TL;DR - What are you running as a means of “antivirus” on Linux servers?
I have a few small Debian 12 servers running my services and would like to enhance my security posture. Some services are exposed to the internet and I’ve done quite a few things to protect the services and the hosts. When it comes to “antivirus”, I was looking at ClamAV as it seemed to be the most recommended. However, when I read the documentation, it stated that the recommended RAM was at least 2-4 gigs. Some of my servers have more power than other but some do not meet this requirement. The lower powered hosts are rpi3s and some Lenovo tinys.
When I searched for alternatives, I came across rkhunter and chrootkit, but they seem to no longer be maintained as their latest release was several years ago.
If possible, I’d like to run the same software across all my servers for simplicity and uniformity.
If you have a similar setup, what are you running? Any other recommendations?
P.S. if you are of the mindset that Linux doesn’t need this kind of protection then fine, that’s your belief, not mine. So please just skip this post.
I also use ClamAV, but only in specific circumstances, such as when a Linux server will be hosting end-user files. Perhaps a SAMBA server with a file share, or a web server which accepts user uploads.
In those cases, I might want to have it monitor the relevant part of the disk, but I also need to make sure my web application won’t fall over when the file it just accepted is unceremoniously ripped away from it. You can test that out using the EICAR file as your payload.
On a jump box, I might also have it turned on for scanning user home directories, by including /home, and then excluding any home directories for applications and daemons which might not deal well with having their IOPS nuked or delayed.
I am not a expert in Linux, and I mostly rely on very strong passwords. I also discovered recently basic stuff like changing the default SSH port. Anyone knows of implementation of 2FA on Linux?
If at all possible, do not expose things like ssh, RDP, etc to the internet. Use traditional VPN or something like tail scale. Just because ssh is on a different port than normal doesnt mean an attacker couldn’t figure out that your running ssh on port 335.
AFAIK this is not what happens on NixOS. Every package gets installed into a directory that’s a hash of its dependencies in the nix store, but there’s no special isolation or anything on NixOS (well, when the packages are built there’s some isolation, but that’s mostly to keep the builds honest). That said, NixOS is a little better than most distros about creating separate daemon users for services with different permissions, but I don’t think it’s done universally. I love NixOS and it has many benefits, but I don’t think this is one.
That’s what modern endpoint security is, really. Traditional AV is dead. There are far too many people making malware for file signatures or heuristics to keep up. Instead, you want to look for behavior on the system and on the network. For example, if a program starts reading every file it can find on the network, and changing then from their current formats to unreadable blobs, that’s probably ransomware and should be stopped. Plain old AV probably won’t catch it on the client because of how frequently it gets modified (plus all the various evasion techniques), nor on the server because nothing unusual is running on the server.
Okay, I think we can wrap this up: OP started with “I don’t want to be convinced of the predominant oppinion about security” and kept their word.
OP: You got your answer. There is no alternative to ClamAV. ClamAV is open source so it will always be slower than apt update in fixing vulnerabilities.
You can wonder why the whole community that created tons and tons of cool shit for Linux with armies of talented people with way more IT knowledge than all of us combined didn’t dedicate their time to Viruses. You can ask yourself how a virus would even get on your server… or you can not. Your choice. But the answer is: There is no alternative to ClamAV and ClamAV is set up mainly to detect Windows-Viruses that get spread by Mail-Attachments and the like.
There are none.
ClamAV is the only one there is, because it has a very specific and narrow purpose.
There are no viruses for Linux.
Chrootkit and rkhunter are also built for very specific things (detecting rootkits - or making them) and are not designed to protect, they are designed to analyse.
My writing here also isn’t specifically to OP, but to all others that may find this thread - Anti Virus for Linux is BS and unless you are running SMB and still have lots of Windows in your network, it’s absolutely not needed, especially if you follow the basics (like not doing stuff as root, using sudo and not giving out any system rights).
I would recommend just setting up iptables & crowdsec. Open only the ports your services need, and add the relevant plugins to crowdsec. Nothing should come through.
If you have services that allow people to upload files, that’s a different story.
To be honest, antivirus software is just not really a security tool. If you’re at the point where malicious software is running on your server you’ve already lost and it’s hard to know what extent the damage will be. Having proper isolation is much more important (something which, tbh, Linux isn’t quite as great at as we’d like to think, at least not with additional effort… mobile operating systems seem to take the isolation of applications a lot more seriously). You could maybe argue that the anti virus software is useful for monitoring, but I’d rather have some stronger guarantees that my application isn’t going to take my lunch money and private keys than a notice a day later that something sketchy is on my machine… I won’t flat out say a virus scanner is completely useless, because of course you can contrive of scenarios where one could be helpful, but they’re kind of dubious.
Also yeah, ClamAV afaik isn’t really used like a typical windows antivirus. It’s mostly used on mail servers to scan email attachments. It’s not necessarily even looking for “Linux viruses”.
Okay sure same thing as Windows. If you aren’t reckless with the things you install and run then you are likely fine BUT there’s always a chance. All it takes is one slip up. Same logic as having a lock in the door knob and a deadbolt. By your logic (and many others), the lock on the door knob is sufficient and that may be okay with you BUT I’m gonna put a deadbolt on too just in case.
We can argue about this all day long. You will have valid points and so will I.
But would you put a deadbolt on your garage door? Or on your fridge door? IMO, arguing by analogy here just obfuscates the points – your servers aren’t physical doorways with locks, and comparing them just confuses the issue.
Can you explain what added security an antivirus package would offer for a Linux server? I haven’t done much with Linux administration, mostly just using Docker images for stuff at work.
I’m not a super Linux expert or anything, but I do grok tech, and I’m curious about this topic.
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !selfhosted@lemmy.world
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don’t control.
Rules:
Be civil: we’re here to support and learn from one another. Insults won’t be tolerated. Flame wars are frowned upon.
No spam posting.
Posts have to be centered around self-hosting. There are other communities for discussing hardware or home computing. If it’s not obvious why your post topic revolves around selfhosting, please include details to make it clear.
Don’t duplicate the full text of your blog or github here. Just post the link for folks to click.
Submission headline should match the article title (don’t cherry-pick information from the title to fit your agenda).
I also use ClamAV, but only in specific circumstances, such as when a Linux server will be hosting end-user files. Perhaps a SAMBA server with a file share, or a web server which accepts user uploads.
In those cases, I might want to have it monitor the relevant part of the disk, but I also need to make sure my web application won’t fall over when the file it just accepted is unceremoniously ripped away from it. You can test that out using the EICAR file as your payload.
On a jump box, I might also have it turned on for scanning user home directories, by including /home, and then excluding any home directories for applications and daemons which might not deal well with having their IOPS nuked or delayed.
I am not a expert in Linux, and I mostly rely on very strong passwords. I also discovered recently basic stuff like changing the default SSH port. Anyone knows of implementation of 2FA on Linux?
removed by mod
Yes I do have fail2ban. Do you mean I could have just (example) a yubikey and no ssh password? As safe as they can be, why remove the other factor?
removed by mod
If at all possible, do not expose things like ssh, RDP, etc to the internet. Use traditional VPN or something like tail scale. Just because ssh is on a different port than normal doesnt mean an attacker couldn’t figure out that your running ssh on port 335.
Well fail2ban went from very active to very quiet. It is definitely worth not leaving 22 (when opening ssh is a must for different reasons)
Honestly, the best antivirus for Linux is Arch.
removed by mod
AFAIK this is not what happens on NixOS. Every package gets installed into a directory that’s a hash of its dependencies in the nix store, but there’s no special isolation or anything on NixOS (well, when the packages are built there’s some isolation, but that’s mostly to keep the builds honest). That said, NixOS is a little better than most distros about creating separate daemon users for services with different permissions, but I don’t think it’s done universally. I love NixOS and it has many benefits, but I don’t think this is one.
Hahahahaha this actually made me chuckle. Thanks for that!
Its a rolling release, so will always have the most up to date and patched packages the fastest. That concept is the antivirus.
Can’t infect your machine if the vulnerabilities are already fixed.
Unfortunately you can’t easily patch the fleshy thing operating the system
Not yet…
deleted by creator
Maybe you’d be interested in https://wazuh.com/
Not sure this is what I’m looking for as it appears to be an XDR SIEM vendor.
That’s what modern endpoint security is, really. Traditional AV is dead. There are far too many people making malware for file signatures or heuristics to keep up. Instead, you want to look for behavior on the system and on the network. For example, if a program starts reading every file it can find on the network, and changing then from their current formats to unreadable blobs, that’s probably ransomware and should be stopped. Plain old AV probably won’t catch it on the client because of how frequently it gets modified (plus all the various evasion techniques), nor on the server because nothing unusual is running on the server.
Okay, I think we can wrap this up: OP started with “I don’t want to be convinced of the predominant oppinion about security” and kept their word.
OP: You got your answer. There is no alternative to ClamAV. ClamAV is open source so it will always be slower than apt update in fixing vulnerabilities.
You can wonder why the whole community that created tons and tons of cool shit for Linux with armies of talented people with way more IT knowledge than all of us combined didn’t dedicate their time to Viruses. You can ask yourself how a virus would even get on your server… or you can not. Your choice. But the answer is: There is no alternative to ClamAV and ClamAV is set up mainly to detect Windows-Viruses that get spread by Mail-Attachments and the like.
I could fork ClamAV and call it OysterAV then there would be a less maintained alternative
There are none. ClamAV is the only one there is, because it has a very specific and narrow purpose. There are no viruses for Linux.
Chrootkit and rkhunter are also built for very specific things (detecting rootkits - or making them) and are not designed to protect, they are designed to analyse.
My writing here also isn’t specifically to OP, but to all others that may find this thread - Anti Virus for Linux is BS and unless you are running SMB and still have lots of Windows in your network, it’s absolutely not needed, especially if you follow the basics (like not doing stuff as root, using sudo and not giving out any system rights).
I would recommend just setting up iptables & crowdsec. Open only the ports your services need, and add the relevant plugins to crowdsec. Nothing should come through.
If you have services that allow people to upload files, that’s a different story.
I am running SMB although it’s not publicly available and setup with specific users having specific access to specific shares.
Good note on crowdsec
removed by mod
To be honest, antivirus software is just not really a security tool. If you’re at the point where malicious software is running on your server you’ve already lost and it’s hard to know what extent the damage will be. Having proper isolation is much more important (something which, tbh, Linux isn’t quite as great at as we’d like to think, at least not with additional effort… mobile operating systems seem to take the isolation of applications a lot more seriously). You could maybe argue that the anti virus software is useful for monitoring, but I’d rather have some stronger guarantees that my application isn’t going to take my lunch money and private keys than a notice a day later that something sketchy is on my machine… I won’t flat out say a virus scanner is completely useless, because of course you can contrive of scenarios where one could be helpful, but they’re kind of dubious.
Also yeah, ClamAV afaik isn’t really used like a typical windows antivirus. It’s mostly used on mail servers to scan email attachments. It’s not necessarily even looking for “Linux viruses”.
Okay sure same thing as Windows. If you aren’t reckless with the things you install and run then you are likely fine BUT there’s always a chance. All it takes is one slip up. Same logic as having a lock in the door knob and a deadbolt. By your logic (and many others), the lock on the door knob is sufficient and that may be okay with you BUT I’m gonna put a deadbolt on too just in case.
We can argue about this all day long. You will have valid points and so will I.
removed by mod
Okay cool. Thanks.
But would you put a deadbolt on your garage door? Or on your fridge door? IMO, arguing by analogy here just obfuscates the points – your servers aren’t physical doorways with locks, and comparing them just confuses the issue.
Can you explain what added security an antivirus package would offer for a Linux server? I haven’t done much with Linux administration, mostly just using Docker images for stuff at work.
I’m not a super Linux expert or anything, but I do grok tech, and I’m curious about this topic.
How does it obfuscate the point? A layered approach to security.