2

I am very new to SANs and have inherited an old EMC CX300 to play with. The CX300 is purely FC, but I'd like to experiment with ISCSI as well without having to purchase an expensive ISCSI gateway device.

I was thinking I could do something like the following:

  1. Connect a Win 2012 server directly to the CX300 via FC
  2. Install Windows native ISCSI target software
  3. Serve up LUNs to the server
  4. Point clients to the server via ISCSI to access the LUNs

Would this be an acceptable production solution? I would be using the server like a ISCSI gateway. But I wouldn't want to use Window's horrendous storage pools\RAID solution I would just want to pass clients straight through to the LUNs I served up to the server.

I apologize if this question is obtuse, but as I said I am very new to SANs. I'm not actually deploying this in production, but was curious if setups like this exist or are acceptable for production environments.

4 Answers 4

4

Would this be an acceptable production solution?

No. It would not. You do say that you'd like to "experiment with iSCSI." It is fine for experimentation.

I would be using the server like a ISCSI gateway. But I wouldn't want to use Window's horrendous storage pools\RAID solution I would just want to pass clients straight through to the LUNs I served up to the server.

Storage spaces actually isn't horrible, but in this scenario you'd be best letting the CX handle the RAID.

I apologize if this question is obtuse, but as I said I am very new to SANs. I'm not actually deploying this in production, but was curious if setups like this exist or are acceptable for production environments.

This would not be typical in a production environment. iSCSI and FC are both fine for production, but not what you're proposing. People buy the right solution for their operation. If you need FC you buy FC kit. If you need iSCSI, you buy iSCSI kit. Like I said, this is fine for learning how iSCSI works, but it's nothing like how you would configure storage in production.

1
  • Thanks for the info, I figured it was a hacky way to do things. Maybe horrendous was a bit harsh, but I read a few articles like this that found serious latency issues as well as the fact that there is no way to re-balance data across physical disks in an array when adding a new disk: "arstechnica.com/information-technology/2012/10/… maybe their better with R2.
    – red888
    Oct 15, 2013 at 17:29
3

Would this be an acceptable production solution?

Well that would work but I wouldn't say it was production because the latencies are going to be awful and there's no resilience - so for a lab it'd be fine but I wouldn't say it was even remotely production spec.

2

I've done this as a strictly temporary stop-gap solution. You're dealing with several layers of abstraction with this method (LUN presented to the Windows server, creating a VHD on the Windows server to be the iSCSI target for the clients, etc.) and the performance is not going to be what I would consider "production acceptable".

1

I've done this...

I've used servers running ZFS through NexentaStor to sit in front of legacy fibre SAN storage and as a way to get new life out of older storage arrays. It's a stopgap solution. Fine for testing... fine for certain workloads... But not the right choice for a permanent solution.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .