Solaris 9 RAID 5 error

[phorbiuz]

I've got a 7 disk RAID 5 device on a Solaris server called d7, totalling 45Gb with 41Gb free. When I do a metastat on it, it returns one of the disks in the state 'maintenance'. All other disks are OK and the metadevice is still working.

What I want to do is remove the 'faulty' disk, in this case c2t9d0s0, from the RAID device and continue to use d7. I'll replace this disk in future but there are no empty slots in the external array to replace it. For now I just want to remove the fault.

Who can tell me how this is done?

Thanks in advance.

 

[coffeysm]

I thought it tells you on the output of metastat what command you need to run.

 

[kHz]

It does show, metareplace -e, if I remember off the top of my head. But that would be to replace it. He wants to remove the disk from the array and run without it for a day, week, month, or whenever the disk is replaced.

If you don't use metareplace you would have to use metadetach.

Why not just leave the disk in until it is replaced?

 

[coffeysm]

Can't you use metatool to remove the disk out of the array?

 

[phorbiuz]

Maybe I should clarify things a little here. When you do the metastat on d7, it does indeed ask you to invoke metareplace and state the replacement device, but we don't have one - the array was built some time back with no resilience. Hence the metareplace command is no good.

I want to remove the device, so that in future I can replace it, then add it.

As for the metadetach bit, if I read the man page correctly it should have been 'metadetach d7 /dev/dsk/c2t9d0s0', but when I run that I get 'metadetach: d7: invalid metadevice type md_raid'.

I that's when I started this post.

Thanks.

 

[coffeysm]

Maybe if you open up metatool and right click the disk that is listed you can choose delete or remove?