Technical Knowledge Base

Parsing SOS Reports

How to parse an SOS Report (RedHat or RHEL) Support Log.

  1. Download the SOS report.
  2. Extract it to a Directory on your local machine. (Note, you will need to extract it twice, there is a zip inside a zip)
  3. Navigate to the sos_reports directory in the extracted SOS report.
  4. Inside this directory is a file called sos.html

    This can be viewed with any web browser.
    It will mostly contain information relevant to the OS, but there are some key files for hardware information.

  5. Open the sos.html file in the browser of your choice and use Ctrl+F to search for any of the following key words.
            1. dmidecode – cannot stress enough this is your friend. It contains all the pieces/parts of the server. Specifically, and especially helpful in finding DIMM info and locations as called out by the system’s SEL. (Uses same location naming scheme)
            2. ipmitool – this will get you to the server’s SEL (System Event Log) self-explanatory.
            3. filesys – this will contain a few links, the 2 most telling are the /proc/mounts and /etc/fstab
            4. dmesg – anything that is written to stdout winds up here, faults, errors, even when a disk is replaced.
            5. lspci – this will give you info on every device seen on the PCI buses. (Some will be integrated or embedded chips on System Board, but you can see the bus IDs and match them to error entries, etc.)
            6. blockdev – another great command to find HDD info.
            7. lsblk – the lsblk -t output will show all devices and the partitions they contain in a neat table format.
            8. fdisk -l - This one shows you the hdd size and any partitions on it. (and that is an “L” not an ”I”)
            9. lvm2 – Logical Volume Manager this will get you the info on any LVM controlled devices and their associated configuration. (Just follow the links)
            10. md – md stands for Meta Device – this is the new standard for disk mirroring in RHEL and other Linux Distributions. On any Linux machine a quick cat of /proc/mdstat will show if meta devices exist and if so, what underlying disk partitions belong to them It also shows their health status.

There you have top 10 things to check on Linux for SuperMicro Calls. (These commands do exist on any Linux server, so try them out, they can get you valuable info.)