I originally posted this in the thread begun 7/8/2005 by Andrew, but the
discussion there seemed to indicate that everybody thought this was no
longer an issue, and I suspect not many are looking at that thread

We also set up an OES lab server on an HP/Compaq with smart array
controller. We made our way past the initial YAST problem with the
modified OES disk 1.

However, we also received the fsck error signal while trying to reboot
after modifying fstab to boot using EVMS. We then tried the manual
console commands Andrew referred to
from http://evms.sourceforge.net/faq.html without success:

mount -o remount,rw / executed ok
/sbin/evms_activate caused a string of console error
messages as
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Warning.. fsck.reiserfs for device /dev/evms/ida/c0d0p1 exited with
signal 6.
fsck.reiserfs /dev/evms/ida/c0d0p1 failed (status 0x8) Run manually!
[blah, blah, other bootup successes]
fsck failed for at least filesystem ( not /).
Please repair manually and reboot.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

I can get the server back to booting normally by modifying the second
line in /etc/fstab
/dev/evms/ida/c0d0p1 /boot reiserfs acl/user_xattr 1 2
back to
/dev/ida/c0d0p1 /boot reiserfs acl,user_xattr 1 2

but, of course, that means EVMS is not really doing the deed for us on

Our OES lab install is running on a Compaq ML370 server with a smart
controller. Since nearly all our servers on campus are Compaq/HP running
raid 5
arrays, utilizing a different type machine is not an option. Any help
would be deeply
appreciated, since we can't consider migrating to OES without creating
NSS volumes,
and of course that can't happen until we get EVMS working properly. My
thus far suggests I could arrange a work-around by dividing the hardware
array into
two logical volumes, thus making OES and EVMS think I have two separate
disks, but
we'd prefer not to do that if EVMS can actually work on a raid 5 array
with one logical volume.

Dale McPherson
Eastern Virginia Medical School (another EVMS)
Norfolk, Va.

andrew@cstgroup.nospam.com.au wrote:

> Looks like the symptoms are at least somewhat of a know issue,
> http://evms.sourceforge.net/faq.html
> The text is listed below, my system reported an unknow device

> but has booted correctly now. if i find out anymore details I will

> them here.
> Andrew
> <snip>
> System Startup
> Q. I rebooted my system, and during boot-up it complains about

not being
> able to fsck the root volume. The boot-up stops and it drops to a root

> shell prompt. What happened to my root volume?
> A. To get your system back to a normal state, run the following

> from the root shell prompt:
> mount -o remount,rw /
> /sbin/evms_activate
> Then reboot your system once more and it should boot normally.
> The reason for this behavior is due to the order in which

> devices are activated. DM assigns device-numbers dynamically as

devices are
> activated. If you made a change to your volume configuration, it's

> that it changed the order in which EVMS discovers and activates all

the DM
> devices. However, the device nodes in /dev/evms/ represent the
> device-number assignments from the previous boot. Since evms_activate
> usually doesn't run until after the root-filesystem is remounted
> read-write, the device node for your root volume may have the wrong
> device-number. Thus the error when it tries to fsck the root volume.
> One of the best solutions to this problem is to use udev to manage

> /dev directory, and to mount a tmpfs on /dev. With this type of setup,

> can actually have your boot scripts run evms_activate before the root
> filesystem is fsck'ed and remounted read-write. This ensures that the
> /dev/evms/ directory is correctly up-to-date with the current DM
> device-number assignments before the root volume needs to be fsck'ed.
> </snip>
> >
> > Greetings All,
> >
> > I have had a read of the recent post "EVMS not handling boot

> > and think my problems may be similar but dont seem identical so I

> > post it seperatly.
> >
> > I have followed the instructions in the OES "installation Guide for
> > Linux" and am having trouble with step 13 on the EVMS configuration.

> > I launched YaST and selected System > Run Level, the screen loaded

> > simple editor but the hour glass remains and no further changes can

> > made.
> >
> > Subsequently I could not change the settings for boot.1vm and

> >
> > Upon rebooting, ( aftering sucessfully configuring several nss pools

> > volumes) the server fails to load correctly. It reports "fsck

failed" and
> > suggested that I can use the "mount -n -o remount, rw /" to remount

> > volume as prompted by the console messages.
> >
> > I can do this and modify files in /etc/
> >
> > Is there a way to manually modify these run levels, in init.d

> >
> > Any other thoughts or suggestion appriciated.
> >
> > Regards, Andrew