Off-network catalog restore is failing

[Ted_Richart]

Just a reality check here

I am testing a restore of our catalog to another server, and I'm really getting nowhere

This was my plan:

Put all imortant files on the test server, including last full catalog, DR files, NetBackup 7.6.0.1 and 7.6.0.3 installation files

Remove test server from domain. I also uninstalled a previous NBU enterprise 7.6 installation.

Disable networking and restart

Rename test server to production master server name

Install NBU Enterprise 7.6.0.1, then 7.6.0.3, license; reboot

recover catalog with recovery wizard

check accordingly

My recover is failing (error 20), and EMM won't keep running (error 77)

Tried to import the image, but phase 1 wants to use a storage unit

I'm not even getting the syntax right trying to use bpimport -drfile -id "D:\Backup"

Storage unit won't create without EMM, and so won't list my test server when I try to create disk storage

I'm thinking this is because I'm off network, but I don't see anything in the docs that discusses that.

This is the first of occassional DR tests that I was asked to perform, so I'm pretty green on this topic

Any thoughts?

Any idea what I might need in the way of logs and/or proceedures?

Thanks in advance!

Ted

[INT_RND]

Where is the catalog image? Tape? Disk pool?

You don't need to import the image if you have the DR file. The DR file will describe where the image is. If the image is on disk then you will need to make the disk pool look the same as it's description in the DR file.

Please make sure the devices and media that contain catalog disaster recovery data are available

"Storage unit won't create without EMM, and so won't list my test server when I try to create disk storage"

I don't know what that means. You will need EMM to do anything. That is the master server. If you don't have EMM then you don't have a master server.

[INT_RND]

Also, if you don't have a media server in your DR environment you will have to edit the DR file like this:

http://www.symantec.com/business/support/index?page=content&id=TECH48819

[Ted_Richart]

Good question

I have a copy (first copy, not primary) on a media server in my environment. simple local NTFS

I copied it to the test master server before I disabled the NIC

I don't have any tape drives available to this test server

Also, I can try to employ your suggestion from Tech48819, I didn't know it applied to me when I read it, but it makes sense now

I will be working on this more tomorrow

Thanks

[jim_dalton]

Ted,

My experience tells me the better route to go is to start with a fresh server , fresh install, patched to same Netbackup level as the original mastwer then the new master you are dealing with is self contained and correctly described in terrms of storage devices etc.

A fully functioning master server is the starting point to do a recovery.

This is what INT_RND is saying essentially.

Then do the catalog recovery, then reconfigure out all the storage unit changes made by the recovery so it again matches reality.

This may not be the only way but it works for me and I believe is the recommended way.

Jim

[revarooo · Accepted solution]

If your Master server is not on a network/NIC not configured you may have issues with it running correctly.

NetBackup does not like running on a system where the network is not running or configured.
 

[Marianne]

Just unplug network, don't disable it. This is probably why EMM is not running. It is called NET-backup for a reason! Add hosts entry for server itself and ensure that it can resolve own hostname to ip and ip to hostname. Only attempt catalog restore when NBU starts up successfully.

[Ted_Richart]

About EMM and a disabled NIC; you're right, I should have thought of that

I did put a bogus IP address on it and enabled the NIC with nearly the same results

I then unintalled/reinstalled NBU and I got much further. I had to recreate the DR File (deceptively simple) and the recovery started OK

The next problem was a bit different. I had critical policies defined in my catalog backup policies. This isn't a problem as long as I have access to the media servers and the media that they reference

I found a post about editing the DR file, and I got closer, but I still have an error 20 trying to import the catalog

So I called Symantec

An engineer called me back within an hour (pretty good, considering I set the case up for 24-hour response)

He stated

1) Critical polices aren't useful anymore because it is assumed that you will have what you need in a DR situation. I'm not sure I follow the logic, but he is probably practically correct.

2) If it takes 10 hours to backup my catalog, it will probably take 10 hours to restore it. This is important because part of this quest is a practice run at replacing the master server with new hardware

I am now repeating my test, with a new full catalog backup and no critical policies defined.

I post my findings

Thanks, all!

[Marianne]

Curious to know what happened with the latest testing....

Always best to have copy of the DR-file sent as email attachment and/or saved on a network share.
This will ensure that the DR-file is readily available and there will be need no to recreate it via import.