Several COH-NUG people have asked me on the side a few Linux questions.....
So, I thought I would bring these queries to cyberspace.
PLEASE answer to help one another out .... and let's start a discussion. Our friends in Central OH need some advice and experience.
I will purposely not use customer's names.
1. Within COH-NUG we have SLES 10SP2 customers with Novell support.
2. Others started with RedHat ... and now are purely CentOS. No support.
3. So the questions are:
a). how difficult with the SLES 10 to 11 upgrade be?
b). how successful?
c). does anyone just use OpenSuSE?
d). What about switch from SLES to CentOS for apps like Oracle, WAMP etc.
Who has wrestled with this?
Hi - One of our customers are using SLES on IBM blades, connected to the storage via FCP, with no internal dribes installed on these blades.
There are a few tricks to be aware of, but for over two years now, we have not had any failure on OS level.
If it was my systems, I would definetely consider of staying on SLES, wether it be 10 or 11. To upgrade from 10 to 11 on the outher hand should not be a problem, especially if your boot disks also reside on SAN.
If you are using FC, you can create a boot LUN, on SLES 10, and upgrade it to 11, to see what can go wrong, and if it works, roll it out to your other servers.
By te way - I forgot the parts about OpenSuSE and Apps.
OpenSuSe is used for development and is NOT supported in a production environment, should you encounter problems.
We are using Oracle and other DB's on SLES at various clients, without performance or any other problems.
We are using Redhat ( various versions ) and Netapp. We are not using any FC support to Netapp except for Media agents for CommVault to a VTL. That having been said; the NFS part should be good. We have had no problems with Oracle under RHEL 3,4 and 5 and NFS. Based on my FC experience with Cisco and Hitachi, I doubt there would be much problems with FC.
I don't see much of a differace between CentOS and RH except from support/version number issues. They are still a recompile of RH right?
But I would ask a follow up question or two, NFSv2? Fiber channel? What is the hardware? Is it on the support matrix?
I would be interested in any answers.
You posed some questions
These are aspects I am seeing with our local customers of NetApp+Linux.
1. AFAIK CentOS is still a recompile of RedHat
2. See http://centos.org/
3. And, KickStarts work fine for me in my little DEV environment.
4. Oracle "Unbreakable" Linux see oss.oracle.com seems to still be a RedHat-based environment
5. I am not aware of anyone using the Oracle Linux product ... even for Oracle. Perhaps someone in COH-NUG does??
6. I am not aware of anyone using OpenSuSE in place of SLES ... except for play. Perhaps someone in COH-NUG does??
7. COH-NUG customer(s) *are* using Linux via FCP & some iSCSI ... but it seems to be CentOS or SLES mostly
-- with Oracle especially and/or WAMP-like Web Servers
8. To see Interop "http://now.netapp.com/matrix/"
9. You will want to use the Host Utils from NetApp.
10. Interestingly Linux seems to be a bit more "chatty" than let's say Windows / Unix ... when *checking* that there secondary paths to the NetApp controllers are alive per se.
11. There are current NetApp releases of ONTAP that do support NFS v2 still.
12. BTW, I have been seeing much better maturity in the SLES mpio sw lately than in previous releases. Also, SLES iSCSI SW Initiator ... I have not played with that yet.
I logged on to http://now.netapp.com/matrix/ ... namely the Interop Matrix.
CentOS (5 / 5.2) is represented there from a VM Guest perspective.
It is not represented (yet?) from a host (physical server) perspective .... (think FCP/iSCSI) ....
Talking with NetApp SEs ... PVR is possible for CentIOS support on your company's behalf.
A PVR is a means by which your NetApp Systems Engineer can request support for your company.