I find that with SLES10, you do NOT have to mkinitrd and zipl. It will happily find the additional DASD and bring them online without needing special intelligence in the INITRD.
The whole INITRD thing ... I will hold my peace for the moment. Now if you added a different KIND of DASD (if you added SAN to prev all CKD or if you added CKD to prev all SAN or if you added FBA (for VDISK)) then you might have to stamp a new INITRD. But even there, as long as your root FS is alive and kicking the other disk types might come to life using modules pulled from there. (I have not tried.) The purpose of INITRD is to have enough smarts to get the root mounted. (And probably /usr and one or two others.) But with /lib/modules present and available, additional kernel modules can be read from there. -- R; <>< On Wed, Feb 18, 2009 at 1:58 PM, Ivan Warren <i...@vmfacility.fr> wrote: > Mark Post wrote: >> >> You need to chroot to /mnt/sysimage before running the mkinitrd. >> >> >> > Sorry to hop on route.. but.. > > I was wondering... > > Why on earth does one need to mkinitrd/zipl after adding a DASD volume > to z/Linux ? > > Most (if not all) 'distributed' platforms are perfectly happy to add a > volume to the configuration without having to rebuild the initial > ramdisk and re-writing the boot record. I can open my workstation, add a > SATA disk, fdisk it, format it, change /etc/fstab - and done.. no need > for any additional trickeries.. > > In a similar fashion, on other IBM z operating systems, adding a volume > doesn't require any form of 'rebuild' (not in their modern day versions > anyway).. I can add a z/VM or z/OS volume without having to rebuild the > nucleus or updating the IPL record. (z/VM requires me to add some info > in SYSTEM CONFIG - but that's far fetched from having to rebuild > anything.. I don't even need to re-run SAPL). > > So it's not something to do with Linux.. It's not something to do with > the architecture.. so *WHY* does it have to be this way for Linux on z ? > > For all I know, all the dasd_xxx modules have automagic probe capabilities. > > And the initrd was never designed for anything else (from what I > understand) to just help out in bootstrapping the operating system > initialization process (by loading those essential modules) so that you > could start a system with a modular architecture (vs monolithic) to at > least mount the essential filesystems : basically the root fs.. and then > perform *real* initialization).. > > Other operating systems use that scheme.. AIX does.. but I never had to > do a 'bosboot' because I added an FCP LUN, SAS drive or whatnot..I only > had to do that if I change the boot device.. but that's different. > > Note : > > I'm not blaming anyone (I don't play the blame game).. Not the linux > folks (Because if I want it differently *then* I can do it myself and > submit a patch to LKML), nor the Novell people (if this is specific to > SLES - and someone decided for whatever reason this is how it ought to > be - OR - maybe they're already looking into it) - and especially not > Mark (who is probably the most helpful person I've seen here). > > Honestly just wondering.. Because I looked.. and can't find any good > reason ! > > --Ivan > > ---------------------------------------------------------------------- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
