Home › Forums › OS X Server and Client Discussion › Questions and Answers › 10.4.11 Server Soft-Locking every few minutes

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • April 8, 2009 at 8:31 pm #375962
    TimBloom
    Participant

    We have a G5 Xserve here that is becoming unresponsive every few minutes, and booting some people from file shares, or not allowing them to connect to different services. I’ve restarted and looked through the logs but nothing jumps out at me as the cause of the problem.

    When I first showed up today atprintd was going crazy and using all free CPU, so I disabled that since it’s hardly used anyway, which freed up the cpu and rebooted.. 30 minutes later it began acting the same without atprintd hogging the cpu. I’m able to use the GUI and the window server looks as though it’s responsive, but buttons won’t press, some apps get the spinning wheel of death. It seems as though it’s waiting for something to time out. I suspect it’s something with authentication because many hangs are when I’m trying to sudo a command in the terminal. I’m quite perplexed as there’s not much in the logs as to why this is happening.

    I used the opportunity of the reboot to apply the new hostname to match our reverse dns entry, but I suspect the issue isn’t related to that since it was around before I changed it with changeip.

    One thing that is especially slow is Server Admin, if that is a clue. Sometime’s it’s refused to load info on services at all or gets disconnected from the server.

    Any thoughts on where to look since the console and system logs are of little help?

    April 9, 2009 at 12:49 am #375965
    chadpilkington
    Participant

    I am having the same issues with Server Admin. I have 3 10.4.11 servers doing it and it just started 2 days ago. It almost seems like anything that requires any sort of authentication is pausing (ftp, pop3, ssh). We have made no server changes and the only hardware in our entire network was we have swapped out a flaky switch 6 days ago. I have changed that switch for another just to rule it out as well. Webserving works well still. WebObjects apps are running as expected.

    We have 2 network cards in each of the servers one for the public network and one for the private although the servers do not act as bridges. The public ip address of the servers correctly reverse dns map. The private ips do not map. It has been that way for years however with no problems.

    Our problem is not happening all the time. It was an issue for about 8 hours the past 2 days but has been fine by the time I left work and was bad again when I returned. I have tried removing all the other computers from the private network but no change.

    April 9, 2009 at 1:14 am #375966
    deemery
    Participant

    Well, this may not help but: I have Leopard Server on a G5 desktop, and that machine is very flakey. It seems to be particularly sensitive to RAM, at times it’s reported errors associated with bad RAM. I also have TimeMachine running on an external drive (not a 10.4.x problem, I know) and large TimeMachine backup sets seem to be susceptible to disk corruption.

    If nothing else, you might want to look for memory diagnostics; even better if you can run something like AppleJack in standalone mode to remove bad RAM as a culprit.

    dave

    April 9, 2009 at 2:15 am #375968
    chadpilkington
    Participant

    That may help Tim but to have ram go bad in 3 Servers at once seems a little far fetched. We upgraded the ram in one of the servers a month or two ago but the other two servers have not had their ram touched.

    I will give it a try anyway just to rule it out.

    January 13, 2010 at 9:11 pm #377812
    TimBloom
    Participant

    Well, I was looking around on here for a post I made a couple months ago, and saw that I had forgotten about this post. I wanted to post the resolution I had for this.

    I was able to track down that the source of this was an intrusion attempt. Something from outside the network seemed to be hitting the SSH authentication system pretty hard with a brute-force attack. I blocked SSH at the Cisco ASA they have, and disabled SSH on the server. After that the server seems to have straightened up.

    January 13, 2010 at 9:20 pm #377813
    chadpilkington
    Participant

    That is exactly what was happening to us. However, we need ssh open (sftp) so I was forced to configure the firewall to only allow ssh to specific external IPs which is not an optimal solution.

  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)
  • You must be logged in to reply to this topic.

Comments are closed

Copyright © 2025 AFP548. All Rights Reserved.