Can you expand your question a bit please? What exactly are you upgrading - hardware or software? It sounds like hardware only. Is the new box same OS type as old one? If yes do you plan to replace old backup server with new one in a way that new one will have same hostname? Such transition should work without any problems - you just need to do a license transfer as hostid will change.

Second part is NDMP backup configuration - however to make a suggestion first you must know what your requirements are (eg. SLA, compliance rules, etc), required or available resources and then come up with solution. So, what you proposed can be both wrong and right - it is impossible to give correct answer based on what you wrote.

However, based on your setup let's imagine group1. It has client1, client2 and client3 and based on what you said difference between them is retention and schedule I guess. I will assume that client1..3 are logical instances of the same filer. I would never to the setup like this as I would try to limit this to 2 retentions; daily (backup each day) and monthly (backup once a month). Unless you would cover different data set I can't see why weekly backup with different retention would be required. Second, I would setup group based on data volume and time required to backup (so that it fits my backuo window), but I would also make sure that group membership determines schedule (daily vs monthly) and that doesn't seem to be the case in your story.

There is number of things here that can be done (from HW and SW point of view) so getting professional services to help would be something advisable in case you get lost.

im mooving from a slower dell server to a newer dell server. Both 2003 standard, new server name and ip so i need to re license since both will be active.

i have copied over the NSR directory but have not placed it into the LEGATO directory on the new server as of yet. I will leave the old server online for restore purposes only, as i do not know as of now if i can transfer complete index's over.

I need a full backup every week. That is why i have the weekly full. The retention will only be 1 month to save on tape cost, and monthlys will be stored off site for 6 months. Our data increases about 500 GB per day, thats why i need weekly full's.

My NDMP is from a EMC 704g that supports i belive 16 ndmp sessions, i am creating 4 groups so i spread the data and access every drive i have available.

I am roughly trying to backup 20TB in less than 3 days. FRI night - Monday morning.

I have a legato support case opened to assist and they have provided me with some knowledge on how to succesfully accomplish this. Also the Install PDF from the legato web site has helped alot.

p.s. - croatia is a lovely country :]

I would say:
- install NW on the new box and use same path and same version as on old one and shut it down
- shutdown old server
- remove licenses using nsradmin (but write down that data somewhere)
- copy NW databases (meaning index, res and mm directories) from old to new server
- start old server - should run in eval mode now (if I remember correctly you will be able to run restores even when eval will expire)
- start new server - enter enablers (you will need to supply new hostid and re-register those licenses after host transfer process)
- test few restores using new server

If above works there is a good chance everything went fine.

When I said daily before I was think of backup on daily basis where at least once a week you have full backup.

Are you using network backup or you write down to dedicated drives assigned via FC to filer (or SnapImage)? What is the speed performance you get per stream from filer? For 20TB you would need cca 72MB/s for 3 days (one stream). Since you will have several session (you are backing up snapshots or..?) and probaly at least two drives that brings us to LTO2 world speeds - the only question is if you can push the data all the time to write that way (probably not - but then you may have more drives so...).

i have 4 lto-2's in a adic scalar i2000 with 100 slots used. Im trying to get 4 more drives.

i see avarege of 25+mb/s usually, i could bump up the cache on the san during off peak times to speed it up a little.

On Networker 7.x you cannot get rid of your licenses unless you throw away your media database.
Even if you delete the licenses, they will get revived at your (system's) first access to the media database.
If the licenses get revived, your old server is on the same network with the new one, both servers will get disabled due to duplicate licenses.
Beware...

You can get rid of the license. What now remains as record in media db is the fact that you used that license, but nothing is stopping you from migrating. Remember - to do restores you do NOT need licensed server. So, if you are planning migration there should be no problem at all.

The trick with media database has been introduced I would guess as too many people were running eval mode for year. Not anymore.

As last thing, I heard (but cannot confirm) that Legato could provide extra license in case that you need to keep old one and get new one for new server, but you must sign that you would use only one server at the time. I did heard that once from one customer who apparently heard that on one training in London @ EMC. Once again, I have no idea if that is correct or not.

I ran into that problem, already trying to run duplicate servers. It actually creates its own client for the new server within the default pool.

The supported way of migrating a Networker server is as a disaster recovery:

* backup your indexes using savegrp -O groupname
* ensure you have an up-to-date bootstrap report
* ensure you have shutdown the Networker services on the old server or taken it off the network

On your new server

* install Networker (ideally same build as before - also try to keep the O/S same version and SP/patch level)
* configure one drive and manually mount the tape with the most recent bootstrap
* run mmrecov specifying the details from the bootstrap report - this recovers the resources (configuration) and media database (what is on the tape)
* stop Networker and rename nsr/res to res.old res.R to res
* restart Networker - you should now have your config back; you will need to reauthorise your licenses within 14? days
* reset and reinventory your jukebox with nsrjb -HEv then nsrjb -Iv
* recover the indexes with nsrck -L7

It is possible to copy the files across but it may be more sensible to follow the supported method.

What HBA's do you have in your Legato Server?

Are you backing SAN Disk directly to the i2000 across the FC?

Do you still backup Server with Direct attached storage to the i2000.

I am interested as we are implementing a i2000 in the next few weeks.

I recently upgraded my backup server and used the quick/shortcut process by copying all the files to tne new server. Now the indexes prior to my migration are not available on the new server, is there a way to append them to the current indexes?

You should open separate thread for your question. Copying only works only under certain circumstances.

If you used the unsupported method and found the indexes were missing but everything else was okay I would suggest using the supported method for index recovery - running nsrck -L7 at the command line.

I would suggest you stop Networker, rename /nsr/index to /nsr/index.old, restart Networker then run nsrck -L7 which is the supported way of index recovery.

And as Hrvoje has said if you have an issue its best to open a new thread - for one thing it is less confusing, and secondly it enables him to get enough points to become a Super Grandmaster and knock a few of those Documentum guys down to size

Thanks for the replies, will try and if there are any more issues, will open a new thread.