vsan ? or other problem ?

[mariangcisco]

Hi, I don't know what is the problem. I have two fabrics covering two sites, IBM FAStT connected to both fabrics on one site, another FAStT connected in the same way on remote site. When I try to setup remote mirroring and I go through fabric A, it is ok. If I go through fabric B, FAStT's don't see each other. The only difference is, fabric A has two 9509s, fabric B has one more 9216 on different site (connected like triangel). But I separated everything on VSAN level. Ports from FAStT are in VSAN 13

This is from interface that trunks to 9216

fc2/12 is trunking

Hardware is Fibre Channel, SFP is long wave laser cost reduced

Admin port mode is auto, trunk mode is on

Port mode is TE

Port vsan is 7

Speed is 2 Gbps

Transmit B2B Credit is 255

Receive B2B Credit is 255

Receive data field Size is 2112

Belongs to port-channel 3

FCSP Status: Successfully authenticated

Trunk vsans (admin allowed and active) (1,7)

Trunk vsans (up) (1,7)

Trunk vsans (isolated) ()

Trunk vsans (initializing) ()

and this from interface connected do opposite 9509 (and this connection should carry trafic for all vsans, including 13)

fc1/14 is trunking

Hardware is Fibre Channel, SFP is long wave laser cost reduced

Admin port mode is auto, trunk mode is on

Port mode is TE

Port vsan is 1

Speed is 2 Gbps

Transmit B2B Credit is 255

Receive B2B Credit is 255

Receive data field Size is 2112

Belongs to port-channel 2

FCSP Status: Successfully authenticated

Trunk vsans (admin allowed and active) (1,3,5,7,9,11,13,15)

Trunk vsans (up) (1,3,5,7,9,11,13,15)

Trunk vsans (isolated) ()

Trunk vsans (initializing) ()

The same for all interfaces in portchannel.

As you see, fc2/12 is only for vsan 7 and it even belongs to vsan 7. fc1/14 is for all vsans, so it belongs to vsan 1. Is that ok ? Do I misunderstand something ?

Please help

Thank you

[mariangcisco]

Fabric OS is 2.0.2b

[mrfrase]

I don't get the complete picture of network and what your trying to do with the design, i do see that your TE EISL's are in different VSAN's as you mention, you can place them all in VSAN 1 for TE's connect to all VSAN's and you control with trunking configuration what VSAN's can connect across TE ISL, this may be the issue for I have not designed TE's in this manor before.

[mariangcisco]

May I contact you directly ? I would email you complete picture then.

[tblancha]

Do a show zoneset active and verify that the devices you want to see have a * next to them in the zones that permit the initiators and the FASTT's to be in the same zone. Look especially hard at the vsan that is not working and verify there is a show flogi database entry for the init and storage you think are supposed to be seeing each other. Then issue the show zoneset active and verify there is a active zone that has those devices and they are logged into the fabric's nameserver. If you need to, create a TAC case for the issue if it goes beyond the scope of the forum.

[mariangcisco]

Yes, that was first what I checked :)

This is from working fabric

zone name FAStTC2-b-FAStTK2-b vsan 12

* fcid 0x1a0100 [pwwn 20:04:00:a0:b8:12:a2:65]

* fcid 0x420100 [pwwn 20:02:00:a0:b8:16:52:40]

(no other zones there, no other ports in VSAN)

and this is from failing one

zone name FAStTC2-d-FAStTK2-d vsan 13

* fcid 0x2e0100 [pwwn 20:05:00:a0:b8:12:a2:65]

* fcid 0x560100 [pwwn 20:03:00:a0:b8:16:52:40]

(no other zones there, no other ports in VSAN)

As you can see it looks perfectly reasonable, I checked WWNs they are OK.

Could you confirm that configuring TE interface into specific VSAN is OK ? Or is it bad idea ?

fabric A (good one) is like this

switch A1 -------- switch A2

all vsans are allowed on portchannel

fabric B is like this

switch B1 -------- switch B2

\ /

\ /

switch B3

B1-B2 - interfaces are in vsan 1, all vsans are allowed

B1-B3 and B2-B3 - interface is in vsan 7, vsans 1 and 7 are allowed

vsan 13 is the one that behaves differently in fabric A and B

I would need to check the concept of vsan traffic filtering if I got it right or not.

[mariangcisco]

sorry for formatting issue, switches B1, B2 and B3 make a triangle. Portchannel B1 to B2 contains 8 interfaces, B1 to B3 and B2 to B3 has just one interface right now.

[tblancha]

If the failing vsan only has those devices, you can see if the initator is quering the nameserver. In one telnet session, shut the port to the HBA. In another session enter config mode and then this command:

fcanalyzer local display-filter mdshdr.vsan==0xd limit-captured frames 0

Then do a no shut on the interface-After about 30 seconds you can hit control-C on the trace. At some point in the trace, the HBA should query the nameserver and be told of the storage. If it does, then it now up to the HBA to go and PLOGI into the storage.

[mariangcisco]

I will try that in the morning. In the mean time, can you confirm or deny my understanding of vsan filtering on portchannels ?

Thank you

[mrfrase]

Yes, you can email me the show tech off your switches and I will comment back on this form. mfrase@cisco.com this may warrant a TAC case be open.

[mariangcisco]

OK, sorry for panic, I was at customer today, tried to shut down and then bring up interfaces on both sites and everything is back to normal.

Thank you for your time and your help.