Fedora Core 5 anaconda no md raid

necrophage · 06-07-2006, 03:34 PM

Hi
i am trying to upgrade fedora core 4 on a machine with 2 37 gb sata drives in md raid 0.
here is dmesg in the operational fc4 environment.when i pop in the fc install disk it does not detect my existing linux installation. the weird thing is that during the upgrade from fc 3 to fc4, the fc4 disk picked up the existing install and upgraded it

md: raid0 personality registered as nr 2
md: Autodetecting RAID arrays.
md: autorun ...
md: considering sdb1 ...
md: adding sdb1 ...
md: adding sda2 ...
md: created md0
md: bind<sda2>
md: bind<sdb1>
md: running: <sdb1><sda2>
md0: setting max_sectors to 512, segment boundary to 131071
raid0: looking at sdb1
raid0: comparing sdb1(36145920) with sdb1(36145920)
raid0: END
raid0: ==> UNIQUE
raid0: 1 zones
raid0: looking at sda2
raid0: comparing sda2(36041728) with sdb1(36145920)
raid0: NOT EQUAL
raid0: comparing sda2(36041728) with sda2(36041728)
raid0: END
raid0: ==> UNIQUE
raid0: 2 zones
raid0: FINAL 2 zones
raid0: zone 1
raid0: checking sda2 ... nope.
raid0: checking sdb1 ... contained as device 0
(36145920) is smallest!.
raid0: zone->nb_dev: 1, size: 104192
raid0: current zone offset: 36145920
raid0: done.
raid0 : md_size is 72187648 blocks.
raid0 : conf->hash_spacing is 72083456 blocks.
raid0 : nb_zone is 2.
raid0 : Allocating 16 bytes for hash.
md: ... autorun DONE.

i am attaching fdisk and /etc/fstab

Device Boot Start End Blocks Id System
/dev/sda1 * 1 13 104391 83 Linux
/dev/sda2 14 4500 36041827+ fd Linux raid autodetect

Device Boot Start End Blocks Id System
/dev/sdb1 * 1 4500 36146218+ fd Linux raid autodetect

|||||||||||||||||||||||FSTAB|||||||||||||||||||||

/dev/VolGroup00/LogVol00 / ext3 defaults 1 1
LABEL=/boot /boot ext3 defaults 1 2
none /dev/pts devpts gid=5,mode=620 0 0
none /dev/shm tmpfs defaults 0 0
none /proc proc defaults 0 0
none /sys sysfs defaults 0 0
192.168.0.1:/home /home nfs soft 0 0
/dev/VolGroup00/Swap swap swap defaults 0 0

i would like to upgrade with the CD and not using yum.
thanks

WhatsHisName · 06-07-2006, 06:43 PM

Try updating /etc/mdadm.conf and see if that makes a difference:

Code:

# cd /etc
# cp mdadm.conf mdadm.conf.`date +%y%m%d`
# echo "DEVICE partitions" > mdadm.conf
# echo "MAILADDR root" >> mdadm.conf
# mdadm -D -s >> mdadm.conf

necrophage · 06-09-2006, 11:56 AM

i tried that but mdadm -D -s didnt work [forgot the error] then i manually wrote in all the info from my fc4 environment...UUID and everything into the mdadm.conf file and $mdadm -A /dev/md0
i then checked dmesg and it worked...but anaconda had an error while [trying to search for previously installed versions] this md0 is in VolGroup00 and has both the / and the swap partitions using lvm...i am basically a gentoo user and have never used lvm...so i dunno what to do with that...if we can mount the logical volume maybe that will work

WhatsHisName · 06-09-2006, 12:17 PM

The mdadm.conf update was a long shot that I didn’t expect to work.

I’ve run into problems with Anaconda not recognizing installations in raids and lvms before, but it usually has to do with the partitions containing the raids and lvms not having the correct filesystem types listed. You don’t seem to have that problem, so I don’t know what’s causing your problem.

The only thing I notice is that your raid0 partitions are size mismatched, but seem to be within the 1% allowance. So that doesn’t seem to be the problem either, unless Anaconda is very picky about it.

I’ve never been a fan of upgrading by way of media and prefer using yum. When it works, a yum upgrade will usually do a better job.

Getting the yum upgrade to go is usually the problem, because of unresolved (or unresolvable) dependencies. It’s usually faster to do a full backup and then do a fresh install.

WhatsHisName · 06-09-2006, 12:34 PM

Forgot to add, if you go the fresh install route and want to keep raid0 functionality, then look into setting up a scattered lvm. I find that the scattered lvm is slightly faster than a mdadm raid0.

Look at the lvcreate man page under the -i and -I options. There is even an example at the end of the man page showing you how to set up a scattered logical volume. You will need to manually configure the lvm before the installation if you want to go this route.

The nice thing about using scattered logical volumes is that you can revert to linear logical volumes at a later date if the raid0 functionality makes you nervous one day. You can also use linear and scattered logical volumes in the same volume group.