Boot from degraded RADI1 fails on RC1
Hi, I just did a test on a boot from a RAID1, works well. Then I degraded the RAID1 by taking out one disk, the system fails to boot. Reattached disk and it booted fine. Took out the other disk in the raid, boot fails. Reattached disk and it booted fine. The Yast boot loader program tells me that the boot loader is installed to both disks and that "Enable redundancy for md array" is checked. Any suggestions to where I start to look? Kind Regards Birger -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Birger Kollstrand wrote:
Hi,
I just did a test on a boot from a RAID1, works well.
Then I degraded the RAID1 by taking out one disk, the system fails to boot. Reattached disk and it booted fine.
Took out the other disk in the raid, boot fails. Reattached disk and it booted fine.
The Yast boot loader program tells me that the boot loader is installed to both disks and that "Enable redundancy for md array" is checked.
Any suggestions to where I start to look?
Kind Regards Birger
If it is a software raid (that is if there is something in /proc/mdstat) then the cause is probably that mdadm needs the special "--run" parameter to assemble(start) a degraded array. This would probably only need one or two extra lines in /lib/mkinitrd/scripts/boot-md.sh but that should better not be all left to end-users => file a bug? -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Yes , it's on a SW raid.
I'll take your advice and file a bug.
2009/10/18 Bernhard M. Wiedemann
Birger Kollstrand wrote:
Hi,
I just did a test on a boot from a RAID1, works well.
Then I degraded the RAID1 by taking out one disk, the system fails to boot. Reattached disk and it booted fine.
Took out the other disk in the raid, boot fails. Reattached disk and it booted fine.
The Yast boot loader program tells me that the boot loader is installed to both disks and that "Enable redundancy for md array" is checked.
Any suggestions to where I start to look?
Kind Regards Birger
If it is a software raid (that is if there is something in /proc/mdstat) then the cause is probably that mdadm needs the special "--run" parameter to assemble(start) a degraded array. This would probably only need one or two extra lines in /lib/mkinitrd/scripts/boot-md.sh but that should better not be all left to end-users => file a bug? -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
-- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Warning: A longish rant follows. One of my test machines is an Mtech laptop with an AMD 450/2 CPU and 256 MB RAM. It now has openSUSE 11.1 on it, but does not run any GUI desktop for obvious reasons. As this machine does not have a DVD drive, I am unable to use that method of installation. Furthermore, Live CDs require an i686 - the AMD 450/2 has the i586 instruction set. As a result, installation that way is not possible. I had attempted to test 11.2 M8 on the machine using the NET install mini-CD and discovered that it too required an i686, even though it was labeled "i586". A bug for that was filed, and the NET install disk for 11.2 RC1 does boot on an i586. When I tried to install using the CD, installation crashes because 256 MB is not enough memory. As a result, there is no way to install 11.2 on this computer! It is possible that I may be able to upgrade by changing the repos and doing a 'zypper dup', but I will not try that until the GM is out. The reason for this rant is to let the Testing Team know that 11.2 is leaving low-end machines behind, even in server applications. Is this something we want to discuss in our next meeting? Larry -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
2009/10/29 Larry Finger
Warning: A longish rant follows. The reason for this rant is to let the Testing Team know that 11.2 is leaving low-end machines behind, even in server applications. Is this something we want to discuss in our next meeting?
For me this is beyond the HW that I think should be supported by the standard distro for new desktop/server GUI release. What maybe could be an idea is if there could be made a "weak HW" distro with very low memory requirements, but with the newest kernel. This distro could then also be used on embedded devices and maybe controlled with the webYast thing. cu -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
participants (3)
-
Bernhard M. Wiedemann
-
Birger Kollstrand
-
Larry Finger