I _would_ dump, then replace all the disks, but we can't buy pre-canistered 9GB SCSI disks anymore. So we're (on the advice of our sales rep) replacing the disks in the canisters ourselves. This is a laborious process, and I don't want to add to the downtime between dump and restore by trying to do it offline, so we're trying to do it one by one while the filer is online.
Are we crazy?
Jason Priebe WRAL OnLine http://www.wral-tv.com/
-----Original Message----- From: Gordon Keegan To: Priebe, Jason; 'toasters@mathworks.com' Sent: 11/12/99 12:01 PM Subject: Re: failing SCSI disks
- Priebe, Jason (priebe@wral-tv.com) done spit this rhetoric:
This is a really stupid question, but I can't find any answers to it in the documentation. We are going to replace the 4GB drives in our F210's shelf with 9GB drives by manually failing the 4GB disks one at a time.
Not sure why you'd want to do that, as this would make only 4G of each
of
the 9G drives useable...
What he said. When a smaller drive fails over to a larger spare, the larger spare is marked down as having the same capacity as the smaller. You would end up with a shelf of 9GB drives and the usable capacity of a shelf of 4GB drives. Better off dumping to tape and restoring to a shelf of empty 9GB's.
Are we crazy?
A kind of ... ;-) Remember, while your system is doing a reconstruct your data isn't raid-protected. If one error more occurs you cannot recompute this data. And doing so many reconstructs increases the probability of running into double disk errors a lot. If you are running 5.3 or above i would suggest that you speak to your local SE/SR and ask him if he gives you a snapmirror license for a day and you mirror the data to the 9g disk volume. Use a one minute schedule. After the level 0 dump is done, wait for some transfers more to complete and then switch off nfs and cifs and wait for the next transfer. If it is completed, rename your src/dest volumes and make the destination volume read/writeable to swap between the identities of src and dest volume.
Oliver
-----Original Message----- From: owner-dl-toasters@netapp.com [mailto:owner-dl-toasters@netapp.com]On Behalf Of Priebe, Jason Sent: Samstag, 13. November 1999 14:21 To: 'toasters@mathworks.com' Subject: RE: failing SCSI disks
I _would_ dump, then replace all the disks, but we can't buy pre-canistered 9GB SCSI disks anymore. So we're (on the advice of our sales rep) replacing the disks in the canisters ourselves. This is a laborious process, and I don't want to add to the downtime between dump and restore by trying to do it offline, so we're trying to do it one by one while the filer is online.
Are we crazy?
Jason Priebe WRAL OnLine http://www.wral-tv.com/
-----Original Message----- From: Gordon Keegan To: Priebe, Jason; 'toasters@mathworks.com' Sent: 11/12/99 12:01 PM Subject: Re: failing SCSI disks
- Priebe, Jason (priebe@wral-tv.com) done spit this rhetoric:
This is a really stupid question, but I can't find any answers to it in the documentation. We are going to replace the 4GB drives in our F210's shelf with 9GB drives by manually failing the 4GB disks one at a time.
Not sure why you'd want to do that, as this would make only 4G of each
of
the 9G drives useable...
What he said. When a smaller drive fails over to a larger spare, the larger spare is marked down as having the same capacity as the smaller. You would end up with a shelf of 9GB drives and the usable capacity of a shelf of 4GB drives. Better off dumping to tape and restoring to a shelf of empty 9GB's.
-- Gordon Keegan, M10-10 Phone: (508) 261-4696 Motorola ING Fax: (508) 261-5757 20 Cabot Boulevard Mansfield, MA 02048 email: lgk011@dma.isg.mot.com
-
Oliver Krause -
Priebe, Jason