NB6.5.4 linux client won't restore dir/files

[nb-nb]

I am using NB6.5.4 client on RHES5.7.  My backups appear to work okay, however in testing a simple directory/files restore via 'bp' command the backup fails.  No errors given, the NB6.5.6 server sees the restore request but nothing happens.  My log file on the client states restore started from backup (date), then it says "TAR STARTED" and never says anything else.

Does anyone know what is going on here?

[mph999]

6.5.4 client should be fine on 6.5.6 master, providing the media server is at 6.5.4 - 6.5.6

It could be worth upgrading the client to 6.5.6 - yes I know it should work, but just very occassionally we see issues that are resolved with an upgrade.

If you are unable, or that doesn't work, can you get these logs

Media server :
bpbrm
bptm

Client :
bpcd
tar

Master :
bprd

Thanks,

Martin

[Marianne]

First step is to see if restore request is received by the master.

We need bprd log on the master. If this folder does not exist, please create it and restart NBU.
Retry the restore and post the bprd log as attachment.
Depending on what we see in this log, we will request some or all of the logs as mentioned by Martin.

[Nicolai]

You mention you start the restore by the "bp" command - Is the command bprestore ?

If yes, then add "-L /tmp/restore.txt" to the command. This will log result from the restore in the text file restore.txt. 

[Pritesh_Pisal]

Hi,

Please mention the master server's OS.

also mention the bprestore command's output as mention by Nicolai

[nb-nb]

Thank you for the responses.  I am gathering the requested information.  I did do an updated to NB_CLT_6.5.6, but that didn't make any difference.  I will add more information and the available requested log files soon.

[Marianne]

As a start, please post bprd log as attachment.

We will request more logs based on what we see in bprd.

[nb-nb]

Here are two text files.  One is the bp job output log and the other is a text file of the bp command running in my processes.

To answer the inquiry about how I'm doing the restore:

/usr/openv/netbackup/bin/bp

This provides a menu driven interface in which I select to restore directories and files, then I select my directory and tell it to initiate the backup which it says it does.

Other inquiries are being worked and I will post when I get the answers from my NB Admin.

[nb-nb]

As soon as I can get a copy of this I will post it.  I do not have direct access to this log and have requested a copy from my Admin person.

[Nathan_Kippen]

bprd log from master would still be helpful.

[nb-nb]

Here are the bpbrm and bptm from the Media Server and the bprd from our Master Server.

Our Master and Media Servers are installed and running on Windows 2003 Server OS.

Note:  My admin didn't put the correct names on the files, but I'm sure you can figure out which is which.

Thank you for all your help and support, I do appreciate it.

[J_H_Is_gone]

 have you check in the master server properties that clients are allowed to do restores?

Have you evern done a restore before - starting it from the client?

[nb-nb]

Master Server policies allow client restores.

Client restores have been done before.  (Not my clients.)

My computers are on different networks and different domains.  All necessary ports are open.  I do get a proper response to 'bpclntcmd -pn'.

Original errors before an initial backup could be made were 58 and 48.  But I have worked through these and have gotten what shows as a successful backup.

As a side note, last year we had a Windows system that had shown being successfully backed up, however when the system died and I need a BMR restore done, it failed; then when I wanted just the files restored all I ever got was the directory structure and a place holder for the files, no files were recovered.  In other words I would get a disc from our NB admin that showed 4GB of data, but all that was there was the directory structure and emtpy folders, but it physically took up 4GB of space.

[Marianne · Accepted solution]

I have spent some time 'digging' through the logs - a bit of a challenge as the files were not renamed to reflect the prcess name.

I found the following in bprd log:

file_path = /usr/local/downloads/

Because of FORCE_RESTORE_MEDIA_SERVER, changing server of bid multilink1_1333642884 from dfrcmedia02 to dfrcveritas01

Restoring from image = multilink1_1333642884

Not quite sure what to make of the following (seems like resource request failure?):

07:11:03.057 [5724.1012] <2> requestFailed: got gotCallback, jmJobStatus = [2075], emmStatus = [2005075], mapped failure status to= [2075]
07:11:05.103 [5724.1012] <16> RequestInitialResources: MultiResReq.cpp:2320 resource request failed [2075]
07:11:05.103 [5724.1012] <2> RequestInitialResources: returning
07:11:05.103 [5724.1012] <2> allocate_fat_media: Basic disk allocation for FAT failed, continuing restore without FAT.

At this point the restore should be seen in Activity Monitor. Please confirm?

Process carries on and bpbrm is started on media server :

07:11:05.103 [5724.1012] <2> start_brm: backup restore server = dfrcveritas01

07:11:05.165 [5724.1012] <2> start_brm: /usr/openv/netbackup/bin/bpbrm bpbrm -restore -mt 0 -k 0 -c multilink1 -hostname multilink1 -rclnt multilink1 -rclnthostname multilink1 -browse_clnt multilink1 -ru root -firstblk 4578306 -cl GTR_RHEL -bt 1333642884 -st 4 -secure 1 -b multilink1_1333642884 -cn 1 -rst 1334153460 -to 0 -mud 180 -ct 0 -S dfrcveritas01 -masterversion 650000 -backup_copy 0 -latest_image -flport 0 -flipc 1855 -p G:\ -rg root -L /root/bplog.rest.003 -clnt_lc_messages C -clnt_lc_time C -clnt_lc_ctype C -clnt_lc_collate C -clnt_lc_numeric C -connect_options 0x01030202 -jobid 119788 -restoreid 119788.001 -job_total 1

Now this is where I have another problem - the bpbrm log that you have attached does not seem to belong to dfrcveritas01.
There is no mention of any restore instruction for this client... or any restore that was started at 07:11.

Same problem with bptm - seems to be from incorrect media server.

We can see in bprd log that Job id was 119788.  What was the status of this job?

Please locate correct bpbrm and bptm logs.

The piece in bprd that I am confused about seems to look for Basic Disk as resource - is this right?
Was backup done to Basic Disk on dfrcmedia02?
Why then FORCE_RESTORE_MEDIA_SERVER to dfrcveritas01?
 

[nb-nb]

Hello all,

Thank you for all your comments, this issue is now resolved.  I wish I could tell you exactly what was done to fix it but I'm not sure.  My NB Admin didn't share with me what he did to make it work, but whatever he did came from your responses, so I thank you very much for your help.

I do know one change he said he made and this probably isn't stated correctly but it's what he said; that he changed an option in NB server so that all restores come from tape archvie and not from the current Basic Disk.

Thank you again.