I was planning on doing the upgrade to my NetWorker servers and storage nodes first. I was planning on letting this run, then going back later and upgrading the clients. Is this a feasible approach?
I was planning on doing the upgrade to my NetWorker servers and storage nodes first. I was planning on letting this run, then going back later and upgrading the clients. Is this a feasible approach?
Server and storage nodes should always be on the same level. Especially when moving to major release number (minor release numbers like patch release can work together even EMC does not suppor that I believe). Client should work fine if they are one major release number below server. So, your approach should be fine.
I am being hammered by my management on the risks associated with the upgrade. I explained that an MMRECOV could be done to the latest backstrap and we could revert back to 7.2.1 ... his next question was, has this been tested.
My question to you would be - what is the reason to go to 7.3? What is that you need so badly that you wish to put initial release to your production environment without any prior test?
will the upgrade break anything that is already running (e.g., due to filter/port changes?) Can someone help to ease my management's, as well as my own, fears in conducting this upgrade?
Oh, perhaps my approach was wrong
I know few customer who went to 7.3 - no problems.
I know few customer who went to 7.3 - big problems.
I know few customer who went to 7.3 - minor problems as patches did resolve their problems.
Without knowing environment it's hard to say which category you will fit. Without test, without plan, without preparations - I would say you may face big problems. 7.3 is very much different than earlier releases and waiting for 7.3.1 is something I would recommend. I was told recently that for the end of May jumbo patch release is planned. Would that be a 7.3.1 or simply jumbo patch - I have no idea. There is at least one cumulative patch which exists already.
The best thing you can do is to: - make a list of you backup clients (application modules and OS) - describe what OS is your backup server and storage nodes - describe what library you use and how - describe extra to basic functions you use (eg. staging, cloning)
and send that to EMC or your support and tell them your management will kill you if you blow something so if there is a patch you require for 7.3 upgrade - they better give it you now.
With all that in mind, I personally would still recommend to wait for 7.3.1 unless you really really need functionality from 7.3 and there is no way back.
I agree that mmrecov can get you back, but how do you answer following question: on day 1 you ugraded to 7.3. On day 1, day 2 and day 3 we had production backups. On day 4 we had major problem. On day 5 we found this is a bug and we need to move back. You do mmrecov and we are back to day 0. What about backups from day 1-5?
Every major release (and 7.3 is) needs some "domestic environment" testing. Smaller environment are usually not affected at all as they use basic features. However, if this is enterprise environment - don't play with it. Enterprise environments use many features and many different HW and SW with all that mixed together you may get unexpected results at the end. This was, it is and will be truth for each major release for every application or OS (so this is general thing - not something specific to EMC or whoever).
"Would that be a 7.3.1 or simply jumbo patch - I have no idea"
I had asked for the v7.3.1 rel date in a support case this week, and was advised that the Jumbo patch was being prepared for release asap, and that the actual public release of v7.3.1 was not yet known....
so given their answer, it would seem they will be separate releases...
I tried to install 7.3 on my AIX machine and I guess I am lucky because it would not install. I didn't have too many problems with the Windows machines so I guess it is mostly a matter of what platform you are using.
Just to clarify what Hrvoje said a little earlier - server should be a later version than the clients and ideally the storage nodes should be the same (though will probably work okay). I have had issues running Networker 5.x clients with Networker 7.x server but not running 5.x with 6.x and 6.x with 7.x! Around the introduction of Networker 7 (possibly before) it was said that it was okay for Networker clients to be at a higher version than the server though I've only seen this in a very small number of instances.
Practically speaking you will only have a small number of Networker servers and storage nodes so you should be able to upgrade those all at once; however you may have quite a significant number of Networker clients and it may not be reasonable to get them all upgraded at once.
No it was on the server. The error was that there was nothing to install from the CD. I tried copying it to the install directory and running intuoc and still nothing, this is an AIX 5.3 machine.
Everything appeared to be there, the files are there, it just says there is nothing to update when trying to install them. This would appear to mean that the date on those files are older than the ones currently installed. I have thought about doing a touch command to change the dates but with people having problems with 7.3 I am not sure if I want to take the chance. I will probably wait for the next update.
There is already jumbo patch which should give us 7.3-STABLE (to use BSDish language). I don't remeber having such problem, but then again I always download the version for net rather than using CD distributed one.
brianluck1
40 Posts
0
May 4th, 2006 10:00
ble1
4 Operator
•
14.4K Posts
0
May 4th, 2006 13:00
servers and storage nodes first. I was planning on
letting this run, then going back later and upgrading
the clients. Is this a feasible approach?
Server and storage nodes should always be on the same level. Especially when moving to major release number (minor release numbers like patch release can work together even EMC does not suppor that I believe). Client should work fine if they are one major release number below server. So, your approach should be fine.
ble1
4 Operator
•
14.4K Posts
0
May 4th, 2006 13:00
I can see you love dangerous life
associated with the upgrade. I explained that an MMRECOV
could be done to the latest backstrap and we could revert
back to 7.2.1 ... his next question was, has this been
tested.
My question to you would be - what is the reason to go to 7.3? What is that you need so badly that you wish to put initial release to your production environment without any prior test?
filter/port changes?) Can someone help to ease my
management's, as well as my own, fears in conducting
this upgrade?
Oh, perhaps my approach was wrong
I know few customer who went to 7.3 - no problems.
I know few customer who went to 7.3 - big problems.
I know few customer who went to 7.3 - minor problems as patches did resolve their problems.
Without knowing environment it's hard to say which category you will fit. Without test, without plan, without preparations - I would say you may face big problems. 7.3 is very much different than earlier releases and waiting for 7.3.1 is something I would recommend. I was told recently that for the end of May jumbo patch release is planned. Would that be a 7.3.1 or simply jumbo patch - I have no idea. There is at least one cumulative patch which exists already.
The best thing you can do is to:
- make a list of you backup clients (application modules and OS)
- describe what OS is your backup server and storage nodes
- describe what library you use and how
- describe extra to basic functions you use (eg. staging, cloning)
and send that to EMC or your support and tell them your management will kill you if you blow something so if there is a patch you require for 7.3 upgrade - they better give it you now.
With all that in mind, I personally would still recommend to wait for 7.3.1 unless you really really need functionality from 7.3 and there is no way back.
I agree that mmrecov can get you back, but how do you answer following question: on day 1 you ugraded to 7.3. On day 1, day 2 and day 3 we had production backups. On day 4 we had major problem. On day 5 we found this is a bug and we need to move back. You do mmrecov and we are back to day 0. What about backups from day 1-5?
Every major release (and 7.3 is) needs some "domestic environment" testing. Smaller environment are usually not affected at all as they use basic features. However, if this is enterprise environment - don't play with it. Enterprise environments use many features and many different HW and SW with all that mixed together you may get unexpected results at the end. This was, it is and will be truth for each major release for every application or OS (so this is general thing - not something specific to EMC or whoever).
cfaller
96 Posts
0
May 4th, 2006 22:00
I had asked for the v7.3.1 rel date in a support case this week, and was advised that the Jumbo patch was being prepared for release asap, and that the actual public release of v7.3.1 was not yet known....
so given their answer, it would seem they will be separate releases...
lalexis
2 Intern
•
253 Posts
0
May 26th, 2006 05:00
ble1
4 Operator
•
14.4K Posts
0
May 26th, 2006 06:00
DavidHampson
2 Intern
•
1.1K Posts
0
May 28th, 2006 12:00
Practically speaking you will only have a small number of Networker servers and storage nodes so you should be able to upgrade those all at once; however you may have quite a significant number of Networker clients and it may not be reasonable to get them all upgraded at once.
ble1
4 Operator
•
14.4K Posts
0
May 30th, 2006 05:00
lalexis
2 Intern
•
253 Posts
0
May 30th, 2006 05:00
lalexis
2 Intern
•
253 Posts
0
May 30th, 2006 06:00
ble1
4 Operator
•
14.4K Posts
0
May 30th, 2006 06:00