Topic
  • 3 replies
  • Latest Post - ‏2009-07-01T06:22:12Z by SystemAdmin
CAOS
CAOS
29 Posts

Pinned topic Problems with smsupdatenode

‏2009-06-29T11:07:22Z |
Hi,

If I execute a "smsupdatenode -a -i RPM", system hangs and I need to reboot server. However, if I execute a "smsupdatenode -N NodeGrp -i RPM", all process runs OK. Why?

For helping me, I'm sending a file that includes both smsupdatenode.log files: smsupdatenode.log.OK-grupo has been generated in a good execution, while smsupdatenode.log.KO-Todo has benn generated in a bad execution.

Help, please, help!!

Bye!
Updated on 2009-07-01T06:22:12Z at 2009-07-01T06:22:12Z by SystemAdmin
  • SystemAdmin
    SystemAdmin
    476 Posts

    Re: Problems with smsupdatenode

    ‏2009-06-30T07:24:43Z  
    From the log file attached, I notice that in the process of "smsupdatenode -a", the rpms are installed onto node19.hpc.local first, then dsh command is called for the others:

    11:10:46 Running Command: /opt/csm/bin/dsh -f 32 -n node19.hpc.local "export SMSLOCAL_CALLED_BY_DSH=1; if -f /opt/csm/csmbin/sms_local ; then /opt/csm/csmbin/sms_local -v -d /csminstall/Linux/SLES/10/x86_64/SP1 -i /csminstall/Linux/SLES/10/x86_64/updates/e2fsprogs-1.38....
    ...........................

    11:11:1 Running Command: /opt/csm/bin/dsh -C CSM -O --device-rsh HASH(0x267b150) -E -f 32 --log -o HASH(0x26403c0) -r HASH(0x267b170) -n node21.hpc.local,node32.hpc.local,node31.hpc.local,node02.hpc.local,node04.hpc.local,node10.hpc.local,node16.hpc.local,node06.hpc.local,node11.hpc.local,node22.hpc.local,node05.hpc.local,node17.hpc.local,node26.hpc.local,node29.hpc.local,node09.hpc.local,node28.hpc.local,node24.hpc.local,node18.hpc.local,node03.hpc.local,node13.hpc.local,node07.hpc.local,node12.hpc.local,node15.hpc.local,node01.hpc.local,node30.hpc.local,node23.hpc.local,node27.hpc.local,node33.hpc.local,node20.hpc.local,node14.hpc.local,node08.hpc.local,node25.hpc.local -F --report -S -t "export SMSLOCAL_CALLED_BY_DSH=1; if -f /opt/csm/csmbin/sms_local ; then /opt/csm/csmbin/sms_local -v -d /csminstall/Linux/SLES/10/x86_64/SP1 -i /csminstall/Linux/SLES/10/x86_64/updates/e2fsprogs-1.38-....

    My question is, is there any difference between node19 and the others?
  • CAOS
    CAOS
    29 Posts

    Re: Problems with smsupdatenode

    ‏2009-06-30T08:55:09Z  
    No, node19 is completely equal to node01...node33.
    From node01 to node16, nodegrp is OldNodesProd
    From node02 to node32, nodegrp is NewNodesProd
    Node33 nodegrp is AllNodesDvl

    But 33 machines are completely equal, hardware, software, network...
  • SystemAdmin
    SystemAdmin
    476 Posts

    Re: Problems with smsupdatenode

    ‏2009-07-01T06:22:12Z  
    • CAOS
    • ‏2009-06-30T08:55:09Z
    No, node19 is completely equal to node01...node33.
    From node01 to node16, nodegrp is OldNodesProd
    From node02 to node32, nodegrp is NewNodesProd
    Node33 nodegrp is AllNodesDvl

    But 33 machines are completely equal, hardware, software, network...
    Could you try to "smsupdatenode -N NodegrpAll"? here NodegrpAll is a nodegroup that contains all your 33 nodes.
    Also try "smsupdatenode -a" to see if this problem is related to "-a" flag or a certain node?

    PS, please let me your CSM version so that I can check with the corresponding code.