Options to export/backup from ISILON to seperate SAN?

you are mentioning fiber channel fabric in your "old SAN fabric and Isilon. Isilon does not understand fiber channel unless you purchase a Isilon NDMP accelerator and start doing traditional NDMP backups.  We need to step back and understand exactly what your requirements are. You said backups, what is it that you want to backup (entire cluster, selected folders) , what is your RTO/RPO.

Oops - sorry at the ISILON - SFP 10gb Fiber x4 to backbone (at that point im not sure of the network/backbone design - not privy to that), then locally i've got fiber to local switch to SFP/Fiber down to the local SAN via fiber.

Ahh to backup, specific folders only.

RTO/RPO? it wont be for live re-placement or DR Site of *if* the ISILON fails, but just a safeguard to say: "Sure! we have a 2nd copy".

For now, eventually after this POC I guess I'd call it, a future plan of what/how can be removed from primary data store perhaps to another media, but for now its just the "Sure!...." can it be done comment.

So all in all, POC *if* it could be done, yea that would be great.

_L_

As Dynamox says, an A100 Backup accelerator has 8G Fiber SAN connection to connect to SAN.  Another idea is to use another NL4xx cluster and SyncIQ many:1 cluster (probably easiest route).

sounds like  you just want a "bunker" copy, can't use it to service customers but the data is there. To me this is not practical but you could stand up a Windows box, present fiber channel storage locally and then use robocopy/emcopy to copy the data from Isilon. 100TB is not trivial especially if it's millions of tiny little files, eventually it would get copied.  So you have this copy, and you run robocopy/emcopy every day or every week. You don't have any versioning so you only have the latest version of something, if you need to "restore" 50 terabytes, it might take you days to copy it back to Isilon.  While this is possible not sure how much value it will provide to your organization.  Get your bean counters to invest into a 3 node HD-series cluster for deep archive and SyncIQ/SnapshotIQ licenses

G'Day dynamox,

That pretty much migh be the go I think, robocopy or even the cron job for now till more $$$ could *found*.

Truth be told, we have 4x HD400's plus a myriad of other nodes - but its a *what if???* though by the bosses.

Furthermore, the types and sizes of files are the killer; from 2kb through to 3tb - its just NASTY.

I must admin previously in our early POC we had two sites for over a year and the SyncIQ was just brill!

__L__

G'Day Phil,

Thanks for the reply.....I'd like to have more nodes - but we are already running 4x HD400's + many more nodes.

but bosses being bosses - *what happens if???*.

So, yea....just fleshing out possibilities with the old *other branded company system* and how I could push/pull some data to it.

Thanks for reading and commenting!

_L_

Robocopy/EMCopy, rsync are proven methods to replicate data.  Those may be the best methods for you.

if you do go down this route, you need to set expectations with your management accordingly.  This solution might be ok for restoring a couple of files here and there but if you have complete meltdown, what good is that data if you can't get to it fast enough.

G'Day guys,

Many thanks for the replies and support.

I hope that we'll never need to even look at it again, and yes dynamox expectations *will* be expressed once ive documented whats been set in play.

Talk again in the future....

_L_