metareplace problem

metareplace problem

Post by Uwe Wolfra » Thu, 09 Jun 2005 05:08:56



I'm running Solaris 9/Sparc (latest patches included) on SVM mirrored
disks. Recently I stumbled over a phenomenon I haven't noticed on Sol8.

After a disk failed I did a metareplace from a scratch spare disk (not
a hotspare) like

metareplace mirror deva devb

The mirror resyncs and runs fine. After replacing the failed disk I
reversed everything:

metareplace mirror devb deva

resync took place and everything runs fine again.
But after a reboot the system marked deva as bad. If I do a
metadetach -f, metaclear, metainit, metattach everything works
again. Has this to do with the device ID and how do I get around
this? The docs give no hints.

Uwe

 
 
 

metareplace problem

Post by Jorgen Moquis » Thu, 09 Jun 2005 07:51:10



> I'm running Solaris 9/Sparc (latest patches included) on SVM mirrored
> disks. Recently I stumbled over a phenomenon I haven't noticed on Sol8.

> After a disk failed I did a metareplace from a scratch spare disk (not
> a hotspare) like

> metareplace mirror deva devb

> The mirror resyncs and runs fine. After replacing the failed disk I
> reversed everything:

> metareplace mirror devb deva

> resync took place and everything runs fine again.
> But after a reboot the system marked deva as bad. If I do a
> metadetach -f, metaclear, metainit, metattach everything works
> again. Has this to do with the device ID and how do I get around
> this? The docs give no hints.

> Uwe

could be metadevadm or metadb problem ?
/

 
 
 

metareplace problem

Post by Uwe Wolfra » Fri, 10 Jun 2005 03:55:54


Quote:> could be metadevadm or metadb problem ?

I wasn't aware of metadevadm, so this should fix it.

Thanks
Uwe

 
 
 

1. metareplace

Hello,

I've got a Sun server using Solaris 9 which has two disks in RAID 1
mirror using Solaris Volume Manager. The problem now is that the machine
somehow didn't see the second disk and booted the OS. Now I've rebooted
again and it sees the two disks but if I do a metastat I see the following:

d6: Mirror
     Submirror 0: d16
       State: Okay
     Submirror 1: d26
       State: Needs maintenance
     Pass: 1
     Read option: roundrobin (default)
     Write option: parallel (default)
     Size: 57617760 blocks

d16: Submirror of d6
     State: Okay
     Size: 57617760 blocks
     Stripe 0:
         Device     Start Block  Dbase        State Reloc Hot Spare
         c0t0d0s6          0     No            Okay   Yes

d26: Submirror of d6
     State: Needs maintenance
     Invoke: metareplace d6 c0t2d0s6 <new device>
     Size: 57617760 blocks
     Stripe 0:
         Device     Start Block  Dbase        State Reloc Hot Spare
         c0t2d0s6          0     No     Maintenance   Yes

Now I've tryed to do a metareplace d6 c0t2d0s6 c0t2d0s6 but this doesn't
work. How can I tell SVM that he has to resync using the same disk as
before ?

Many thanks

2. Can't paste in FVWM2:

3. metareplace again

4. Root Password

5. metareplace - can't find component in unit

6. Error with e2fs?

7. metareplace question (Solaris 9)

8. ZIP drive goes nuts on bad disk

9. metareplace -e d40 c0t0d0s0 what will happen

10. md.tab not correct after metareplace

11. metareplace -e and 280R not working: can I use this trick ?

12. Metareplace Bad RAID 5 Disk W/Hotswap Disk?

13. Metareplace one disk into a mirror striped array?