SUMMARY: using metareplace to replace failed disk -Reply

From: Dennis Martens (
Date: Sat Nov 07 1998 - 19:10:55 CST

Got 3 good replies from:

    Lee Boon Heong <>
    Rajeev Kumar <>

Who confirmed my thought that the 'good' slice on the failed disk had to
be hotswapped out before the disk was replaced.

Thanks for the replies.

Original post follows:

>>> Dennis Martens <> 30/October/1998 11:10am

   I have checked the archives and could not find the exact syntax that
I am looking for. Hope someone can help.

   E5000 running Solaris 2.5.1 using DiskSuite 4.0

   Checks of log files show that a slice in a RAID5 set has errored, and
was replaced by a hot spare. This will be replaced when the engineer
arrives. I am just double checking the commands to resynch the disks
BACK after the physical replacement.

  # metastat d15

d15: RAID
    State: Okay
    Hot spare pool: hsp002
    Interlace: 16 blocks
    Size: 4142000 blocks
Original device:
    Size: 4143168 blocks
        Device Start Block Dbase State Hot Spare
    c1t1d1s7 170 No Okay
    c1t3d1s7 170 No Okay
    c1t5d1s7 170 No Okay c1t1d3s7

c1t5d1 is a 2.1Gb disk, split into 2 slices. Only 1 slice has errored
and switched to a hot spare.

If I replace the disk with the bad slice, what happens to the 'good'
slice on it - will that automatically switch to a hot spare as well?

And then, after the physical replacement of the disk, would I use these

     metareplace -e d15 c1t5d1s7
     metareplace -e d20 c1t5d1s6

I am not sure of the above syntax - for the 'component', do I specify
the failed slice as I have done? (according to the man page this slice
will be resynched and the hot spare slice will be returned neatly to the
hot spare pool - is this correct??)

Thanks and I will summarise.

Dennis Martens
unix admin.
queensland health, brisbane, australia

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:52 CDT