This post is more than 5 years old
76 Posts
0
1471
Discover DMX4 from HP Blade Virtual Connect passthru
Trying to launch an RHEL5 Blade server with four Emulex FC adapters in the server, connected to the Blade enclosure's 4GB VirtualConnect (VC) pass through. The switch sees the VC with no problem, and I created a fabric and zoneset for the VC. Now stuck on how to get the actual FCs in the Blade server to see the SAN.
FYI.. from one of the switches:
zone name CSF_BLADE_ENC2_2_DMX4 vsan 202
* fcid 0x4a0004 [pwwn 50:06:04:8a:d5:31:ee:27] [Symm5528_FA8c1]
* fcid 0x4a0014 [pwwn 20:00:f4:ce:46:ae:5a:b2] <=== this is the VC on the Blade enclosure
Director Identification : FA-8C
Director Port : 1
User-generated Logged On
Identifier Type Node Name Port Name FCID In Fabric
---------------- ----- --------------------------------- ------ ------ ------
10000000c95ef8b9 Fibre buttercup fcs1 4a0006 Yes Yes
10000000c9961492 Fibre csf-tan fcs1 4a0013 Yes Yes
10000000c99ea53c Fibre sp-sql-prod Slot2 4a0010 Yes Yes
2000f4ce46ae5ab2 Fibre NULL NULL 4a0014 No Yes <===== and here it is (of course not logged in)
Not sure where to go from here. Thanks.
RobertDudley
448 Posts
0
August 8th, 2012 07:00
You will only see it if the blade has been activated (powered on) and you will always get the WWN's on module 1 first. We have the C7000 chassis and with the two SAN ports per module the chassis does a round robin balance between the modules.
There are two ways the WWN's are generated. There is a check box to use HP generated WWN's out of a range they own or let the chassis use the default WWN on the blade itself. We let the chassis generate the HP WWN's so they are always in the range I noted.
If the blade is configured and powered on you can check the switch port to see what WWN's are there. You need to have NPIV support on (node port virtualization) so you can have multiple WWN's on the same san port. Our SAN is brocade and NPIV is on by default on brocade.
RobertDudley
448 Posts
0
August 8th, 2012 06:00
On the HP blade chassis you do not zone the VC module parent WWN. There should be WWN's assigned to the blade itself. if you are using the HP default WWN range they should start like this 50:06:0B: . The chassis uses FCoE from the blade to the virtual connect module, module 1 always comes up first. There is an internal cross so in my setup module 1 has two SAN connections to the odd fabric and module 2 has two SAN connections to the even fabric.
markdf
76 Posts
0
August 8th, 2012 07:00
Okay, so I can undo the zone to the VC parent WWN. However, on the switch, should I be seeing the virtual passthru WWNs (i.e. 50:06:0B..) sitting in VSAN 01? If so, I do not.
markdf
76 Posts
0
August 8th, 2012 09:00
NPIV was not enabled on the switches (MDS 922i). As soon as I enabled it, I can see the WWNs from the enclosure and the blade itself.
Both were in the default VSAN on each switch. Now I can create the zone/fabric using the virtual WWNs.
Thank you!
RobertDudley
448 Posts
0
August 10th, 2012 10:00
Thanks for the feedback good to know I was able to help you.
markdf
76 Posts
0
September 18th, 2012 17:00
So I hate to revive a dead thread, but now we're presented with another problem. I got the first Blade connected to the SAN via the HP VC and the Cisco 9222i router, but now we're launching another Blade using the VC, and I cannot see it in either switch. NPIV is enabled.
One thing I'm not sure about though, which might be a configuration error on my part.. but the VC is also sitting in the same VSAN as the first blade. I think it's a result of when originally configuring, I moved the entire FC into that vsan.. i.e.
1. VC physically wired to fc1/13. New Blade installed.
2. "show fcns database" correctly shows the WWN for the VC and the first blade sitting in VSAN 1.
3. In attempting to create the fabric, I moved the entire FC to vsan 202 (where everything else is).. "vsan 202 interface fc1/13". This moved the original Blade's HBA into VSAN 202 AND the VC's WWN. I think this is where I messed up, because based on what you said earlier, the VC should always be in VSAN 1. Yet, I'm not sure how to just move the Blade's HBAs into VSAN 202, leaving the VC in VSAN 1, and not the entire port (which would include the VC).
If I'm being confusing, let me know and I'll try to break it down. Basically, I don't see the second Blade's HBA cards (either in VSAN 1 or VSAN 202), which is also going through the HP Virtual Connect.
thanks.
-mark
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
September 18th, 2012 21:00
while i don't have VC in my C7000 C-Class HP enclosures, we use Cisco 9124e in NPV mode and 9513 as the core switch with NPIV enabled. I always put my connections from the enclosure switches to be in the same VSAN as my production fabric, you will see your switch login first and then you see your blade servers login as well after they have been powered on.
markdf
76 Posts
0
September 24th, 2012 12:00
I'm sorry. I should have updated this thread. The Emulex Bios was disabled on the boards themselves. I temporarily turned them by entering the Emulex bios during the post, and I was able to see the cards in the switch. I disabled them again, and let the OS (RHEL) enable them and I was able to see the cards in the switch and subsequently create the fabric for them. My solution to this problem is to first install the OS before checking to see if the cards are logged into the switch. I had assumed that once power was given to the box, the cards should automatically light up and connect to the switch, but this time I had to install the OS first (or enable the Emulex bios during post).
Thanks for your suggestions!
RobertDudley
448 Posts
0
September 24th, 2012 12:00
How many SAN connections do you have per VC module? We have 2 per module (4 total per chassis) so the chassis does an automatic round robin balance; first blade is on 1st connection, second on #2, third blade back to 1st connection etc. If you have more than 1 FC connection per module the second one will come out on fiber connection 2. Just pointing this out in case you are looking at the wrong switch port for the login.
If you only have 1 then it should be logging in on the same port as the first connection.