Solaris

I have a Netra T5220 Solaris 10 server with LDoms installed and enabled. RAID1 is enabled in the control domain with the following slices:

d50 -m d51 d52 1
d51 1 1 c1t0d0s5
d52 1 1 c1t1d0s5
d10 -m d11 d12 1
d11 1 1 c1t0d0s0
d12 1 1 c1t1d0s0
d20 -m d21 d22 1
d21 1 1 c1t0d0s1
d22 1 1 c1t1d0s1
d40 -m d41 d42 1
d41 1 1 c1t0d0s4
d42 1 1 c1t1d0s4
d60 -m d61 d62 1
d61 1 1 c1t0d0s6
d62 1 1 c1t1d0s6
d70 -m d71 d72 1
d71 1 1 c1t0d0s7
d72 1 1 c1t1d0s7

Each slice has a guest domain built. In other words, when the guest domain is built, something similar to /dev/md/dsk/d30 is assigned; the commands look like:

# ldm add-vdsdev /dev/md/dsk/d70 vol7@primary-vds0
# ldm add-vdisk vdisk1 vol7@primary-vds0 ldg1

My question is whether there is an issue with having LDoms and RAID1 at the same time? The reason I am asking is, after I am done with the control domain and guest domains cofguration, and metast looks good, once I reboot the control domain I get an error:

Hostname: control_dom
/dev/md/rdsk/d40 is clean
/dev/md/rdsk/d70 is clean
checking ufs filesystems
/dev/md/rdsk/d50: PARTIALLY ALLOCATED INODE I=2
/dev/md/rdsk/d50: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
/dev/md/rdsk/d60: NEGATIVE SIZE -5476372287601639394 I=2
/dev/md/rdsk/d60: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
THE FOLLOWING FILE SYSTEM(S) HAD AN UNEXPECTED INCONSISTENCY: /dev/md/rdsk/d50 (/exp ort/home/sgw), /dev/md/rdsk/d60 (/export/home/mrf)
fsckall failed with exit code 1.
WARNING - Unable to repair one or more filesystems.
Run fsck manually (fsck filesystem...).
mount: Please run fsck and try again
mount: Please run fsck and try again
svc:/system/filesystem/local:default: WARNING: /sbin/mountall -l failed: exit status 1
Reading ZFS config: done.
Feb 4 09:52:41 svc.startd[7]: svc:/system/filesystem/local:default: Method "/lib/svc/method/fs-local" failed with exit status 95.
Feb 4 09:52:42 svc.startd[7]: system/filesystem/local:default failed fatally: transitioned to maintenance (see 'svcs -xv' for details)

After that, ldm commands returns "Failed to connect to logical domain manager: Connection refused". Running fsck does not help; it even makes thing worse. Appreciate your help...

Hi,

before y reboot the control domain, stop the guest domains

I tend to agree with you but my understanding from Sun documentation is that guest domain should not be affected when the control domain is rebooted or down. The only thing is that we won't be able to manage the guest domain when the control domain is not available. I noticed though when the control domain is down, guest domain response becomes very slow (to typing commands, etc.)!

hi,

no, if y have only 1 io/service/control-domain and y reboot this domain, all guest domain lost his virtuell devices (vdisk, vnet ... ). With 1 io/service/control-domain y must shutdown the guest domain before y reboot the control domain.

Y want active guest domains, if y reboot the control domain. Then y need for example T5240 with 2 BUS controler and 2 io-domains.

I did stop all the guest domains (their status was in "bound" state); and got the same error! It is very time consuming cause I have to re-install the whole thing again.

{0} ok boot
Boot device: /pci@0/pci@0/pci@2/scsi@0/disk@0,0:a File and args:
SunOS Release 5.10 Version Generic_138888-03 64-bit
Copyright 1983-2008 Sun Microsystems, Inc. All rights reserved.
Use is subject to license terms.
Hostname: js-client_t5220
/dev/md/rdsk/d40 is clean
/dev/md/rdsk/d70 is clean
checking ufs filesystems
/dev/md/rdsk/d50: PARTIALLY ALLOCATED INODE I=2
/dev/md/rdsk/d50: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
/dev/md/rdsk/d60: PARTIALLY ALLOCATED INODE I=2
/dev/md/rdsk/d60: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.

THE FOLLOWING FILE SYSTEM(S) HAD AN UNEXPECTED INCONSISTENCY: /dev/md/rdsk/d50 (/export/home/sgw), /dev /md/rdsk/d60 (/export/home/mrf)
fsckall failed with exit code 1.

WARNING - Unable to repair one or more filesystems.
Run fsck manually (fsck filesystem...).

mount: Please run fsck and try again
mount: Please run fsck and try again
svc:/system/filesystem/local:default: WARNING: /sbin/mountall -l failed: exit status 1
Reading ZFS config: done.
Feb 6 09:29:11 svc.startd[7]: svc:/system/filesystem/local:default: Method "/lib/svc/method/fs-local" failed with exit status 95.
Feb 6 09:29:11 svc.startd[7]: system/filesystem/local:default failed fatally: transitioned to maintena nce (see 'svcs -xv' for details)

js-client_t5220 console login: root
Password:
Last login: Thu Feb 5 14:59:40 on console
Sun Microsystems Inc. SunOS 5.10 Generic January 2005
You have new mail.
root.js-client_t5220# ldm list
Failed to connect to logical domain manager: Connection refused

is js-client_t5220 the guest ldom? you are rebooting the control domain and the control domain is failing to come online?

only thing i can think of is stop the guest domains, tear down d50, d60, boot off cdrom single user, fsck the rdsk's of d50 and d60, run this several times to see if it repairs, mount root fs, remove md's d50 and d60 from vfstab and then reboot. see if you can get a stable os. make sure the guest ldoms don't come up because they will obviously be looking for the metadevices.

i'm not sure about ldom's because i haven't had the pleasure of working with them as of yet but i'm just thinking of this as if a stand alone sol10 install wasn't working after a reboot with metadevices.

hi,

can y show me y /etc/vfstab and `ldm list-bindings -p`

and why y have mount d50 (/export/home/sgw) or d60 (/export/home/mrf)?