Not able to see presented SAN disks

[Jonathan Caplette_1]

Hi guys,

I'm trying to see which new disks our SAN team have assigned to my HP-UX box.

It's an rp8420 with Tachyon XL2 running on HP-UX 11.11, the SAN is an Hitachi it's all I know about...

I've asked our SAN team to present me a 100GB disk. They told me that it was done... Normally I do :

ioscan -fnC disk -> I don't see the new disk normally I should see a device that have not been created.

insf -C disk -> this should create a new device, right? but right now it is not doing anything.

ioscan -fn | grep rdsk | inqraid -CLI -fxd -> This should show me all the presented disk to the box with their LDEV number. I can't find the new one that the SAN team gave to me...

Is there a way I can find if my box can see or have the new SAN disk presented???

Regards
Jonathan

[TwoProc · Accepted solution]

Any chance that they put a program called "xpinfo" on your box? I know that an XP storage system is an HP, but those are just rebranded Hitachis. If you've got that program (usually in /opt/xppa ), you might be able to get more information on the problem.

[David Bellamy]

Jonathan, just try a straight ioscan first, then do insf -e -C disk. I have noticed that if you don't do a straight ioscan that sometime the devices don't show up.

[Jonathan Caplette_1]

TwoProc: No chance I don't have that utility

David: I did a ioscan followed by the insf -e -C disk and it didn't work either...

Any other ideas?

[Tim Nelson]

Jonathan,

Looks like you hit all you can. The SAN guys forgot a step or misconfigured something.

Have them recheck their work.

[TwoProc]

I'm wondering if they created the LUN, but didn't correctly configure the secure portion of the SAN management to let you see it. Ask the admins to review the secure assignments to your server.

[Sandman!]

Did you run ioscan without any arguments? Only that will bind the drivers to the LUNs. An "ioscan -fnC disk" won't i.e.

# ioscan
# insf -e -C disk

[Jonathan Caplette_1]

I already asked the SAN team... they return to me and said they already did everything they should and they did it right...

Is this anything to do with the driver of the FC cards?

Sandman: Yes I did your suggestion, as David suggested earlier... Didn't work...

[Tim Nelson]

If they say they are cool and you have no tools to prove otherwise from your end then we are at an impass.

I cannot speak for Hitatchi tools but in the EMC world we have solutions enabler that allowed you to look at the presentation and masking database from the server side.

We are also assuming that this server already sees other storage presented to this server and we are not at some weird lun limit and that all switch zoning is correct as well.

How about them presenting to a different server to see if there is the same issue ?

The only last hope could be a reboot. I have seen some cases where it takes a logoff logon of the FC HBA to see the change.

[Sandman!]

Ask your SAN team for the CU:LDEV number of the LUN they have presented to the HP-UX host. If it does not match the output of inqraid then it's most likely a LUN zoning/presentation problem.

[TwoProc]

I think Tim's suggestion to present it to another server is about all you've got left for troubleshooting. Unless they will let you look over their shoulder and show you first a working LUN and it's CU:LDEV and security allowances, and then show you how they did the same for the new allocation, paying close attention to the WWN that was associated with the allocation. I'd probably be pushing for both right now, the latter first. At times like these, buying someone lunch and explaining to them about the impact and the steps you've taken thus far will usually get them on board in problem solving mode to assist you. I know you'd probably feel a lot better about the setups if you could see for yourself that the new allocation setup is the essentially same as a working storage allocation that you've already got.

The other thing to do is to ask them to open a call with Hitachi, and you open a support call with HP. Something is missing from one side or the other, and it has to turn up.

[Andy Torres]

You didn't mention whether there were already disks presented to the host from the same array on the SAN. If you already have LUNs presented from the same array and these new ADDITIONAL LUNs aren't showing up, I'd investigate the SAN team's LUN assignment more closely (I know it can be a delicate dance between inquiry and sensibilities, but let's assume everyone's on the same team and working together professionally). Maybe they created an additional host group that doesn't have LUN security or something like that. It's very easy to misconfigure the SAN. They shouldn't mind having another pair of eyes to help solve the problem.

The host side is pretty straight forward - it sees what is presented to it or it doesn't. A plain ioscan and insf -e -C disk should do the trick there.

Now, if this is the first time presenting LUNs to this host, you might throw HBA firmware mismatch into the troubleshooting mix. Double-check the HBA is compatible with the array. Is it SAN attached or direct-connect? So many variables involved.

[Jonathan Caplette_1]

Good morning guys!!

Yes the server already have some LUNs presented to it and they work fine...

I logged on the server this morning and now I can see the newly presented LUN, that yesterday I'm positively sure that I didn't see... I really suspect that our SAN team did something yesterday or this morning...

And now I did the insf -e -C disk and see the new device and can access it!!

Thanks for your input!

[Jonathan Caplette_1]

Closed!