How to rebuild Networker Server?

NMC is not relevant in this case.

The index information (/nsr/index) is also not important right now.

What you need to do is a standard disaster recovery:

- Install NW and NMC to the same directory as before.

- Create/define a compatible tape drive.

- I hope you have the "bootstrap" report so you have the info on which tape it has been stored. Insert it.

- Run "mmrecov" and answer the questions carefully to recover the bootstrap.

   This will restore the important databases.

- Follow the instructions to reactivate the old resources.

- If you are lucky this is all you need to reactivate the full server.

- If the "hostid" changed (most likely) then you need to apply for the "auth codes" to keep the license alive. Do this within 2 weeks.

These are the mandatory steps before you address the file index.

Let get to this point before you proceed.

Thanks a million Bingo.

I currently don't have the bootstrap report but will your steps will help me a great deal on where to start.

I will report back with a status update.

Well - If you know at least the tape where your latest Bootstrap has been stored, you need to scan it with

  scanner -Bv device

Do not forget the 'v' - it could well be that the latest bootstrap is 'corrupt'. Then you have to go for an earlier one.

With '-v' you get them all.

Before you start scanning let me suggest you ask the admins onsite who might have received a bootstrap report. Usually, the bootstrap notification will be configured to send this info by mail.

@bingo  Yeah, the former sysadmin left the company but I have his email address and will ask him if he was getting the daily emails with the bootstrap reports.  I know that emails were sent when the backups failed.  This exercise has pointed out obvious flaws on the way my company has the backups setup in case we need to recover the backup server.  ;-)

@Bill  That's a good idea.  I think I will get the backups working again and then focus on restoring the historical backups later.

On that topic, while I was trying to do some test backups on this newly rebuilt server, it kept failing with the error "no clientid found".  I'm sure it's a setup issue..

Thank you both for your help so far.

Jntexas,

Basically yes – as you don’t have any info for the Bootstraps or the tapes as it’s been lost this is what you can do, or you can decide to reconfigure everything from scratch and forget about historical backups.

Typically these remote office set ups are small therefore if historical backup data is not an issue you can rebuild, get the thing up and running and then maybe scan the tapes in later to see what is on them You would have to put some new tapes in for the backups starting now of course.

You could try scanning in the highest slot number first as NW would typically have most recent backups on that volume providing there is not more than 1 pool, if there was then it is truly “pick a number time”

Regards,

Bill Mason

This might be a rhetorical question but if I don't know which tape has the latest Bootstrap, then I need to scan all the tapes that are in the library?

If so, is there anything I can do to speed up the scanning process?

bingo wrote: "no clientid found" - this is weird as NW assigns the clientid automatically whenever it creates a client resource. What exactly did you do?

I had this recently due to mistake where new storage node was placed into production and went live prior to setting backup server interface for correct VLAN. This caused devices to go on and off and some clients to report this error though I fail to see why.  What I ended up doing was set communication path correctly, stop NW, remove nsrladb on on storage node, start it again and all was ok.  The issue as I could see later was that due to this storage node didn't create NSR peer record at all on server and it seems like this was causing the issue.

"no clientid found" - this is weird as NW assigns the clientid automatically whenever it creates a client resource.

What exactly did you do?

@bingo  When I configured the Windows client, the 'clientid' field is populated but it's when it's when I try and run a backup job that fails with the "no clientid found" error in the details page and logs.

Unfortunately, I have another problem with the library now.  This office can't catch a break. 

Stay tuned...

Well it seems that Hrvoje got the answer. shutdown the services and move /nsr/res/nsrladb.

Then start NetWorker again.

I never verified that hut I trust Hrvoje :-)