A rescue images contains the minimal version of all the basic diagnostic
tools, drivers and other utilities that would be useful for checking a previously
existing logical partition.
To use the rescue image that you created on NWSSTG, use the following
steps:
- Disconnect the failed partition's virtual storage space (if applicable)
through WKRNWSSTG.
- Connect your rescue storage space as the first drive to the NWSD
and reconnect the original storage space (where applicable) as the second
drive.
- Edit the NWSD for the failed partition to start from IPL Source
*NWSSTG. Also edit the IPL Parameters Field to reflect the root partition
on the rescue storage space. For most distributions this is a parameter
like 'root=/dev/hda3' or 'root=/dev/vda1' Refer to your Linux documentation
for assistance or look at other running partitions using the command 'cat
/proc/iSeries/mf/B/cmdline'.
- Boot the partition.
- If the existing root partition is on native disk, you may need
to insert the ibmsis driver using the command "insmod ibmsis".
- Create a mount point to which you will mount the root partition
of the network storage space you are trying to rescue. You can
use a command like "mkdir /mnt/rescue".
- Now mount the root partition of the network storage space you are
trying to rescue. You mount a drive using the command "mount
-t your partition's type partition's location mount point" where
the partition's type is the format of the partition like ext2 or reiserfs,
the partition's location is something like /dev/hdb3 (for non-devfs), /dev/hd/disc1/part3
(for devfs server) or /dev/sda2 in the case of a partition on native disk.
It is important to note that the drive you are trying to rescue, when using
virtual disk, will be the second drive rather than the first drive. (i.e.
If the drive was /dev/hda3 when the partition was running normally, it will
be /dev/hdb3 in the rescue server.) Here again, your Linux documentation or
the configuration files you created when you created the rescue NWSSTG will
help you determine the device for the root of the partition you are trying
to rescue. Finally, your mount point will be something like /mnt/rescue if
you use the example from above.
After following the steps listed above, you can either use the rescue
tools provided in your rescue storage space against the mount point you created
or you can change root to that partition using "
chroot mount point"
to work on the partition from within its own storage space. See
Build a rescue image on an NWSSTG for more information.