Skip to main content

I have some CIFS Shares on a NAS that I want to backup and I currently do this by using a subclient on a Windows File iDA that simply connects to the UNC path.

I’m replacing the box that the existing subclient sits on so I need to re-create it and I have two options.

  • Same as before but use a new File iDA.
  • Add a NAS/Network Share client and use the new Windows File iDA as a data access node.

The end result is my UNC path is backed up so I’m not entirely clear what the difference is between the two methods and why I might choose one over the other?

sorry (cannot edit my post now): when I said “NDMP” above I meant “NFS share, with UNC being used for the backup config””


I have a similar situation. The previous admin had some things set up as NDMP, but then changed them all over to filesystem subclients…I had a few more questions..

Q1: it looks like the “proxies” have to be MediaAgents?

Q2: Suppose I have an NDMP backups set up (several NDMP shares/subclients), and all the proxies are configured in a group at the NDMP level (global for all subclients). Then I get a new NDMP backup setup that’s very large (a ton of small files/or a lot of data). When a backup of this large share/subclient occurs, does it choose only 1 proxy server from the group, or can I override the global proxy setting and choose 2 or more proxies (or a special group) for just that share and all proxies will take part in backing it up?  Why: Normally we try to have “small standard size vanilla proxies” vs “making one very large and special”

Q3: its mentioned above “you can hang ddbs off the proxy which I would assume is much more robust that standard client dedupe signature generation.”. can you elaborate on what this means?  Currently for our NAS backups we have several separate storage policies tied to DDB’s for these (prob so we don’t have a massive single DB) and we kinda balance out the storage policies by choosing a different one (based on data size) every time we add a new CIFS/NAS backup


I honestly think this is a question for your Tam, different licensing agreements have different caveats.

So your tam should be able to answer the question Re: Licensing


Does anyone have any more thoughts on this please?

Particularly around the license requirements piece as I’m not 100% clear if the NAS client consumes an additional FS iDA license or if it’s just the backend “proxy” that needs the FS iDA?


Ah that makes sense 👍🏻

So the downside is a potential extra license.

The upside is the NAS “object” helps abstract the client from the proxy so future changes might be less painful.

 

Any other pros or cons?


The proxy doesn’t need to have the ida licence active.


Thanks Christopher this is just a dumb CIFS backup so I don’t think intellisnap comes into for what we need it but otherwise I understood it to be exactly as you said it in so much as doing it this way seems to be a bit more future proof.

Downside looks to be that the NAS client uses an extra iDA license even if the Windows proxy already has an iDA assigned to it, unless I’m missing something?


I suspect the biggest differences are that with the psuedoclient you specify proxies and enable intellisnap.

These are pretty big differences in that you can specify multiple proxies and thus split up the load. You can even specify a group that contains proxies and therefore if you ever need additional processing power you can add or remove proxies at will so you will never have to do the exact thing you are now doing. If you need to upgrade the box doing the boxes its not bound to a specific client and the ability to do intellisnap speaks for itself.

also you can hang ddbs off the proxy which I would assume is much more robust that standard client dedupe signature generation.


deferring to my alt account: @christopherlecky 


Reply