2

I have replaced failed drives on Solaris 10 that are mirrored with SVM before. I am wondering if I have been doing it incorrectly... or taking a huge risk. I follow the usual metadetach, metaclear, metadb -d, and cfgadm -c unconfigure steps. My question is should I be doing this in single user mode? I haven't in the past the and haven't noticed any issues but I am curious. I haven't really found anything in the docs that corresponds to this procedure and nothing on the run level.

0

Your procedure is absolutely fine. I do this all the time, generally without issue. [1]

Solaris is designed to handle this sort of thing and does it very well. Of course, you may want to leave the re-mirroring until the system is less busy so that you don't get an I/O performance hit while it's re-mirroring.

[1] The only time I've had issues are when the disk has failed in an odd way, so that Solaris doesn't recognise that it's dead, hence won't allow you to offline it, and doesn't recognise that it's gone when you pull it. In these cases, all you can do is a reconfigure reboot (boot -r or touch /reconfigure && init 6).

0

SVM is meant to be used in multi-user mode (and many servers have hot-swap disks so that you don't even have to take the machine down to replace them). Depending on the hardware, there are specific commands to offline the disks before you remove them, as well.

  • He already is offlining the disks: cfgadm -c unconfigure. The procedure is fine and is what I do all the time on SPARC systems. – James O'Gorman Mar 2 '12 at 22:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.