|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-users] guest boots sucessfully only one time then lvm complains no uuid found [ SOLVED]
There were two problems I had in my configuration:1. When the guest was first installed, it was not set up with raid, ( I added it later ) therefore initrd does not have raid1.ko and the initrd's script init does not load raid1.ko as well as not run mdautorun. So when the system boots /dev/md0 is never detected therefore the VG cannot be loaded. I am not sure if this applies to other distros, in my case, it is RHEL5. I fixed this by adding those changes into initrd.img. 2. I used whole disks /dev/xvdb and /dev/xvdc for raid1. This makes /dev/md0 work only *after* the real partition / is mounted, and then /etc/mdadm.conf is applied. At boot time, before / is mounted, mdautorun does not detect these two devices as raid devices. I fixed this by using partitions with their types FD ( auto raid detect ) instead. This makes /dev/md0 is detected at boot time. I guess I can still use whole disks for raid 1 to be detected at boot time if I change the init script in initrd to rebuild md0. What I still do not understand is why my original guest image works the first time after boot and only that time. It should not work at all. Will check it later, though. :) Anyway, it feels good when this is fixed, and feels stupid when looking back at it :) Thanks, Vu Vu Pham wrote: christoffer@xxxxxxxxx wrote:It should be possible - I do precisely that: LVM on software raid1. I think more detail is needed. What are your device names and what do you export to your domU? Cheers, /ChrisI set up a guest ( under RHEL5 dom0) that has a root volume group extends onto /dev/md which is a raid 1. The first time after setting it up, that guest boots just fine but thenext reboot will panic with the root lvm cannot the uuid of the /dev/md0.Because I have a backup of that image, so I can restore and restart it again but then even without doing anything, the next reboot will panic.I do not have problem when my main VG ( the VG that has the LV containing the / partition ) is built directly on the /dev/md0. I have another guest with that configuration and it runs just fine.This guest with problem has the main VG built just on one drive and it boots normally. Then I build the /dev/md0 on two other drives, and *extend* the main VG into this /dev/md0 's PV : just extend the VG, not yet resize the LV inside it. I did run mdadm --detail --scan > /etc/mdadm.conf.The first time I boot, the LVM can see /dev/md0 and its PV, so the VG is found ok. The 2nd time I boot, it complains no UUID ( for /dev/md0 )is found. All of the below configuation output is from the first boot:Below is my guest config: [root@xen2 ~]# more /etc/xen/g1 name = "g1" uuid = "2731f3ac-3bde-7105-46b9-794004cb3894" maxmem = 1024 memory = 1024 vcpus = 1 bootloader = "/usr/bin/pygrub" on_poweroff = "destroy" on_reboot = "restart" on_crash = "restart" vfb = [ ]disk = [ "tap:aio:/data/xen/images/g1.img,xvda,w", "tap:aio:/data/xen/images/g1_ _______________________________________________ Xen-users mailing list Xen-users@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-users
|
 |
Lists.xenproject.org is hosted with RackSpace, monitoring our |