Results 1 to 13 of 13
  1. #1
    Join Date
    May 2006
    Posts
    64

    Server keeps crashing every few days

    One of my servers at FDC Servers is crashing (= kernel panic) every few days since we got it.

    The specs of the machine are:
    C2D E6550, 2 GB

    Linux .. 2.6.24.4-64.fc8 #1 SMP Sat Mar 29 09:15:49 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux

    I opened several tickets with FDC, after running memtest and clocking down memory they say it's not a hardware issue but software related.

    I see that according to this ticket, memtest has been run without error and the memory was even slightly clocked down to avoid problems. It appears that the errors you are reporting are software related...
    I already tried several kernels and different application versions.

    The server is only running lighttpd.

  2. #2
    Join Date
    Apr 2005
    Location
    Cochin
    Posts
    2,446
    post some more logs from dmesg and var/log/messages so that someone can shed some light on it
    Sam
    Supportlobby.com - Expertise in Windows/Linux server support, IaaS and PaaS Management
    Spiralbean.com - Custom Software Development
    Email: exec @ activelobby.net | Skype: unni_active

  3. #3
    Not sure what distro you are running but look for a package called "netdump". It's available for RHEL and HP-UX, and others. It will allow you to dump the contents of kernel memory over the network to another system when a kernel panic occurs. You can later read the contents of the dump file to help troubleshoot what is happening.

    From experience, I highly doubt lighttpd is the culprit. Typically random kernel panics are caused by bad system memory, system board, or processor.
    ServiceFlex - High Performance Web Solutions

  4. #4
    Join Date
    May 2006
    Posts
    64
    Quote Originally Posted by serviceflex View Post
    Typically random kernel panics are caused by bad system memory, system board, or processor.
    same here .. but fdc wont believe it

    netdump is a great idea. i'll install it.. after the machine comes back up :/

  5. #5
    Join Date
    May 2006
    Posts
    64
    hmm no netdump in fedora repositories?

  6. #6
    ServiceFlex - High Performance Web Solutions

  7. #7
    Join Date
    Jul 2007
    Location
    Dallas, TX
    Posts
    320
    +1 on memory. I'd start there.

    Good luck,
    Joe
    https://ServersAndHosting.com
    Dedicated Servers | Guaranteed Hosting

  8. #8
    Join Date
    May 2006
    Posts
    64
    netdump running. based on what i saw in the logs before its nothing that points to a single application or kernel issue.

    is there something like memtest that i can run on a server without ipkvm ?

  9. #9
    Join Date
    Jun 2004
    Location
    Oregon
    Posts
    1,282

  10. #10
    Join Date
    May 2006
    Posts
    64
    looks like this is gonna be a never ending story.

    after countless crashes and manual reboots from FDC I asked for the memory to be swapped... but the problem still persists

    http://img.techpowerup.org/080428/Capture153.gif
    how fast can you swap memory on your machine?

  11. #11
    Join Date
    Apr 2002
    Location
    chicago
    Posts
    717
    Quote Originally Posted by W1zzard View Post

    http://img.techpowerup.org/080428/Capture153.gif
    how fast can you swap memory on your machine?
    if you ever visit FDC datacenter you`ll see that hardware swaps can be done very quickly. Most of our dedicated servers do not have side covers. Techs can just walk up to the machine, power it down and swap the ram. It can be done pretty much under 1 minute

    If the ram swap does not fix your problem I recommend that you request your whole server to be replaced. That way you can eliminate hardware side of the server crashing problem.

    Feel free to PM your ticket number if you have any problems getting your server replaced
    Petr Kral - FDCServers.net LLC
    Los Angeles - Seattle - Denver - Chicago - Miami- New York - Paris - London - Madrid - Amsterdam - Frankfurt - Vienna - DATACENTERS
    10+Tbps of internet connectivity --- 10G, 20G, 40Gbps unmetered services in 17 US & EU locations

  12. #12
    Join Date
    May 2006
    Posts
    64
    thanks peter. the server has been replaced a few hours ago. everything stable so far *knock on wood*

  13. #13
    Join Date
    Jun 2007
    Posts
    99
    I had a box at FDC that was crashing every 60 min, was a bad grub entry

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •