11-03-2010 03:52 AM
Hi All,
There was one Disk failed in RAID 5 DG. Which was replaced , Current vxprint -htg raid 5 output shown as below.
bash-2.03# vxprint -htg raid5
dg raid5 default default 73000 1019571721.1080.DEV-ADC
dm raid501 - - - - REMOVED
dm raid502 c1t1d0s2 auto 2888 71121402 -
dm raid503 c1t5d0s2 auto 2888 71121402 -
dm raid504 c1t4d0s2 auto 2888 71121402 -
v vol01 - ENABLED ACTIVE 8388672 RAID - raid5
pl vol01-01 vol01 ENABLED ACTIVE 8389632 RAID 4/32 RW
sd raid501-01 vol01-01 raid501 0 2796552 0/0 - RMOV
sd raid504-01 vol01-01 raid504 0 2796552 1/0 c1t4d0 ENA
sd raid503-01 vol01-01 raid503 0 2796552 2/0 c1t5d0 ENA
sd raid502-01 vol01-01 raid502 0 2796552 3/0 c1t1d0 ENA
v vol02 - ENABLED ACTIVE 12582912 RAID - raid5
pl vol02-01 vol02 ENABLED ACTIVE 12584448 RAID 4/32 RW
sd raid501-02 vol02-01 raid501 2796552 4194828 0/0 - RMOV
sd raid504-02 vol02-01 raid504 2796552 4194828 1/0 c1t4d0 ENA
sd raid503-02 vol02-01 raid503 2796552 4194828 2/0 c1t5d0 ENA
sd raid502-02 vol02-01 raid502 2796552 4194828 3/0 c1t1d0 ENA
v vol03 - ENABLED ACTIVE 104857632 RAID - raid5
pl vol03-01 vol03 ENABLED ACTIVE 104861952 RAID 4/32 RW
sd raid501-03 vol03-01 raid501 6991380 34954011 0/0 - RMOV
sd raid504-03 vol03-01 raid504 6991380 34954011 1/0 c1t4d0 ENA
sd raid503-03 vol03-01 raid503 6991380 34954011 2/0 c1t5d0 ENA
sd raid502-03 vol03-01 raid502 6991380 34954011 3/0 c1t1d0 ENA
v vol04 - ENABLED ACTIVE 4194336 RAID - raid5
pl vol04-01 vol04 ENABLED ACTIVE 4194816 RAID 4/32 RW
sd raid501-04 vol04-01 raid501 41945391 1398276 0/0 - RMOV
sd raid504-04 vol04-01 raid504 41945391 1398276 1/0 c1t4d0 ENA
sd raid503-04 vol04-01 raid503 41945391 1398276 2/0 c1t5d0 ENA
sd raid502-04 vol04-01 raid502 41945391 1398276 3/0 c1t1d0 ENA
v vol05 - ENABLED ACTIVE 83333184 RAID - raid5
pl vol05-01 vol05 ENABLED ACTIVE 83333184 RAID 4/32 RW
sd raid501-05 vol05-01 raid501 43343667 27777735 0/0 - RMOV
sd raid504-05 vol05-01 raid504 43343667 27777735 1/0 c1t4d0 ENA
sd raid503-05 vol05-01 raid503 43343667 27777735 2/0 c1t5d0 ENA
sd raid502-05 vol05-01 raid502 43343667 27777735 3/0 c1t1d0 ENA
How to enable newly added disk into existing raid5 subdisk. If i try to delete highligted subdisk will the data get destroyed?
the volumes were started forcefully when the disk was failed using
vxmend -g raid5 fix stale vol03-01
vxmend -g raid5 fix clean vol03-01
vxprint -htg raid5
vxvol -g raid5 -f -o unsafe start vol03
Solved! Go to Solution.
11-03-2010 05:52 AM
Use vxdiskadm and select replace a failed disk option. Deleting subdisk which is in RMOV state ideally should not be allowed by VxVM. I would not do this atleast. You must create log plex while using RAID 5 Volumes to avoid such issues.
11-03-2010 05:52 AM
Use vxdiskadm and select replace a failed disk option. Deleting subdisk which is in RMOV state ideally should not be allowed by VxVM. I would not do this atleast. You must create log plex while using RAID 5 Volumes to avoid such issues.
11-04-2010 03:50 AM
vxdiskadm is safe choice... option 5 would be there to replace the failed/removed disk
same can be achieved manually by command "vxdg -g <diskgroup> -k adddisk <dmname=daname>"
& yes, using log plex would be a strong recommendation in order to add reslilience..
Gaurav
11-04-2010 04:31 AM
It is difficult to try and understand how exactly you managed to get to this state. If I look at Symantec's recommendations, there is not a single point where disks or subdisks end up in REMOVED state...
Please share detailed steps that you have taken to remove disk - did you use vxdiskadm to remove without specifying replacement disk?
Reference:
Removing and replacing disks: http://www.symantec.com/business/support/index?page=content&id=HOWTO22047
Best practice to recover a failed RAID-5 volume: http://www.symantec.com/business/support/index?page=content&id=TECH22735
It might be possible to specify the replacement disk with 'vxdg -k adddisk ....', but right now my recommendation is to rather open a Support call ASAP... A Symantec engineer can via a Webex session determine what the current state is and suggest the best way to recover.
11-06-2010 01:54 AM
-o unsafe option is unsafe to use if log plex is not enabled you may loose your data
11-08-2010 10:13 PM
Hi All,
Thanks for help....
Varun