> A few years ago we did an upgrade from, I believe sles10 to sles10sp1, and > the upgrade automatically changed all our drives to "by disk id" versus this > format /dev/hdxx. Upon reboot our boxes could not find the DASD and they > did the pw000000 lockup stuff. I subsequently went through each server and > changed the formats back - zipl, and fstab. Was it something about the id's > not happy with zvm DASD's, I CAN'T REMEMBER, yes I knew at the time... > Now sles11 upgrade is giving me a dire warning that I need to go back to "by > id" and the only thing that sticks in my head in not the reason, just the all > night changing the stupid configs. Help.
In a word: ignore them. /by-path is still the most reliable method of referencing specific disks on Linux on Z. The problem was that the Linux DASD driver at the time assigned all minidisks on a physical 3390 a by-id string related to the physical 3390 holding the minidisk. If you changed disk units or moved a minidisk, the physical UUID of the disks changed and you lost access to the data because the by-id string no longer matched. They may have fixed that this time, but just in case, you can find the script to fix it at http://download.sinenomine.net/sane-dasd-update/ The trigger case to write that script was that someone got a new disk array, moved their systems to it and tried to bring up their guests -- and none of the Linux filesystems would mount because they were all referencing the old physical platters. The poor folk had to cable up the old disks, copy the data and fix the fstabs en masse. ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For more information on Linux on System z, visit http://wiki.linuxvm.org/
