SUMMARY: Restoring from tape using Legato's Networker

From: Rasana Atreya (atreya@library.ucsf.edu)
Date: Thu Jan 09 1997 - 15:33:00 CST


Firstly, my grateful thanks to "Jeffrey S. Jewett" <spider@umd5.umd.edu>
who single-handedly helped me through my problems.

My original post.

> Someone blew away a disk (/home), so she had to restore from tape.
> Unfortunately, the last full back was from 11/20/96 (The networker software
> was being upgraded in the middle). We need something more recent. She did
> restore the last level 9 backup, but that did not quite do the trick.
>
> We do have 155 level 9 backups since then. My question is: how do I integrate
> the whole thing, so I can bring the /home up to date. I tried using the save
> set recover feature, but it seemed to just hang.
>
> Networker server: Solaris 2.5.1
> Networker version: 4.2.4 Turbo/30

Well the deal was that I had a few different problems.
From: "Jeffrey S. Jewett" <spider@umd5.umd.edu>

I was trying to use the "saveset" feature to recover. I was under the mistaken
impression that I needed to do 155 individual recovers (for each level 9
backup since my last full backup). Apparently, this is not necessary. Jeff
pointed out that I could use the standard NetWorker recovery interface
"nwrecover", to bring up the GUI and that I should let NetWorker decide what
files need to be reloaded, and from what tapes.

To test this out I tried to recover just my home directory using "nwrecover".
It said "Recovering 5051 files into their original locations Total estimated
disk space needed for recover is 50 MB."

I started this recover on Friday afternoon, and by Monday morning it was still
saying the same thing (it's not completed). This is a 2gig disk which is about
16% full, so space could not be the problem.

Also, it needed 3 tapes. I loaded all three onto my jukebox and did an
inventory. They were recognized. But the pending messages on Netowrker told
me that it was waiting for tape #1. This is inspite of the fact that I had
explicitly mounted this tape.

If I tried to unmount it told me" error, Load operation already in progress
(35019)".

So I bugged Jeff again, and he was kind enough to send a long mail
pointing out a few things:

Since it was so informative, I'm including his entire email (since it fixed
my problem).

My heartfelt thanks to Jeff.

Rasana

Jeff's response:

The symptoms are that NetWorker believes that it has been unsuccessful in
loading the first tape of the sequence.

It sounds confused. The "nsrjb" entry in the "ps" listing shows that
NetWorker is still trying to get the jukebox to load the tape. If you
tried to load the tape explicitly while it was trying then you may have
a problem. Typically in such situations, one of the processes times-out.

I may have neglected to say that "nwrecover" has to be run on the client,
not the server, but I'm sure that you knew that. Were *ANY* files
recovered in the process? Or did it really stall out before the 1st tape
was loaded?

Do: "nsrjb -C" to see what NetWorker thinks is in the jukebox, and loaded
in the drives. Do you have more than one drive? Did you use NetWorker
to load (or attempt to load) the tape into the drive? If the nsrjb -C
shows that Networker thinks that the desired tape (Athena.005) is in one
drive and it's actually in another, then you should move it to the drive
that NetWorker thinks the tape is in.

In any event, the recovery should take at most an hour, and certainly not
several days.

Also, I see that /usr/sbin/nsrexecd is *NOT* running on your server.
Beginning with NetWorker 4.2, nsrexecd has to run on the server as well
as the clients. The syntax is:

        /usr/sbin/nsrexecd -s <server-name>

So, you may want to reset everything. That is, reboot the NetWorker
software.

On the CLIENT:

   Kill the nwrecover run (click on all cancel buttons)

On the SERVER (as soon as the nwrecover messages go away):

   Do "nsr_shutdown" to kill the deamons, wait about 30 seconds, and then do

        /usr/sbin/nsrexecd -s your-server
        /usr/sbin/nsrd

And then, when networker starts functioning (as seen by the nwadmin
or nsrwatch windows), unload all drives, verify that NetWorker knows
(correctly!) where the 3 recovery tapes are), and then re-start the nwrecover
run on the client.

- jeff

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~ Rasana Atreya Voice: (415) 476-3623 ~
~ System Administrator Fax: (415) 476-4653 ~
~ Library & Ctr for Knowledge Mgmt, Univ. of California at San Francisco ~
~ 530 Parnassus Ave, Box 0840, San Francisco, CA 94143-0840 ~
~ Rasana.Atreya@library.ucsf.edu ~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:42 CDT