AIX Open Source

 View Only
  • 1.  system dump device configuration related issue

    Posted 15 days ago

    Hi Team,

    We are getting issues when system getting rebooted automatically but dump is not generated successfully. Reason could be space issue or some other issue.
    We tried to check one server with command sysdumpdev -e where estimated dump size was huge. But on that server primary dump device was not that big but still dump generated successfully. What is the criteria to set the size of dump device. Can you help and suggest. What is the best recommended approach to set the system dump. We have to set this on across the environment. 

    Example of few servers given below and we have same issue almost across the board

    Dump_Device_Name REQUIRED_SIZE AVAILABLE_SIZE Status

    hd6 1685627862 41943040 Warning: dump device size is not insufficient

    hd6 1302226534 4194304 Warning: dump device size is not insufficient

    hd6 6869619834 4194304 Warning: dump device size is not insufficient



    ------------------------------
    Manoj Kumar
    ------------------------------


  • 2.  RE: system dump device configuration related issue

    Posted 14 days ago

    Take a look at this technote, as it may help you with this task: https://www.ibm.com/support/pages/managing-system-dump-devices#Determining_size



    ------------------------------
    Chris Gibson
    ------------------------------



  • 3.  RE: system dump device configuration related issue

    Posted 5 days ago

    Thank you Chris for this technote.  I have review this and will follow the same. Currently we are using hd6 paging space as primary dump device so we need to move into dedicated dump device.

    Have a question on that.

    I hope the below settings are the best recommended one for system dump? How much free space do we need in /var file system to copy the system dump? 

    and I hope the calculation for creation of system dump for larger LPAR (Big Memory size) is the same

    primary              /dev/lg_dumplv
    secondary            /dev/sysdumpnull
    copy directory       /var/adm/ras
    forced copy flag     FALSE
    always allow dump    TRUE
    dump compression     ON
    type of dump         fw-assisted
    full memory dump     disallow



    ------------------------------
    Manoj Kumar
    ------------------------------



  • 4.  RE: system dump device configuration related issue

    Posted 2 days ago
    Edited by Chris Gibson 2 days ago

    Yes, that dump device config looks OK.

    For fw-assisted dump, the space in /var is not initially that important, only the size of the dump device. When the system completes a successfully fw-assisted dump, the image is saved to the dump device (for example /dev/lg_dumplv). Unlike traditional dumps, or systems configured to dump to paging space (hd6), the dump is not automatically written to /var. So, to extract the dump image, after a successful dump, you run the savecore command. This command will save the system dump files to the directory that you specify. Meaning as long as you specify a directory/fs with sufficient free space, it'll be successful.

    To estimate how large the dump might be, it is a good idea to run the following commands on the systems when they are busy:

    sysdumpdev -e

    /usr/lib/ras/dumpcheck -p

    This can help ensure your dump device is large enough to hold a potential dump and give you an idea of how large the dump could be.

    HTH.



    ------------------------------
    Chris Gibson
    ------------------------------



  • 5.  RE: system dump device configuration related issue

    Posted yesterday

    Thank you. Really appreciate for the help.



    ------------------------------
    Manoj Kumar
    ------------------------------