Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I have a hosting client who is reporting back to me that he is seeing port scan attempts from the IP of the server I host his companies' website on. I'm unaware of what program he is using, but apparently the company's primary IP is being port scanned over port 443 (SSL).

The log snippets he sent over match up with the SSL access log of the site, so is this a fault monitoring program, or is my server pinging their IP back during SSL access?

share|improve this question
    
Um...any more details on what your server is running outside of normal system processes? Your own logfiles with their IP? Something? –  Bart Silverstrim Aug 9 '10 at 12:40
    
who is he? and why is your server pinging? –  StampedeXV Aug 9 '10 at 12:40
4  
I don't understand what you mean by port scan over a single port. By definition a port scan is over a range of ports. –  John Gardeniers Aug 9 '10 at 12:58
1  
Are these 'scan' attempts coming to 443, or from 443? Is it possible that he's seeing all the response traffic from your sever showing up as a scan from your 443, to a random high number port? This could be caused by a poor configuration of a non-stateful IDS. –  Christopher Karel Aug 9 '10 at 13:29
    
@ChristopherKarel is correct, see my answer below for a more in depth description of why this happens. –  bahamat Jul 22 '12 at 19:11

5 Answers 5

Port 443 is actually the HTTPS port. It is common for hosts to allow SSH access to port 443 because most company web proxies will allow a transparent connection to be made to this port, and this port only.

What you are seeing is somebody attempting to gain SSH to the box, either through the normal port (22) or through a commonly-used alternative (443).

share|improve this answer
3  
I wouldn't call 443 a "commonly-used" alternative port to SSH... –  MDMarra Aug 9 '10 at 13:08
    
Its not uncommon, at least in my experience. –  Sirex Aug 9 '10 at 14:45
    
i understand what PP is trying to say. +1 –  djangofan Aug 9 '10 at 23:18

A) run a check on your system with ClamAV to see if anything out of the ordinary is detected.

B) Run rootkit checkers to see if they find anything out of the ordinary.

C) Run a packet sniffer (tcpdump, wireshark) on your server in question to see if there's unusual traffic running through it.

D) check /tmp, /var/log, etc. to see if there are any unusual temp files or logfiles deposited there.

E) happen to have any programs like tripwire installed to look for unusual changes to system files?

F) Check with other admins to see if they've been doing something with that server that could trigger a check on the remote system.

What exactly does the log SAY is happening? Requesting a particular file? Just pinging it? Are they running a listening process on 443 that logs what is happening, or do they have a legit process handing requests on that port? You can narrow down a little bit (or eliminate other possibilities) by looking at what is in the request. Also, is your server's IP showing up somewhere else in their logs, like requesting non-SSL web pages or something like that?

Is your system scanning other machines in your network? Is the IP showing up for unusual activity in other server's logs?

You can also set up something like Snort or Honeyd within your network for continued monitoring to see if something's going on. My advice on the clam and rootkit checkers are only half-solutions since if your system IS compromised, only an offline check may actually find malware if the system binaries have been compromised (unless you have another server of the same configuration against which you can run md5 checksums on certain binaries to see if they've been tampered with).

share|improve this answer

you can run a full port scan using nmap (http://nmap.org/). It has a GUI called Zenmap GUI (http://nmap.org/zenmap/).

It's one of the best security scanners out there.

share|improve this answer
    
I thought the problem is the poster has a client reporting that the client's computer is showing a connect attempt on port 443 from a server belonging to the poster, and the poster doesn't know why his server would do this... –  Bart Silverstrim Aug 9 '10 at 13:24

I would bet that the attempts are coming from a web browser.

share|improve this answer

He is misinterpreting what he is seeing in his firewall logs.

Here's what is happening:

  • A connection is made from his side to you, from random port to port 443.
  • Data is passed.
  • Somehow his firewall stops tracking the connection.
  • The session is still active and your side sends a packet that is part of this ongoing connection.
  • His firewall incorrectly sees this as a new connection without a SYN coming from you, source port 443 destination port the same port this connection used originally.

Because his firewall somehow stopped tracking the connection this last packet from you is seen as a new connection, which it isn't.

This is happening a lot to him, possibly hundreds or thousands of times a day/hour/minute. He or his IDS is incorrectly detecting this as a port scan, all with the source port 443.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.