Networker

[Networker] New install on RH Linux (from Solaris)???

2009-09-22 17:47:56
Subject: [Networker] New install on RH Linux (from Solaris)???
From: George Sinclair <George.Sinclair AT NOAA DOT GOV>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Tue, 22 Sep 2009 17:43:14 -0400
Hi,

We'd like to install the latest "stable" version of NetWorker Network Edition for RHEL on a new Linux server and storage node. We have the new equipment all set and ready to go. These machines have never run NW before so this will be a fresh install. We're migrating from an old release (running on Solaris), and I have some questions below.

We've been way behind on this as we didn't have the equipment in place prior when EMC phased out support for the old release. I've seen a few posts over the last few months regarding upgrading from old releases so please hold those tomatoes! I'm looking mainly for advice and/or any gotchas before proceeding.

First, here are the details for the new and current (old) equipment:

New server
----------
Hardware: Dell PowerEdge 1950
OS: Red Hat Enterprise Linux Server release 5.4, 64 bit
Memory: 8 GB RAM; dual quad processor (Intel)
Disk space: Lots!
Kernel (`uname -a`): Linux server_name 2.6.18-164.el5 #1 SMP Tue Aug 18 15:51:48 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux

New snode
---------
Hardware: Dell PowerEdge R905
OS: Red Hat Enterprise Linux Server release 5.4, 64 bit
Memory: 16 GB RAM; four quad processor (ADM)
NIC: 4 GIG LACP Trunk
Disk space: Plenty!
Tape library: SAS attached Dell ML6000 tape library with six LTO-4 drives.
Kernel (`uname -a`): Linux snode_name 2.6.18-164.el5 #1 SMP Tue Aug 18 15:51:48 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux

Current server
--------------
Hardware: Sun Enterprise 450
OS: Solaris 2.9, 64-bit
NetWorker: 7.2.2.Build.422 Network Edition

63 client indexes, totaling 21 GB
Media database is 325 MB.
All of /nsr is 16 GB.
Note: we have some indexes that live on two other disks, totaling about 6 GB (16+6=21)

Current snode (1 of 2)
----------------------
Hardware: Dell PowerEdge 6650
OS: Red Hat Enterprise Linux ES release 3, 32-bit
NetWorker: 7.2.2.Build.422 Network Edition
Tape library: LVD SCSI attached StorageTek L80 with eight IBM Ultrium LTO-3 drives.

Current snode (2 of 2)
----------------------
Hardware: Dell PowerEdge 6650
OS: Red Hat Enterprise Linux ES release 3, 32-bit
NetWorker: 7.2.2.Build.422 Network Edition
Tape library: LVD SCSI attached Quantum M1800 with four SDLT600 drives (two M1500s stack linked w/ two drives each)


I guess we'll be running the new release in evaluation mode over a few weeks to test things out. Our current set up features NW 7.2.2, running with a Solaris primary server and two RH Linux storage nodes. Our plan was to move the Quantum library over to the new snode and run it on a separate card from the one(s) that will managing the new Dell LTO-4 tape library and then shut down the old snode (2 of 2). This would be done once testing of the new server, snode and Dell tape library were completed. We would keep the other snode (1 of 2) running in parallel for a quite a while, though, and, of course, those LTO-3 tapes can still be read on the new LTO-4 library.

My last understanding was that EMC does not officially support migrating from Solaris to Linux but that many have done so successfully. I would like to obtain the appropriate media-kit from our Legato rep. and proceed to read the documentation, do the install and then run some tests on the new server, library, etc. and, of course, also test migrating the old database and client indexes from the Solaris box. The
new server will have a different IP name, at least during testing.

Here are some important questions that I have that might not be covered in the installation notes:

1. What is the most recent version of NW that would be recommended?

We've considered just requesting the latest so we won't have to upgrade again for a while, but I've seen some negative feedback concerning some of the later minor releases, but 7.4.4 seemed to be in general favor with everyone, so wanted to know what media kit to request. I'd rather be stable and avoid headaches of having to patch or deal with problems and then another upgrade as soon as the next release fixes that.

2. We considered upgrading the Solaris box first and then doing the migration to Linux, but we decided against this because if something goes wrong then we'd be in trouble with our existing backups, and the upgrade might be irreversible. Seemed better, therefore, to simply install a whole new release on the new Linux servers, test that and leave the old Solaris box alone, which we know works. Also, since we're moving to a whole new OS anyway, maybe using a newer version at the same time isn't that much more of a hurdle.

Does this sound reasonable?

3. Not sure about little Endian versus big Endian. I know this has been discussed before, so my thinking was simply to back up all the client indexes (level=full) using the old server and then restore them from tape to the new server. Ditto for the media database (disaster recovery, I guess) and hope that all this should be OS independent. Not sure whether it was the media database or the client indexes that could just be copied over? Regardless, since our total disk usage on the old Solaris server is not that large, I thought maybe I would just recover the whole thing from tape. Obviously, I will have to scan the tape on the new server, so maybe I'd use a new tape when I back up the indexes and bootstrap on the old server so the tape won't contain anything more than necessary, thereby reducing the scan time.

Does this all sound correct, and is this the recommended approach?

4. As far as an update enabler, since we'd be doing a fresh install, that would not apply, right? Would there be any concern about recovering the old Solaris 7.2.2 database on the new Linux server, running the newer release, because we didn't first upgrade the old Solaris NW version to a newer version?

5. I'm unclear on how exactly the license keys will work (will talk to licensing), but I was going to ask for temporary Linux enablers for all the ones we use now on the Solaris server and then once we're ready to go production mode, permanently enable those by converting the old Solaris licenses. Moreover, once we're at that point we'd be dropping one of two current snodes, so that would free up that snode license, too.

That sound right?

6. What about the host name for the new server once we're ready to start "real" live nightly backups?

Should we change the old server's host name, back it up from the new server as a precaution, power it off and then continue running backups on the new server using the new name, OR change the new backup server's
name to match the old one?

7. Any gotchas or pitfalls we should be aware of in regards to the migration that might not be explained in the notes?

Thanks!!!

George

--
George Sinclair

- The preceding message is personal and does not reflect any official or unofficial position of the United States Department of Commerce -
- Any opinions expressed in this message are NOT those of the US Govt. -

To sign off this list, send email to listserv AT listserv.temple DOT edu and type 
"signoff networker" in the body of the email. Please write to networker-request 
AT listserv.temple DOT edu if you have any problems with this list. You can access the 
archives at http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

<Prev in Thread] Current Thread [Next in Thread>