[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] unclean file system

>   We got some mail saying that RAM disks were not the way to go.  So we
> re-installed, made several partitions to use, and rebuilt xen.  However,
> we're still having trouble mounting partitions when xeno-linux boots.
> We give a physical grant of one of our newly made paritions (/dev/hda14 in
> particular) to the domain.  But when we start a xenlinux domain we get
> error messages saying that "Your systems appears to have shut down
> uncleanly".  It then proceeds to check the root file system and runs fsck
> on it, which goes fine.  It tries to activate swap but it can't seem to
> find /dev/hda3.  These seem odd but not critical.

You need to allocate a swap partition per domain, and grant each
domain physical access to its particular swap partition. The
alternative is simply not to configure the guest OS to use swap -- we
commonly operate in this mode with no problems.

Once we have reimplemented 'virtual disc' support over our new
unstable toolset, it will then be possible to very easily allocate a
'virtual swap device' to new guest OSes without any need to
repartition your hard disc or worry about limitations on the number of
partitions supported by a disc. This is still work in progress, however.

> Then it tries to do fsck on /dev/hda1 (boot) and we get a error message
> saying "The superblock could not be read or does not describe a correct
> ext2 filesystem".  It does this for all filesystems in the fstab file.
> For the one we give a physical grant (/dev/hda14, mounted on /xen1), it
> goes into a seemingly infinate loop where we see the message "/xen1:
> recovering journal" over and over unitl it runs out of file descriptors.
> As usual we'll take any suggestions.  We think we might need to unmount
> the filesystems from the domain0 before letting other Xenolinuxes get a
> hold of them, but that doesn't explain where our other errors come from.

The only filesystem that both DOM0 and the new domain should
simultaneously have mounted is /usr, which they should both have
mounted read-only. Any other configuration (e.g., DOM0 still having
hda14 mounted) is rather unlikely to work as both OSes will fight over
the superblock.

 -- Keir

This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/
Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.