problems accessing server's

Discussion in 'Technical' started by Torsson, Apr 18, 2010.

  1. Torsson

    Torsson Member

    Hi, i have 2 identical servers both with ubuntu and Xen to host virtual servers on.. and on the both dom0 i have drbd, nfs and heartbeat to get files replicated between them both..

    And now to the problem
    Both the servers stops working sometimes and refuse to do anything, but i can still ping them both. but can't access any port like http, ssh, ftp. and i have checked syslog and kernel log but can not se anything that tells me that something went wrong.

    The both servers is have now the same problem and i can't write anything from the logs.

    Next weekend im gona go to the place where they are hosted (450 km from here). and it would be helpfull if i could get any hints about what the problem could be.

    i have a gut feeling that this has something about drbd/nfs to do becaus one time when i was moving 30-40 gb to it both died in the same way.

    Any tips are welcome.
  2. falko

    falko Super Moderator Howtoforge Staff

  3. Mosquito

    Mosquito New Member

    Are you using a bridge with static IPs to the Virtual hosts or a bridge with DHCP assigned IPs?

    I had this problem on a CentOS 5.4 server with Xen when I was using DHCP assigned IPs. Switched to static (outside of my DHCP range, but still within the subnet) and everything started working fine.
  4. Torsson

    Torsson Member

    Falko, ill try that

    Mosquito, all the virtual servers and xen host's have static ip address

    Saw something strange today on the servers.. tried to access ssh but it dosen't answer anyting.. not "No route to host" or "Refused". so the machine seems to be alive still.
    Last edited: Apr 19, 2010
  5. Torsson

    Torsson Member

    This is the only strange thing i find in the server log. Can this be the reason that the server is always dieing?

  6. Torsson

    Torsson Member

    I think i have found the problem.

    I installed the same version of Xen (3.3) on Ubuntu 9.10, same as the server and i get exacly the same problem as on the servers.

    the problem is that when the kernel is booting it just get black screen when it is starting/been online for a few hours. i can ping the server but nothing else.

    im using kernel: linux-image-2.6.24-16-xen_2.6.24-16.30zng1_i386

    So im gona try to install Xen 4.0.0 with a newer kernel and se if that works
  7. edge

    edge Active Member Moderator

    When they are down are you trying to access the server(s) by domain or by IP?
    If by domain it could be a DNS isue.
  8. Rapid2214

    Rapid2214 New Member

    Also do they go down at the same time?
  9. Torsson

    Torsson Member

    Now this night i got the same problems. both died almost at the same time. between 6-7 am. the second server died first and everyting was pointed to the first server.

    Yes i try to ssh directly to the ip and not to the dns
  10. falko

    falko Super Moderator Howtoforge Staff

    Maybe this is caused by a cron job? Maybe by one of the scripts in /etc/cron.daily/?
  11. Torsson

    Torsson Member


    i think you have right, it haves to be some crontab that is messed up

    when i gave it a hard think i realized that i have tryed to move everything to the servers 2 time

    Both times night between Thursday and friday
    And both servers died 06:00-07:00

    And this has happend twice, so it needs to be cron.

    ill let you know when i have tryed that!
  12. Torsson

    Torsson Member

    i have now stopped cron.daily and have created a script that creates 40 gb files of different size, and then i tar everything and untar it, then it starts over, so now i just need to wait and se what happends :)
  13. Torsson

    Torsson Member

    I stopped everything in cron.daily, hourly, weekly and monthly and everything seemd to work ok. but this morning server 2 died. and then i looked on server 1 and found out that motd-update and vnstat vas in cron.d folder and stopped it on the server 1 and it has not died yet.

    So i started digging in the logs on server1 for information from the resent crash and found out these things. and after these things the logs dosent say anything untill manually reboot. it would be nice if someone had the time to look at the logs and se if you can se why it dies

    Kernel log:
    Messages log:
  14. Torsson

    Torsson Member

    And now it seems that server 1 is dead, when i try to ssh the xen virt server on server 1 it asks for the password but when type the password nothing happends and trying to ssh server 1 it dosent do anything, not timeout or something like that.
  15. Torsson

    Torsson Member

    I have been looking around on Xen lists and i have noticed other pepole with similar problems. and it seems to be a bug in the xen kernel on DELL R200-R300 machines. so i think i need to fix that or upgrade to 4.0 or something

Share This Page