HPUX an EMC clarion

[Sreejith Kaliyam]

Hi all,

We have one rp5470 running on HPUX 11.00 attached to clarion box [new to me :-(] through brocade 2800.There are two luns and presented to this server.I can access these disks (luns) only through one path.ioscan(attached) shows CX500WDUNB as descr for one path.There are many servers using this clarion box and same san setup,In all machine we could add alternate PV links.Can any body help us??

Can anybody make sense what is CX500WDUNB and
CX500WR5 diff.

Thanks in advance

Regards,
Anoop

[Steven E. Protter]

For alternate links to work the san admin needs to have the ownership of the LUNS available to two World Wide Names(WWN).

This is needed for alternate links to work.

Commonly one owner is assigned and that won't work for alternate links.

SEP

[Sreejith Kaliyam]

Hi Stev,

Thanks for your reply.I cant use this device files for extending my VGs only one particualr machine.But all other machine I was able to extend my VGS with alternate PV links.

could please explain little more..!!

Regards,
Anoop

[Tom Geudens]

Neil,
A ClarIIon (which type btw ?) usually has an SP/A and an SP/B. Each of these has several ports. Now, to "see" alternate links, you need a zone (or physical connection) from your server to an SP/A port and a zone (or physical connection) from your server to an SP/B port.

Hope this helps,
Regards,
Tom Geudens

[Sreejith Kaliyam]

Hi Tom,

Thanks for your help.As you can see my ioscan out as attached,I have created zone and hpux see the hardware.But through the second path the luns are available CX500WDUNB instead of CX500WR5,Which is not allowing me to get the disk details (diskinfo) or extending the vg with additional PV liinks

I hope this clarify my issue. little further.

Regards,

Anoop

[Devesh Pant_1]

Neil,

to me this looks like a zoning problem since I have Clariion connected to my HPUX11.11 servers but I don't see any CX500 stuff and I have more than 30 servers connected. This is some other storage that is in the SAN and it's ports have accidentally made it to this server's zone. Please verify the zoning on the SAN switches and after correcting rescan the devices.
I am pretty sure nothing is wrong on your server side.

thanks
Devesh

[Devesh Pant_1]

Neil,
do you see this CX500 stuff on your other servers too ?
I think you won't see that.
Clariion for all purposes I have seen shows up as EMC SYMMETRIX
It tells me that something else either another servers' port has made it into the zone of this server
DGC in your ioscan output somehow tells me it has to do with Veritas on another server that has got zoned into this one.

Please let me know what you found

thanks
Devesh

[Sreejith Kaliyam]

Hi Devesh,

I have one EMC SYMETRX BOX and clarion box both connected to same SAN switch.All working servers ioscan shows as CX500WR5 for both path .Please find ioscan from different server which is connected to same clarion box.

Regards,
Anoop

[D Block 2]

Neil, what about your Brocade 2800 ? Can you get on your brocade and run 'SwitchShow' Brocade command ?
do you see LIP or FLOGI repoted ?


this will show you the WWNs that Stev talked about above.

can you runt:
# ioscan -fnkC fc

to verify your WWNs, then run:

# fcmsutil /dev/td#
where # is the device # reported by ioscan.

[Sreejith Kaliyam]

Hi Tom,

As you can see in ioscan out with my query,I can see the clarion box through 0/10 path. but it shows as CX500WDUNB which I am not able to extend to exising VG as alternate patch

I am attaching switch show of 2800 on which 0/10 card is connected

Anoop

[D Block 2]

Neil,
I see you are going PtoToPoint (PTTOPT) and not Fabric-Logon.

from your SwitchShow:

I don't see the /dev/td0 or FibreChannel Port: N_Port Port World Wide Name = 0x50060b000007978a from SwitchShow command.

but the /dev/td1 - looks fine on the other hand which shows on port 8: sw Online F-Port 50:06:0b:00:00:07:97:24

I'm thinking here, why don't you see the N-Port for t/dev/td0 on the 2800 ?

another question: are you using: PowerPath ? which allows the alternative disk paths for vgextend command ?

can you run: inq

[Sreejith Kaliyam]

HI tom

You wont be able see td0 on this switch because it is connectd another 2800 through which I am able use clarion disk.

No. I am not using any other software like power path

Regards,
Anoop

[D Block 2]

Neil, after I posted the above comment, I did realize it really should be on a 2nd Switch for isolation or SPOF-prevention.

anyway, by all means, verify that this link is up and it works.

how about running a simple: inq emc command and post.

[Mauro Gatti]

A little note...
Haven't you got powerpath?
EMC suggest to use powepath licensed or unlicensed (in this case the software is free but you can't use load balacing and it is named somthing like "EMC Utlity kit for HP-UX").
Powerpath modify kernel and you don't have to be worried about alternate path.
If you have powepath you should see something like:
# powermt display
Symmetrix logical device count=0
CLARiiON logical device count=1
==============================================================================
----- Host Bus Adapters --------- ------ I/O Paths ----- ------ Stats ------
### HW Path Summary Total Dead IO/Sec Q-IOs Errors
==============================================================================
6 0/6/2/0.1.17.0.0 optimal 1 0 - 0 0
7 0/6/2/0.1.26.0.0 optimal 1 0 - 0 0

Bye

[Devesh Pant_1]

Neil,
thanks for the clarification. I now see what you meant. Did you make any progress on this. Please let us know
all the best
Devesh

[Sreejith Kaliyam]

Hi all,

I am working with EMC as well to sort this out.Let you guys know once I find solution.

Regards,
Anoop

[Sreejith Kaliyam]

Hi all,

The problem was resolved by:
â â 1. Removing the host from the Storage Group.
â 2. Exiting Navisphere Manager.
â 3. Restarting Manager.
â 4. Adding the host back into the Storage Group.
â 5. Rerunning ioscan.
â
At this point, both paths could be seen.â

Thanks to all for your valuable hel

[Sreejith Kaliyam]

Thanks guys,

Anoop