I'd like to put the cat among the pigeons here(!), and ask why you feel it is necessary or desirable to "contain each volume to a shelf of its own". The same situation you have seen here is likely to occur again if another disk fails (it will, eventually).
One reason to buy a filer is to minimise administration and maximise uptime, and this process will increase administration as well as introduce unnecessary downtime.
One possible explanation is that you need to move some disks later to another filer - in which case identify at that time which disks comprise the volume you wish to move, and then simply move those. As mentioned by others, if you halt the filer, you can move the disks around with no fear whatsoever of confusing the filer - the disk's identity is not related to the slot in which it finds itself, rather to the label on the disk.
It is also desirable to minimise the number of "disk fail" operations, following the maxim of "if it ain't broke, don't fix it" :-)
Regards, Andrew
-----Original Message----- From: Fox, Adam Sent: 10 July 2000 15:55 To: toasters@mathworks.com Subject: RE: volume/shelf containment
There is a simplier way to do this, but it involves halting the filer.
You could simply, halt the filer, place the disks where you want them, then reboot.
Sorry I didn't mention this obvious way before, but I tend to think in terms of no downtime.
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
-----Original Message----- From: Fox, Adam Sent: Monday, July 10, 2000 10:38 AM To: makin@dnrc.bell-labs.com; toasters@mathworks.com Subject: RE: volume/shelf containment
-----Original Message----- From: Bhavnesh Makin [mailto:makin@dnrc.bell-labs.com] Sent: Monday, July 10, 2000 10:04 AM To: toasters@mathworks.com Subject: volume/shelf containment
We have a F740 filer with two 18 gig shelves. Each shelf is dedicated to a volume.We have reserved disk on each shelf as a
hotspare for that
particular volume. So in total we have two hot spares.
Last week one of the disks failed on the first volume and
failed disk
was recontructed on the spare which was physically
present on second
shelf. So we ended up with the volume which transgresses over to another shelf. Our goal was to contain each volume to a
shelf of its
own.
First,is there a way to contain the volumes with hot
spares? Secondly
how do I make the first volume give up the disk on second
shelf and
recontruct it on spare which is physically present on first shelf.
As far as I understand it, spares are global to the filer.
The filer
will pick the smallest spare disk that will fit. So unless you have different sized disks on different volumes, you can't make a filer pick one disk over another.
As far as fixing your current situation. Config your system now so that it only have 1 spare (it may be that way anyway now), then fail over the spare in the 2nd shelf which is now a data drive for the
volume on the
first shelf. In my opinion, the best way to do this is to run the 'disk fail' command rather than pull the drive. I think it's cleaner since you won't have to cause your lop to reset. Then you will force your volume to rebuild on the only spare in the system, which not so coincidentally is the one you want.
Hope this helps.
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
The volume/shelf configuration I used before to describe my problem is not completely accurate. Infact I have multiple volumes with multiple shelves. I just wanted to keep everything clean since other people fill for me sometimes.
Thanks for all the responses.
Bhavnesh
abond@netapp.com wrote:
I'd like to put the cat among the pigeons here(!), and ask why you feel it is necessary or desirable to "contain each volume to a shelf of its own". The same situation you have seen here is likely to occur again if another disk fails (it will, eventually).
One reason to buy a filer is to minimise administration and maximise uptime, and this process will increase administration as well as introduce unnecessary downtime.
One possible explanation is that you need to move some disks later to another filer - in which case identify at that time which disks comprise the volume you wish to move, and then simply move those. As mentioned by others, if you halt the filer, you can move the disks around with no fear whatsoever of confusing the filer - the disk's identity is not related to the slot in which it finds itself, rather to the label on the disk.
It is also desirable to minimise the number of "disk fail" operations, following the maxim of "if it ain't broke, don't fix it" :-)
Regards, Andrew
-----Original Message----- From: Fox, Adam Sent: 10 July 2000 15:55 To: toasters@mathworks.com Subject: RE: volume/shelf containment
There is a simplier way to do this, but it involves halting the filer.
You could simply, halt the filer, place the disks where you want them, then reboot.
Sorry I didn't mention this obvious way before, but I tend to think in terms of no downtime.
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
-----Original Message----- From: Fox, Adam Sent: Monday, July 10, 2000 10:38 AM To: makin@dnrc.bell-labs.com; toasters@mathworks.com Subject: RE: volume/shelf containment
-----Original Message----- From: Bhavnesh Makin [mailto:makin@dnrc.bell-labs.com] Sent: Monday, July 10, 2000 10:04 AM To: toasters@mathworks.com Subject: volume/shelf containment
We have a F740 filer with two 18 gig shelves. Each shelf is dedicated to a volume.We have reserved disk on each shelf as a
hotspare for that
particular volume. So in total we have two hot spares.
Last week one of the disks failed on the first volume and
failed disk
was recontructed on the spare which was physically
present on second
shelf. So we ended up with the volume which transgresses over to another shelf. Our goal was to contain each volume to a
shelf of its
own.
First,is there a way to contain the volumes with hot
spares? Secondly
how do I make the first volume give up the disk on second
shelf and
recontruct it on spare which is physically present on first shelf.
As far as I understand it, spares are global to the filer.
The filer
will pick the smallest spare disk that will fit. So unless you have different sized disks on different volumes, you can't make a filer pick one disk over another.
As far as fixing your current situation. Config your system now so that it only have 1 spare (it may be that way anyway now), then fail over the spare in the 2nd shelf which is now a data drive for the
volume on the
first shelf. In my opinion, the best way to do this is to run the 'disk fail' command rather than pull the drive. I think it's cleaner since you won't have to cause your lop to reset. Then you will force your volume to rebuild on the only spare in the system, which not so coincidentally is the one you want.
Hope this helps.
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
I agree. One big advantage of the filer is that you don't have to worry about this kind of stuff! It might get confusing, but a simple sysconfig -r will tell you where all your disks and volumes are.
Moshe
abond@netapp.com wrote:
I'd like to put the cat among the pigeons here(!), and ask why you feel it is necessary or desirable to "contain each volume to a shelf of its own". The same situation you have seen here is likely to occur again if another disk fails (it will, eventually).
One reason to buy a filer is to minimise administration and maximise uptime, and this process will increase administration as well as introduce unnecessary downtime.
One possible explanation is that you need to move some disks later to another filer - in which case identify at that time which disks comprise the volume you wish to move, and then simply move those. As mentioned by others, if you halt the filer, you can move the disks around with no fear whatsoever of confusing the filer - the disk's identity is not related to the slot in which it finds itself, rather to the label on the disk.
It is also desirable to minimise the number of "disk fail" operations, following the maxim of "if it ain't broke, don't fix it" :-)
Regards, Andrew
-----Original Message----- From: Fox, Adam Sent: 10 July 2000 15:55 To: toasters@mathworks.com Subject: RE: volume/shelf containment
There is a simplier way to do this, but it involves halting the filer.
You could simply, halt the filer, place the disks where you want them, then reboot.
Sorry I didn't mention this obvious way before, but I tend to think in terms of no downtime.
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
-----Original Message----- From: Fox, Adam Sent: Monday, July 10, 2000 10:38 AM To: makin@dnrc.bell-labs.com; toasters@mathworks.com Subject: RE: volume/shelf containment
-----Original Message----- From: Bhavnesh Makin [mailto:makin@dnrc.bell-labs.com] Sent: Monday, July 10, 2000 10:04 AM To: toasters@mathworks.com Subject: volume/shelf containment
We have a F740 filer with two 18 gig shelves. Each shelf is dedicated to a volume.We have reserved disk on each shelf as a
hotspare for that
particular volume. So in total we have two hot spares.
Last week one of the disks failed on the first volume and
failed disk
was recontructed on the spare which was physically
present on second
shelf. So we ended up with the volume which transgresses over to another shelf. Our goal was to contain each volume to a
shelf of its
own.
First,is there a way to contain the volumes with hot
spares? Secondly
how do I make the first volume give up the disk on second
shelf and
recontruct it on spare which is physically present on first shelf.
As far as I understand it, spares are global to the filer.
The filer
will pick the smallest spare disk that will fit. So unless you have different sized disks on different volumes, you can't make a filer pick one disk over another.
As far as fixing your current situation. Config your system now so that it only have 1 spare (it may be that way anyway now), then fail over the spare in the 2nd shelf which is now a data drive for the
volume on the
first shelf. In my opinion, the best way to do this is to run the 'disk fail' command rather than pull the drive. I think it's cleaner since you won't have to cause your lop to reset. Then you will force your volume to rebuild on the only spare in the system, which not so coincidentally is the one you want.
Hope this helps.
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
-
abond＠netapp.com -
Bhavnesh Makin -
Moshe Linzer