Networker

Re: [Networker] Question regarding NetWorker's recovery strategy

2013-07-23 16:34:38
Subject: Re: [Networker] Question regarding NetWorker's recovery strategy
From: "Clark, Patricia A." <clarkpa AT ORNL DOT GOV>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Tue, 23 Jul 2013 16:34:32 -0400
On 7/23/13 4:05 PM, "Tim Mooney" <Tim.Mooney AT NDSU DOT EDU> wrote:


>In regard to: [Networker] Question regarding NetWorker's recovery
>strategy,...:
>
>> I'm not sure how to phrase this question, but my experience with NW has
>> always been host based. In other words, if you want to recover a file
>> from some host from long ago, and that host no longer exists, then you
>> have to clown around a little bit to recover it to another current host.
>> In other words, having the physical tape, a drive that can read it, the
>> NetWorker software and knowing the ssid for the save set is not enough.
>> It seems there was always some mischief with having to create an NSR
>> client resource having that same client name and clientid in order to
>> recover it, maybe even to another host? Is that right? I'm not talking
>> about a browsable recovery. For that, I believe what I mentioned is the
>> case. Instead, I'm only talking about a save set recovery, so I may have
>> my information wrong there?
>
>Pretty much all true, yes.
>
>> Here's my question: I want to be able to back up specific directories
>>on a 
>> given client, and I don't care about the client.
>
>This doesn't fit NetWorker's security model at all, so you're going to
>have a difficult time here.
>
>Think about it: in the vast majority of the cases, admins actually *do*
>care about what host the data is on.  /var/lib/mysql on hostA isn't
>generally going to be the same as /var/lib/mysql on hostB, and unless the
>NetWorker admin specifically configures it, hostB should not be able to
>recover hostA's data.
>
>> In other words, even though NW will
>> record the client name in the media database, is there a way to recover
>>the 
>> save set (save set recover, not browsable recover) without having to
>>know 
>> anything about the client that backed it up (particularly in cases
>>where the 
>> client is long gone) or having to play any games with making NW think
>>there's 
>> such a client or named resource or having to know the clientid?
>
>I've never tried to do this (again, you're really off the beaten path
>here) but it's possible that you could skip the client creation process
>entirely and just use uasm, possibly in combination with dd or some other
>command.
>
>Tim
>-- 
>Tim Mooney                                             Tim.Mooney AT ndsu DOT 
>edu
>Enterprise Computing & Infrastructure                  701-231-1076
>(Voice)
>Room 242-J6, IACC Building                             701-231-8541 (Fax)
>North Dakota State University, Fargo, ND 58105-5164


I find this to be typical for storage devices where you don't want to do
an NDMP type backup and don't really want a client per se as the owner.
While not ideal, my networker server used NFS mounts and acted as the
client for the backups.  My server was not too busy to successfully manage
it all and it kept the management of the data under my control including
upgrades and migrations.  If not your actual backup server, perhaps a
dedicated client that can be a surrogate?  Possibly a virtual machine.

Patti Clark
Linux System Administrator
Research and Development Systems Support Oak Ridge National Laboratory