[Webinar] Streamline your web hosting managementRegister Today

x
?
Solved

Dell Powervault 3600f SAN configuration

Posted on 2013-12-17
7
Medium Priority
?
1,032 Views
Last Modified: 2016-11-23
Hello

We have a Dell 3600f SAN with dual controllers.
Additionally 6x wintel servers each with HBA with single fibre connection.

We started with a single server attached to controller 0 and all works fine, all LUNS are seen.  Afterwards a second server was attached to controller 1 and LUNs all seen apart from the first server that then lost all LUNs.  Ownwership/preferred path also moved to controller 1.

This is bad as I need all LUNs available on all directly attached servers on either controller.

According to the Dell info we can directly attach up to 4 servers in HA mode or 8 servers without HA.  I can only presume we are currently running HA (have no idea how to check this) and need to convert it.

The first server on controller 0 is running live data therefore have to be careful not to cause further problems.

Any ideas how to deal with this one?  We cannot contact Dell as the support contract is with another organisation that is struggling to find an answer too.

Thank you.
0
Comment
Question by:nmxsupport
  • 4
  • 2
7 Comments
 
LVL 17

Expert Comment

by:Andres Perales
ID: 39724228
You are the customer, you should be able to visit the Dell website and ask for support on your hardware!

Here is a starting point - http://www.dell.com/support/my-support/us/en/04/product-support/product/powervault-md3600f after the page displays click on Manuals to expand that sections there should be something to help you there!

You can also go to support.dell.com and enter your arrays service tag for more specific information.
0
 
LVL 56

Expert Comment

by:Handy Holder
ID: 39724330
You cannot have two servers connected to the same LUN without clustering software, at least not using NTFS since the master file table is cached in RAM and there is no file locking mechanism. Data corruption is pretty well instant since each overwrite the others data, because it doesn't know it's there since it's not in it's local copy of the MFT.
0
 
LVL 56

Expert Comment

by:Handy Holder
ID: 39724376
Additionally there are only 4 host ports per controller and you have 6 hosts, so you can't have HA even if you slice a LUN off for each server unless you but a pair of FC switches and get second HBAs for each server (or buy 3 dual channel ones so that 3 servers have 2 single channel HBAs and 3 have the new dual port ones).

How to deal with it? Get someone else to design the solution most likely since you don't have the right kit or software at the moment. What are you trying to achieve with your setup?
0
The 14th Annual Expert Award Winners

The results are in! Meet the top members of our 2017 Expert Awards. Congratulations to all who qualified!

 

Author Comment

by:nmxsupport
ID: 39724754
Thanks for comments- cannot go to Dell as was purchased through another agent and support has to go through them and they don't do config assistance -- long story!

Fibre switch and dual HBAs are out for the time being due to cost.

Dell site states that 8 servers can be connected in "non HA" mode.  Can you get this link if you need it.

Cannot find anything in the manuals.
 
These are 6x Hyper-V 2012 R2 clustered servers and LUNs will be shared using CSV.  We don't require HA at the moment as happy to run off a single controller/half the cluster nodes should the worst happen.

Just need to work out how turn HA on/off and enable 8 directly attached servers as Dell seem to indicate we can.
0
 
LVL 56

Accepted Solution

by:
Handy Holder earned 2000 total points
ID: 39724819
Once you have Hyper-V / CSV setup you won't get the corruptions - you'll have clustering support. With CSV one host is effectively the owner and only lets a single other host access a particular VHD file at any one time

You can have 4 hosts per controller and LUNs are owned by one controller or the other - until failover occurs due to hardware failure. You could have 2 clusters of 3 nodes each, one with the LUNs and hosts connected to controller A and the other cluster all connected to controller B. That would maximise your host usage but you would not have HA on the SAN side since the hosts would only be connected to one controller.

You can have 4 hosts connected to a single controller for a 4 node cluster, but again you could not have HA on the SAN controller side although if one failed you could re-patch the fibre manually to the other.

You could have a fully redundant 3 node cluster by putting the HBAs in 3 machines so they could connect to both controllers.

You don't turn HA on/off on the SAN, it remains on but you set the ownership of the LUNs all to one controller and don't connect anything to the other controller, it's just backup for the controller's write cache. You have effectively turned it off by not presenting the LUN to anything on the other controller.

You can have a 4 node cluster plus a separate 2 node one of course but they would have separate LUNs since a LUN is not presented from both controllers at the same time, which is why you lost connection to the storage from your first server when you connected the second one to the second controller.

I'd be tempted to max out 4 of the servers with RAM from the other two and sell them to buy another couple of HBAs and go for a 4 node HA cluster.
0
 
LVL 56

Expert Comment

by:Handy Holder
ID: 39725348
As far as Dell's claim that 8 servers can be connected in non-HA mode it's true, but they cannot share the same LUN.
0
 

Author Comment

by:nmxsupport
ID: 39737999
Points understood.  We are using all nodes on a single controller now and due to limitations have gone ahead with 2x Brocade silkworm 200e fibre switches.
Thanks!
0

Featured Post

The 14th Annual Expert Award Winners

The results are in! Meet the top members of our 2017 Expert Awards. Congratulations to all who qualified!

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Among the most obnoxious of Exchange errors is error 1216 – Attached Database Mismatch error of the Jet Database Engine. When faced with this error, users may have to suffer from mailbox inaccessibility and in worst situations, permanent data loss.
Organisation is organized in a pattern to flow the day to day business, every application and system is interdepended on each other and when very important “Exchange Server downtime” happened.
This tutorial will walk an individual through the process of installing the necessary services and then configuring a Windows Server 2012 system as an iSCSI target. To install the necessary roles, go to Server Manager, and select Add Roles and Featu…
This Micro Tutorial will teach you how to reformat your flash drive. Sometimes your flash drive may have issues carrying files so this will completely restore it to manufacturing settings. Make sure to backup all files before reformatting. This w…
Suggested Courses

612 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question