UNSOLVED

ronald_lijs

updated

15 years ago

R

ronald_lijs

1 Rookie

15 Posts

0

1630

September 29th, 2011 01:00

Client file index ownership question

Hi there,

Just a simple one as I have been unable to make it work on the field.

I've got stora

ge node A defined, which writes to Data Domain devices.

I'd like to backup client B (Windows host), by mounting a snapshot volume onto storage node A and backup that volume from there. This volume is an NTFS volume, nothing particular about it. I just want the data to be indexed against a different client.

I've been trying to use the save -c

> option, but Networker keeps on doing a full only thinkning the data is owned by storage node A.

I have tried also -without success- to use the pathownership variable, but as expected this has not worked.

Is there any way to backup data from client B from a storage node and index that data against client B?

It seems in my case a full is always run, no incremental backups. It seems savefs

42506 09/29/2011 10:37:38 AM  nsrd savegroup info: fcl-bck-p004.freshfieldsbruckhaus.com:E:\Test: No full backups of this save set were found in the media database; performing a full backup

Unforuntatly Networker thinks storage node A is the owner of the data and fires up a full savefs

7236:savegrp: Group will not limit job parallelism

32494:savegrp: fcl-bck-p004.freshfieldsbruckhaus.com:probe   started

savefs -s fcl-bck-p001.freshfieldsbruckhaus.com -c fcl-bck-p004.freshfieldsbruckhaus.com -g DC1DD1_TST_NWC_COR_D_F -p -l full -R -v -F "E:\\Test"

fcl-bck-p004.freshfieldsbruckhaus.com:E:\Test level=full, dn=0, mx=1, vers=pools, p=36

Thanks,

Ronald

  • ble1

    6 Operator

    14354 Posts

    56186 Points

    607

    0

    Posted September 29th, 2011 13:00

    Your backup server is:  fcl-bck-p001.freshfieldsbruckhaus.com

    Your storage node is: fcl-bck-p004.freshfieldsbruckhaus.com

    Is this correct?  If in backup command you place save -L -c ..freshfieldsbruckhaus.com (which I assume is client B in this case), does that work?

    If not, try the same from command line first and then check mminfo and nsrinfo outputs to see if it worked.   Then we can go back to server initiated backup.

  • ronald_lijs

    1 Rookie

    15 Posts

    607

    0

    Posted September 30th, 2011 04:00

    Thanks Hrvoje, hope you are doing fine...

    fcl-bck-p001: server
    fcl-bck-p504: storage nodes
    fcl-dpa-v501: client

    Runing save from the storage node, i.e. a manual save works fine, it only backups what has changed (incremental):

    C:\Install\Networker\test>save -s fcl-bck-p001 -b dc2dd1backup -L -c fcl-dpa-v501.freshfieldsbruckhaus.com -t "09/30/2011 11:54:19" "C:\Install\Networker\test"
    57777:save:Multiple client instances of fcl-dpa-v501.freshfieldsbruckhaus.com, u
    sing the first entry
    75146:save: Saving files modified since Fri Sep 30 11:54:19 2011

    save: C:\Install\Networker\test  0 KB 00:00:01      0 files
    7167:save: save completion time: 30/09/2011 11:57:16

    Daemon.log shows:

    70896 09/30/2011 12:54:19 PM  nsrd fcl-dpa-v501.freshfieldsbruckhaus.com:C:\Install\Networker\test saving to pool 'dc2dd1backup' (dc2dd1backup.002)
    71659 09/30/2011 12:54:19 PM  nsrd fcl-dpa-v501.freshfieldsbruckhaus.com:C:\Install\Networker\test done saving to pool 'dc2dd1backup' (dc2dd1backup.002) 3337 KB


    A server initiated backup always does a full :-( And it thinks it has to index the files against the storage node and not the client (look at the first line "info: ".

    42506 09/30/2011 01:04:23 PM  nsrd savegroup info: fcl-bck-p504.freshfieldsbruckhaus.com:C:\Install\Networker\test: No full backups of this save set were found in the media database; performing a full backup
    70896 09/30/2011 01:04:39 PM  nsrd fcl-dpa-v501.freshfieldsbruckhaus.com:C:\Install\Networker\test saving to pool 'dc2dd1backup' (dc2dd1backup.002)
    71659 09/30/2011 01:04:40 PM  nsrd fcl-dpa-v501.freshfieldsbruckhaus.com:C:\Install\Networker\test done saving to pool 'dc2dd1backup' (dc2dd1backup.002) 6670 KB

    I've tried changing the level and the schedule on the group without luck! I forced an incremental from the command line and same thing. It still thinks it needs to index the files against the storage node:

    [root@fcl-bck-p001 ~]# savegrp -vv -I -l incr DC2DD1_TEST_GROUP
    32451:savegrp: fcl-bck-p504.freshfieldsbruckhaus.com:C:\Install\Networker\test level=incr
    7236:savegrp: Group will not limit job parallelism
    32494:savegrp: fcl-bck-p504.freshfieldsbruckhaus.com:probe   started
    savefs -s fcl-bck-p001.freshfieldsbruckhaus.com -c fcl-bck-p504.freshfieldsbruckhaus.com -g DC2DD1_TEST_GROUP -p -l full -R -v -F "C:\\Install\\Networker\\test"
    fcl-bck-p504.freshfieldsbruckhaus.com:C:\Install\Networker\test level=full, dn=0, mx=1, vers=pools, p=36
    77567:savegrp: DC2DD1_TEST_GROUP * fcl-bck-p504.freshfieldsbruckhaus.com:Probe  See the file /nsr/tmp/sg/DC2DD1_TEST_GROUP/sso.fcl-bck-p504.freshfieldsbruckhau.ELDqA2 for output of save command.
    7341:savegrp: fcl-bck-p504.freshfieldsbruckhaus.com:Probe succeeded.
    32494:savegrp: fcl-bck-p504.freshfieldsbruckhaus.com:C:\Install\Networker\test started
    save -L -c fcl-dpa-v501.freshfieldsbruckhaus.com -s fcl-bck-p001.freshfieldsbruckhaus.com -g DC2DD1_TEST_GROUP -LL -m fcl-bck-p504.freshfieldsbruckhaus.com -a "device interface=data domain" -l full -W 78 -N "C:\\Install\\Networker\\test" "C:\\Install\\Networker\\test"
    ...

    But it spawns a full as it tries to check for existing backups of the storage node instead of the client:

    nsrd savegroup info: fcl-bck-p504.freshfieldsbruckhaus.com:C:\Install\Networker\test: No full backups of this save set were found in the media database; performing a full backup


    This is pissing me off, looks like a bug to me! I'm running 7.6.2.1 Build 638.

    Thanks for your help, this is basic functionality and I can't make this work now :-(

    Thanks!
    R

  • ronald_lijs

    1 Rookie

    15 Posts

    607

    0

    Posted September 30th, 2011 04:00

    Backup command is set to:

    save -L -c fcl-dpa-v501.freshfieldsbruckhaus.com

  • ble1

    6 Operator

    14354 Posts

    56186 Points

    607

    0

    Posted September 30th, 2011 07:00

    Does root on storage node have permissions to touch client's index file?  It should work though as I use something similar, just automated with PowerSnap (in such case, you must also on storage node's servers file place client to make it work as by masking data you get this sort of logical relationship).

  • ronald_lijs

    1 Rookie

    15 Posts

    607

    0

    Posted October 7th, 2011 00:00

    Hi again, yes servers file is empty, so no problems there.

    Can you have a look maybe there is an LGTsc or similar for this? I have not yet tested on older versions of Networker... only 7.6.2

    Thanks,

    R