Sunday, July 1, 2012

Boot error from second disk after mirrored disks on Solaris 5.10 x86 Virtual machine on VM Ware

i'm having some problem with boot from second disk after i've configured two disk in mirror on a virtual machine Solaris 10 x86 . 
it seems the metadevice d20 is corrupt. 
I want to simulate the breaking of the primary disk and start the system from the secondary disk connected as primary on second IDE channel. 
Could you help me? 
The slices on disk are 6 and I built the half 1 = 4 metadb for every disk on slice 7. It's right? the syncing works well but... 
....when the system start with the second disk: 


WARNING: md: d10 (Unavailable) needs maintenance 
Hostname: etc..... 
Insufficient metadevice database replicas located. 
Use metadb ti 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. 
WARNING: md: d15 (Unavailable) needs maintenance 
Apr 2 13:18:31 svc.startd[7]: system/filesystem/usr:default failed fatally: tra 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): 
Apr 2 13:18:31 svc.startd[7]: svc:/network/service:default: Method "/lib/svc/methid/net-svc start" failed with exit status 1. 
Apr 2 13:18:31 svc.startd[7]: svc:/network/service:default: Method "/lib/svc/methid/net-svc start" failed with exit status 1. 
Apr 2 13:18:32 svc.startd[7]: svc:/network/service:default: Method "/lib/svc/methid/net-svc start" failed with exit status 1. 
Apr 2 13:18:32 svc.startd[7]: network/service:default: failed: transitioned to maintenance (see 'svcs -xv' for details) 
Have you any idea ? 

If you create two replicas on each disk in a two-disk configuration, Solaris Volume Manager 
still functions if one disk fails. But because you must have one more than half of the total replicas 
available for the system to reboot, you cannot reboot. 

How many database replicas have you got? 

I beleive the following line in /etc/system gets around this ">51%" quota 

set md:mirrored_root_flag=1 

Obviously you will have to reinstate the simulated failed disk and boot the server / add the above entry to /etc/system then reboot for it to take effect.

Thank you Andi, 
your claim has worked, however, this is the structure of metadb: 
metadb -i 
flags first blk block count 
a p luo 16 8192 /dev/dsk/c0d1s7 
a p luo 8208 8192 /dev/dsk/c0d1s7 
a p luo 16400 8192 /dev/dsk/c0d1s7 
a p luo 24592 8192 /dev/dsk/c0d1s7 
a p luo 32784 8192 /dev/dsk/c0d1s7 
a p luo 40976 8192 /dev/dsk/c0d1s7 
a m p luo 16 8192 /dev/dsk/c0d0s7 
a p luo 8208 8192 /dev/dsk/c0d0s7 
a p luo 16400 8192 /dev/dsk/c0d0s7 
a p luo 24592 8192 /dev/dsk/c0d0s7 
a p luo 32784 8192 /dev/dsk/c0d0s7 
a p luo 40976 8192 /dev/dsk/c0d0s7 

As I said theslices on disk are 6 
cannot paste the output of metastat command because the screen is a virtual machine. But all submirror are ok.
Thank you Andi, 
your claim has worked, however, this is the structure of metadb: 
metadb -i 
flags first blk block count 
a p luo 16 8192 /dev/dsk/c0d1s7 
a p luo 8208 8192 /dev/dsk/c0d1s7 
a p luo 16400 8192 /dev/dsk/c0d1s7 
a p luo 24592 8192 /dev/dsk/c0d1s7 
a p luo 32784 8192 /dev/dsk/c0d1s7 
a p luo 40976 8192 /dev/dsk/c0d1s7 
a m p luo 16 8192 /dev/dsk/c0d0s7 
a p luo 8208 8192 /dev/dsk/c0d0s7 
a p luo 16400 8192 /dev/dsk/c0d0s7 
a p luo 24592 8192 /dev/dsk/c0d0s7 
a p luo 32784 8192 /dev/dsk/c0d0s7 
a p luo 40976 8192 /dev/dsk/c0d0s7 

As I said theslices on disk are 6 
cannot paste the output of metastat command because the screen is a virtual machine. But all submirror are ok.
the Andi's claim has worked, however, this is the structure of metadb: 
metadb -i 
flags first blk block count 
a p luo 16 8192 /dev/dsk/c0d1s7 
a p luo 8208 8192 /dev/dsk/c0d1s7 
a p luo 16400 8192 /dev/dsk/c0d1s7 
a p luo 24592 8192 /dev/dsk/c0d1s7 
a p luo 32784 8192 /dev/dsk/c0d1s7 
a p luo 40976 8192 /dev/dsk/c0d1s7 
a m p luo 16 8192 /dev/dsk/c0d0s7 
a p luo 8208 8192 /dev/dsk/c0d0s7 
a p luo 16400 8192 /dev/dsk/c0d0s7 
a p luo 24592 8192 /dev/dsk/c0d0s7 
a p luo 32784 8192 /dev/dsk/c0d0s7 
a p luo 40976 8192 /dev/dsk/c0d0s7 

As I said the slices on disk are 6 
cannot paste the output of metastat command because the screen is a virtual machine. But all submirror are ok.
Excuse me Andi, You can provide a simulator to learn the creation and administration LUN? 
Know also links to documentation on this?
Maybe you don't have change boot device location in /etc/system
It looks like your metadb on second disk are having some error , you have to 
delete all metadb on second disk and boot. 
once server is up you should create the matadbs on 2nd disk. 
Please let me know if you have any concerns. 

0 comments:

Post a Comment

 
Design by BABU | Dedicated to grandfather | welcome to BABU-UNIX-FORUM