Re: metadevice

From: Mr. Johan Andersson (johan_at_solace.mh.se)
Date: 09/05/03 

Date: Fri, 5 Sep 2003 16:34:59 +0200

On Thu, 4 Sep 2003, Per Knutsen wrote:

>
> "Scott George" <sgeorge@partshounds.com> skrev i melding
> news:Pine.GSO.4.50.0309032105550.8234-100000@phwww.partshounds.com...
> >
> > On Wed, 3 Sep 2003, Per Knutsen wrote:
> >
> > > I have a system running Solaris 9 on a Sun 280 R, it have sevreal
> > > metadevices that works fine except for one witch i build up by four
> > > partisions, when I run the command metastat d7 i reports that at sub
> mirror
> > > needs maintance, and it says to run the command metareplace d7 c2t3d0s0
> <new
> > > device>. So i wounder is it something physical with the disc or does it
> need
> > > a fsck or something equal
> > >
> > > What can I do
> > >
> > > Oracle stores it files on the d7 disc
> > >
> > > Thanks
> > >
> > >
> > >
> > Please provide the output to "metastat". You didn't leave us enough
> > information.
> Output from metastat
> d7: Mirror
> Submirror 0: d71
> State: Needs maintenance
> Submirror 1: d72
> State: Needs maintenance
> Pass: 1
> Read option: roundrobin (default)
> Write option: parallel (default)
> Size: 284332491 blocks
>
> d71: Submirror of d7
> State: Needs maintenance
> Invoke: metareplace d7 c3t8d0s0 <new device>
> Size: 284332491 blocks
> Stripe 0: (interlace: 64 blocks)
> Device Start Block Dbase State Reloc Hot Spare
> c3t8d0s0 0 No Maintenance Yes
> c3t9d0s0 2889 No Maintenance Yes
> c3t10d0s0 2889 No Maintenance Yes
> c3t11d0s0 2889 No Maintenance Yes
>
>
> d72: Submirror of d7
> State: Needs maintenance
> Invoke: after replacing "Maintenance" components:
> metareplace d7 c4t8d0s0 <new device>
> Size: 284332491 blocks
> Stripe 0: (interlace: 64 blocks)
> Device Start Block Dbase State Reloc Hot Spare
> c4t8d0s0 0 No Last Erred Yes
> c4t9d0s0 2889 No Last Erred Yes
> c4t10d0s0 2889 No Last Erred Yes
> c4t11d0s0 2889 No Last Erred Yes

Looks like you had some problems here.
Are the disks listed available at all?
i.e. can you see them and look at their partiontables in format?

Does your /var/adm/messages* files have any failures recorded (they
should)

what does metadb give?

If all disks seem available, i would start with just enabling the disks
oen by one, letting them sync up before i take the next.
Starting with thoose that DOESNT have last erred, sicne the system sees
thoose as the most correct mirror half.

i.e. in your case, IF the disk is available and working.

metareplace -e d7 c3t8d0s0

The waith until that is synced (will take some time) and check whci disk
it now wants you to replace, do the same for that and so forth until all
disk in MAINTAINACE state are ok.

Then you can start with the Last Erred disks, either doing the same thing
going through them one by one, or just detach the mirror and reattach it
to sync them all at once.

You will probably need to use the force mode to detach it, and DO wait
until one mirror is all OK before you try it.

Now, this routine is dependant upon that nothing really is wrong with the
disks and that you just need to reeenable them, if any disk is truly
faulty you need a proper replacement, call your service vendor.

If you have a contract, you could let them do it :-)

/Johan A

Relevant Pages