We at the University of Guelph just upgraded our FAS920 Cluster recently. We went from 7.0.6 to 7.2.3. We run NFS and CIFS and had to go the regular upgrade path - disruption to services was at most 5 minutes per head (NFS was unavailable for the reboot only - which was about 70 secs) and was done at an early morning hour on a Sunday. The only thing that stood out was that it required approx 24hrs to build all the i2p databases for the volumes (low priority background task). Other than a few commands working differently we have not really faced anything major (or minor for that matter) related to the upgrade.
Because of our use of CIFS we really have never done a NDU - our requirements are not so stringent that we can't schedule a small outage in non prime hours.
Hope this helps
Hugh A. Smith System Analyst Campus Services Computing and Communication Services University of Guelph
phone: (519) 824-4120 x56351 fax : (519) 763-6143
****
"of all the things I have lost.. I think I miss my mind the most..." - Ozzy
On Tue, 25 Sep 2007, Kennedy, Jeffrey wrote:
I'm curious what is so valuable about it? As I said, CIFS connections are still lost and NFS and iSCSI both will easily survive a reboot, so where does the NDU value come into play? I'm not certain what it buys.
Jeff Kennedy
QCT Engineering Compute
858-651-6592
From: Leeds, Daniel [mailto:dleeds@edmunds.com] Sent: Tuesday, September 25, 2007 12:28 PM To: Kennedy, Jeffrey; Borzenkov, Andrey; philip.boyle@eircom.net; toasters@mathworks.com Subject: RE: ONTAP UPGRADE - FEEDBACK REQUEST
NDU is highly valuable when you are running 24x7 environments with live active data. We have had alot of success with this but still prefer a scheduled upgrade downtime if possible. Both work great. :)
-- Daniel Leeds Manager, Storage Operations Edmunds.com
-----Original Message----- From: owner-toasters@mathworks.com on behalf of Kennedy, Jeffrey Sent: Tue 9/25/2007 11:16 AM To: Borzenkov, Andrey; philip.boyle@eircom.net; toasters@mathworks.com Subject: RE: ONTAP UPGRADE - FEEDBACK REQUEST
NDU really isn't is it? CIFS connections are still severed and NFS survives both an NDU and a reboot easily enough.
I've never understood the value of NDU honestly but if someone has a success story I'd be interested to hear it.
Thanks.
Jeff Kennedy
QCT Engineering Compute
858-651-6592
From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Borzenkov, Andrey Sent: Monday, September 24, 2007 9:55 PM To: philip.boyle@eircom.net; toasters@mathworks.com Subject: RE: ONTAP UPGRADE - FEEDBACK REQUEST
As you have clusters, I would recommend first updating to 7.0.6 and then - if required - to 7.2.3 or whatever version is available at this time. Then both of them are NDU. NDU between major releases are supported first starting with 7.2.3 and minimal base release is 7.0.6.
Personally I do not see any point jumping through 7.1 additionally.
Otherwise update from 7.0.4 directly to 7.2.3 was quite straightforward.
? ????????? / With best regards / Mit freundlichen Gr�?en
Andrey Borzenkov Senior system engineer
From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Philip Boyle Sent: Monday, September 24, 2007 9:19 PM To: toasters@mathworks.com Subject: ONTAP UPGRADE - FEEDBACK REQUEST
Hi guys,
I'm currently running a number of FAS940 clusters with DATA ONTAP Version 7.0.4 (Sun Feb 5 00:52:53 PST 2006 )
Primarily NFS with some ISCSI.
I'd like to hear back from you re your experiences upgrading to the current s/w releases available on the NOW Site for the FAS940
series system.
http://now.netapp.com/NOW/cgi-bin/software?product=Data+ONTAP&platform=F....
Our immediate options are to take a small step to 7.0.6, move up a little further to 7.1.2.1 or go for the latest and greatest, 7.2.3.
Have you guys seen any issues along the upgrade path to any of the above releases?
Looking at the stats there are a decent number of systems running 7.2.3, however the average run-time days per system
is quite low.
Feedback appreciated.
Thanks,
Philip