I have set up a mirrored root and swap in a Logical Domain by setting
up the configuration during a JumpStart process. I'm using Solaris 10
8/07 and Logical Domains 1.0.1. The manual says that you can use
SVM from within a domain, and I am using two flat files that the
I/O domain provides to the logical domain as two separate disks.
The mirroring seems to work according to metastat, and all seems
fine from reboot to reboot except for the following warnings at
boot time:
NOTICE: mddb: unable to get devid for 'vdc', 0x7
NOTICE: mddb: unable to get devid for 'vdc', 0x7
NOTICE: mddb: unable to get devid for 'vdc', 0x7
NOTICE: mddb: unable to get devid for 'vdc', 0xf
NOTICE: mddb: unable to get devid for 'vdc', 0xf
NOTICE: mddb: unable to get devid for 'vdc', 0xf
Then I try to simulate a disk failure. I shut down the domain
normally, then remove the second disk in the mirror using
the following command:
ldm remove-vdisk vdisk1 guest4
Then I boot the domain. It recognizes that the disk is missing
but seems to have inconsistent database replicas:
WARNING: md: d12: /dev/dsk/c0d1s0 needs maintenance
Hostname: s4
Insufficient metadevice database replicas located.
Use metadb to delete databases which are broken.
Ignore any Read-only file system error messages.
Reboot the system when finished to reload the metadevice database.
After reboot, repair any broken database replicas which were deleted.
Apr 11 12:23:19 svc.startd[7]: svc:/system/metainit:default: Method
"/lib/svc/method/svc-metainit" failed with exit status 96.
Apr 11 12:23:19 svc.startd[7]: system/metainit:default misconfigured:
transitioned to maintenance (see 'svcs -xv' for details)
Apr 11 12:23:22 svc.startd[7]: svc:/system/filesystem/usr:default: Method
"/lib/svc/method/fs-usr" failed with exit status 95.
Apr 11 12:23:22 svc.startd[7]: system/filesystem/usr:default failed fatally:
transitioned to maintenance (see 'svcs -xv' for details)
Requesting System Maintenance Mode
(See /lib/svc/share/README for more information.)
Console login service(s) cannot run
Root password for system maintenance (control-d to bypass):
When I look at what metadb has to say, it appears that the surviving
disk only has one valid replica:
# metadb
flags first blk block count
a m p lu r 16 8192 /dev/dsk/c0d0s7
a p l r 8208 8192 /dev/dsk/c0d0s7
a p l r 16400 8192 /dev/dsk/c0d0s7
M p luo r 16 unknown /dev/dsk/c0d1s7
M p luo r 8208 unknown /dev/dsk/c0d1s7
M p luo r 16400 unknown /dev/dsk/c0d1s7
Does anybody have a clue here as to what's going on? I'm suspecting
that this is an SVM problem, but not sure at all so I'm copying the
LDoms discussion as well.
--
This message was posted from opensolaris.org
