metareplace - can't find component in unit

cnurbw_at_yahoo.com
Date: 02/16/05 

Date: 16 Feb 2005 10:59:44 -0800

Hi,
 I have a Sun 3500 running Solaris 5.8 and DiskSuite 4.2.1 .
I had a bad submirror ( d24 ) , so I detached the submirrors,
brought the system down, and replaced the disk. I brought the system
backup, reformatted the drive, and then tried the "metareplace"
command. But
I get an error "can't find component unit" . Please see the output
below.

First is the metastat before disk replacement ( disk c3t7 was replaced
) , then
are commands I tried , then a new ( abbreviated ) metastat. Now the
former
submirror slices show up as Concat/Stripe and it doesn't show them as
being a submirror as before.

Any ideas ?
Thanks,
Carl

Initial ( abbreviated to show c3t7 disk ) metastat before disk
replacement.

d24: Submirror of d22
    State: Needs maintenance
    Invoke: metareplace d22 c3t7d0s4 <new device>
    Size: 14687190 blocks
    Stripe 0:
        Device Start Block Dbase State Hot Spare
        c3t6d0s0 0 No Okay
    Stripe 1:
        Device Start Block Dbase State Hot Spare
        c3t7d0s4 0 No Maintenance

d30: Submirror of d28
    State: Okay
    Size: 2100735 blocks
    Stripe 0:
        Device Start Block Dbase State Hot Spare
        c3t7d0s0 0 No Okay

d33: Submirror of d31
    State: Okay
    Size: 1019844 blocks
    Stripe 0:
        Device Start Block Dbase State Hot Spare
        c3t7d0s1 0 No Okay

d36: Submirror of d34
    State: Okay
    Size: 8252118 blocks
    Stripe 0:
        Device Start Block Dbase State Hot Spare
        c3t7d0s3 0 No Okay
    Stripe 1:
        Device Start Block Dbase State Hot Spare
        c3t7d0s5 0 No Okay

System to single user

metadb -d -f /dev/dsk/c3t5d0s7

# metadetach -f d22 d24
/sbus@2,0/SUNW,socal@d,10000/sf@1,0 (sf1):
        Forcing lip
d22: submirror d24 is detached
# metadetach -f d28 d30
d28: submirror d30 is detached
# metadetach -f d31 d33
d31: submirror d33 is detached
# metadetach -f d34 d36
d34: submirror d36 is detached

Powered down and replaced disk.

root@ [/] 154# metadb -a -c 1 c3t7d0s7
root@ [/] 155#

root@ [/] 155# metareplace -e d22 c3t7d0s4
metareplace: pasteur: d22: c3t7d0s4: can't find component in unit

metastat

d24: Concat/Stripe
    Size: 14687190 blocks
    Stripe 0:
        Device Start Block Dbase
        c3t6d0s0 0 No
    Stripe 1:
        Device Start Block Dbase
        c3t7d0s4 0 No

d30: Concat/Stripe
    Size: 2100735 blocks
    Stripe 0:
        Device Start Block Dbase
        c3t7d0s0 0 No

d33: Concat/Stripe
    Size: 1019844 blocks
    Stripe 0:
        Device Start Block Dbase
        c3t7d0s1 0 No

d36: Concat/Stripe
    Size: 8252118 blocks
    Stripe 0:
        Device Start Block Dbase
        c3t7d0s3 0 No
    Stripe 1:
        Device Start Block Dbase
        c3t7d0s5 0 No

Relevant Pages

  • Re: metareplace - cant find component in unit
    ... and that didn't complain. ... sense is that the disk shows up in a metastat as ... Concat/Stripe but I can't make it a submirror. ...
    (comp.unix.solaris)
  • Re: Older StorageWorks Parts Not Available
    ... > Who did the parts replacement - CS or the Customer? ... you do also not simply swap bad drives with good drives ... The parts were DOA. ... Disk drives are designed to survive certain ...
    (comp.os.vms)
  • Re: Code density and performance?
    ... > unsatisfactory experience with traditional paging. ... MVS had two separate issues ... more efficient to select non-changed page for replacement (before ... writing a non-changed page out to disk (since the previous copy was ...
    (comp.arch)
  • Live Upgrade and metadisks
    ... I detach 2nd disk submirros. ... d0: Mirror ... Submirror 0: d10 ... Stripe 0: ...
    (comp.unix.solaris)
  • Re: The zoaran curse returns
    ... I never used to do it much, but since my last replacement I ... but it seems like the logic board fairy is visiting me ... a repair from someone else's logic board failure). ... >> First I have to decide whether I should take out the 80GB disk I put in ...
    (uk.comp.sys.mac)