08-12-2011 03:19 PM
We currently have 2 HP DL380 servers running ESXi 4.1 and they are connected to an HP P2000 G3 FC/iSCSI SAN. The connection is through dual FC controllers on the HP2000 to dual port HBAs on the two servers. Each server connects to a each of the controllers for fault tolerance.
We have 25 VMs running on the 2 hosts (DL380s) and use vCenter to enable a single cluster and HA (High Availability). Data stores are vmfs3.
We now want to grow so we purchased another DL380 server and another HP P2000 SAN. Now we wouldn't have enough ports to tie this all together so we bought a HP Storageworks 8/8 SAN Switch (Brocade 300).
1. For such a simple setup with a single switch I probably want to use EZSwitchSetup and use switch mode and not Access Gateway mode is that correct?
2. Will I be able to connect FC cables from the existing servers and the SAN FC controllers into this switch and everything will continue to work? Or am I dreaming?
3. If the answer to question 2 above is no, is it fairly straight forward to get this configured so it will work?
Thanks so much for any assistance you can provide.
08-12-2011 10:16 PM
1) You should run the Brocade 300 in Switch mode correct. EZSwitchSetup is a tool to help you to set it up. If you read the admin guide you will be able to configure it on your own. Ask your HP reseller if they can help with the base setup. You can ask questions here in the forum as well.
2) I recommend that you buy new cables with 50um 125 in OM3 quality if you use 8Gbit. If your current cables are OM1 cables please replace them. You have to configure your switch before you plug the cables into the switch. Basicly you need a zoning inplace.
Create for each device an alias based on the WWPN. Create for each server <--> storage connection a single zone and don't forget to activate your zoning configuration.
3) Yes it is straight forward.
Please considre following that you create a redundancy on server level but you have a single point of failure. It is your switch with a single power supply.
I would recommend that you buy a second Brocade 300 and create two independend fabrics (switches are not connected to each other)
If you would like to monitor the switch you should buy the FabricWatch and Brocade Advanced Performance Monitoring lic as well.
I have attached the current admin guide for your reference.
There you find many useful information, please read it it. This will save you much time :-)
If you are happy with this information I appresiate if rate or mark the thred as answered.
In other cases I will try to give some more information.
08-15-2011 10:44 AM
Thanks so much for the response back. I've gone over what you've said and looked over the Guide you attached. So it looks like when I've correctly set up the switch with the zones and activated them, there would not be any configuration necessary on the hosts running ESXi or wihin VMWare vCenter.
Also, if the switch did go down I could plug the cables right back into the controllers from the servers as a temporary measure until the switch was fixed/replaced?
I will propose to management that we purchase another Brocade 300.
08-17-2011 01:07 AM
The problem I would discuss with the management is what happens to the buisiness if all VMs died at the same time and you can not gurantee that the File System is OK.
If you need further infomation please let me know if not please mark or rate the thread.
08-22-2011 04:02 AM
The ports in the P2000 can via the web interface individually be set to either "direct connect" or "to san-switch" mode.
So yes, you should be able to swap the cables. But yes, do think about the likelihood of the filesystem getting corrupted if all the vms drop offline.