i get this error ” clamd: ERROR: LOCAL: Socket file /var/run/clamd.scan/clamd.sock is in use by another process.”
can you tell me where i could make mistake?
1:31 PM
Easiest step to validate this is using the lsof utility and see what process keeps it open. Let us know when you found it, so it might helps others as well.
2:30 AM
If it helps… I just fixed this with killall clamd, then systemctl restart clamd.
8:00 PM
It seems the `killall` followed by `systemctl restart clamd` fixes things because there were 2 instances of clamd trying to run. The second one is the one that keeps failing and trying to start.
After I did the initial `systemctl start` commands as suggested in the main article, ps showed two running clamd’s:
# ps -aef | grep clam
clamscan 4507 1 88 10:31 ? 00:00:22 /usr/sbin/clamd -c /etc/clamd.d/clamd.conf –foreground=yes
clamscan 4522 1 98 10:32 ? 00:00:18 /usr/sbin/clamd -c /etc/clamd.d/clamd.conf –foreground=yes
root 4555 7049 0 10:32 pts/10 00:00:00 grep –color=auto clam
clamupd+ 7213 1 0 May22 ? 00:00:09 /usr/bin/freshclam -d -c 4
Later, when I did the `killall`, `tail /var/log/messages` showed:
May 23 13:47:06 gs619-g001189 clamd[21844]: LOCAL: Socket file /var/run/clamd.scan/clamd.sock is in use by another process.
May 23 13:47:06 gs619-g001189 clamd: ERROR: LOCAL: Socket file /var/run/clamd.scan/clamd.sock is in use by another process.
May 23 13:47:06 gs619-g001189 systemd: clamd@scan.service: main process exited, code=exited, status=1/FAILURE
May 23 13:47:06 gs619-g001189 systemd: Unit clamd@scan.service entered failed state.
May 23 13:47:06 gs619-g001189 systemd: clamd@scan.service failed.
May 23 13:47:06 gs619-g001189 systemd: clamd@scan.service holdoff time over, scheduling restart.
May 23 13:47:06 gs619-g001189 systemd: Stopped Generic clamav scanner daemon.
May 23 13:47:06 gs619-g001189 systemd: Started Generic clamav scanner daemon.
May 23 13:47:06 gs619-g001189 clamd[21871]: Received 0 file descriptor(s) from systemd.
May 23 13:47:06 gs619-g001189 clamd[21871]: clamd daemon 0.101.2 (OS: linux-gnu, ARCH: x86_64, CPU: x86_64)
May 23 13:47:06 gs619-g001189 clamd[21871]: Running as user clamscan (UID 982, GID 973)
May 23 13:47:06 gs619-g001189 clamd[21871]: Log file size limited to 1048576 bytes.
May 23 13:47:06 gs619-g001189 clamd[21871]: Reading databases from /var/lib/clamav
May 23 13:47:06 gs619-g001189 clamd[21871]: Not loading PUA signatures.
May 23 13:47:06 gs619-g001189 clamd[21871]: Bytecode: Security mode set to “TrustSigned”.
May 23 13:47:14 gs619-g001189 clamd: — Stopped at Thu May 23 13:47:14 2019
May 23 13:47:14 gs619-g001189 clamd: Socket file removed.
May 23 13:47:14 gs619-g001189 clamd[4507]: — Stopped at Thu May 23 13:47:14 2019
May 23 13:47:14 gs619-g001189 clamd[4507]: Socket file removed.
In particular, note that last PID of 4507 – it’s the same PID that had been running since the first `ps`. The other PID kept changing (most recently 21871) as it kept failing on the socket file and restarting (under a new PID)
I suspect the problem is rooted in doing both starts as directed in the article:
[root@centos7 system]# systemctl start clamd.service
[root@centos7 system]# systemctl start clamd@scan.service
If so, that makes the explicit `start` of the `clamd@scan.service` superfluous.
4:48 PM
Hi, thank you so much for this great article! I have the following question:
How would I configure clamav to automatically kick off a weekly virus scan of the entire hard drive?
Thanks in advanced!
1:30 PM
That would be via a cron job or timer.
5:17 PM
Hello ~ thank you for this great article.
Now that I have all 3 services up and running, is it possible for me to create an automated weekly scan of the full hard drive (or at least the important areas of the drive)?
Thank you!
1:30 PM
Hi Jason. Sure, you could run the clamscan utility to do a full system scan. Add it to your cronjobs (or timers).
10:02 PM
The freshclam works fine, but when I follow the instructions above here for clamd, it won’t start. When I call the status with: systemctl status clamd.service I see the following:
|root@centos72.testdomain.com > /usr/lib/systemd/system |->systemctl status clamd.service
● clamd.service – clamd scanner daemon
Loaded: loaded (/usr/lib/systemd/system/clamd.service; enabled; vendor preset: disabled)
Active: failed (Result: start-limit) since Sun 2016-02-14 21:52:10 CET; 6min ago
Process: 16999 ExecStart=/usr/sbin/clamd -c /etc/clamd.d/clamd.conf –nofork=yes (code=exited, status=1/FAILURE)
Main PID: 16999 (code=exited, status=1/FAILURE)
Feb 14 21:52:10 centos72.testdomain.com systemd[1]: Unit clamd.service entered failed state.
Feb 14 21:52:10 centos72.testdomain.com systemd[1]: clamd.service failed.
Feb 14 21:52:10 centos72.testdomain.com systemd[1]: clamd.service holdoff time over, scheduling restart.
Feb 14 21:52:10 centos72.testdomain.com systemd[1]: start request repeated too quickly for clamd.service
Feb 14 21:52:10 centos72.rope-parkstad.nl systemd[1]: Failed to start clamd scanner daemon.
Feb 14 21:52:10 centos72.rope-parkstad.nl systemd[1]: Unit clamd.service entered failed state.
Feb 14 21:52:10 centos72.rope-parkstad.nl systemd[1]: clamd.service failed.
Can someone tell me what is going wrong here? I Get the same error when i want to start clamd@scan
1:45 AM
Hi Robert, was the problem fixed? any experience to share?
6:23 PM
Anyone? Anyone?
10:32 PM
Was there a fix to this?
1:48 PM
Please notice that you don’t any extra chart in:
/etc/clamd.d/clamd.conf
I have in the start “/” , after I removed and systemctl restart clamd.service it’s worked
1:34 AM
Robert,
Try running the command alone…
$ ExecStart=/usr/sbin/clamd -c /etc/clamd.d/clamd.conf –nofork=yes
In my case, the response was “ERROR: Please define server type (local and/or TCP).”
So I enabled/uncommented “TCPSocket 3310” in /etc/clamd.d/clamd.conf and it now works.
Will
6:39 AM
Hi Michael. This post includes a big error, the service is not missing and didn’t need to be created. Take a look here, it’s solved my problems:
https://www.adminsys.ch/2015/08/21/installing-clamav-epel-centosred-hat-7-nightmare/
5:07 PM
Thanks, that might be helpful for others. Not everyone will do their installation the same way, so keeping things up for readers to consider what option they prefer.
9:11 AM
update your blog post, no need to remove or change clamd@.service is the service template file, it means you can run few different config files. thats why it has “%i” inside. you name config and you start the service. for example 1 – /etc/clam.d/monitor.conf -> service clamd@monitor start
example 2 – /etc/clam.d/mailscan.conf -> service clamd@mailscan start
first one will monitor folders with on-access feature, and run as root user, the second will be using together with milter socket to scan incoming e-mails, runnind as clamscan user. and so on…
thnks
p.s. http://blog.clamav.net/2016/03/configuring-on-access-scanning-in-clamav.html
1:43 PM
Thanks for sharing
12:39 PM
Why exactly are creating a freshclam daemon?
The default clamav package creates a CRON file (namely “/etc/cron.d/clamav-update”) that calls “/usr/share/clamav/freshclam-sleep” every three hours. Creating an additional freshclam daemon by hand seems to be a bit redundant to me.
9:40 AM
replace “–nofork=yes” with “–foreground=yes”.
https://bugzilla.redhat.com/show_bug.cgi?id=1364253
3:42 AM
The blog converted your double dashes to long dashes / single em dashes.
I understand that you meant to say replace the “dash dash nofork=yes” with “dash dash foreground=yes”
That automatic conversion of your typed characters by the CMS caught me for a second… :-)
That fixed my problem. Thanks for the note…
9:59 AM
Hi,
Thx for the great article! It really helped me get everything up&running.
One thing though. there are no databasemirrors in my country apparently, so, as advised by the Clamav tutorial (http://www.clamav.net/documents/mirrors), I added
DatabaseMirror db.be.clamav.net
DatabaseMirror db.local.clamav.net
This different from what the default contents of the freshclam file says
cfr.
# database.clamav.net is a round-robin record which points to our most
# reliable mirrors. It’s used as a fall back in case db.XY.clamav.net is
# not working. DO NOT TOUCH the following line unless you know what you
# are doing.
DatabaseMirror database.clamav.net
I guess they updated the round-robin record, but failed to update the freshclam.conf default ?
Best Regards,
Sven
2:21 PM
Thanks for sharing that insight. Might be the case, yes.
11:50 AM
Thats a great guide there but it needs to be updated. after a recent update for clamav it no longer supports the flag
‘–nofork=yes’. It was replaced by ‘–foreground=yes’. This will need to be done like this: navigate to /usr/lib/systemd/system/clamd.service : in there replace ‘/usr/sbin/clamd -c /etc/clamd.d/clamd.conf –nofork=yes’ with ‘/usr/sbin/clamd -c /etc/clamd.d/clamd.conf –foreground=yes’ . then run ‘systemctl daemon-reload’ and start clamd
4:49 PM
Thanks for the feedback
7:33 PM
freshclam shouldn’t be run as a ‘service’ in systemd but as a timer. By default it comes with a cron job that will run it automatically as well. rpm -ql clamav-update for details.
4:45 PM
I’m getting this error now:
systemd[1]: /usr/lib/systemd/system/clamd@scan.service:1: .include directives are deprecated, and support for them will be removed in a future version of systemd. Please use drop-in files instead.
8:33 PM
Why are you doing this
cp /usr/share/clamav/template/clamd.conf /etc/clamd.d/clamd.conf
sed -i ‘/^Example/d’ /etc/clamd.d/clamd.conf
but starting the service with another config file (scan.conf)
systemctl start clamd@scan.service
10:53 PM
Thanks a lot for this tutorial. This solved my problem of clamd not starting.
6:03 PM
Hey Sebastian, did you get it to work by disabling clamd.service?
systemctl enable clamd@scan.service
systemctl disable calmd.service