VMware Cloud Community
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Only 1 hba sees luns after upgrading to u3

Hi all,

I upgraded one of my hosts to ESX 3.5 U3 today using Update Manager. After the update, only one of the HAB's sees the luns. Before the upgrade both HBA's saw all the luns assigned. Using the command esxcfgboot -q vmkmod and comparing the output with an U2 host that works fine, I don't see any difference. I'm not sure where to go from here and what I can check. Any help and tips to what I can do/check to figure why this has happened and hopefully fix it will be greatly appreciated. The server in question is HP BL480c with Qlogic QMH2462 FC card.

I'm a Windows guy so please be fairly detailed about any Linux/Esx console commands that might help me.

Thanks in advance.

GB

0 Kudos
1 Solution

Accepted Solutions
Texiwill
Leadership
Leadership
Jump to solution

Hello,

Verify the presentation and zoning of your LUNs to the HBAs.


Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education.

Blue Gears and other Blogs: http://www.astroarch.com/wiki/index.php/Blog_Roll

--
Edward L. Haletky
vExpert XIV: 2009-2023,
VMTN Community Moderator
vSphere Upgrade Saga: https://www.astroarch.com/blogs
GitHub Repo: https://github.com/Texiwill

View solution in original post

0 Kudos
20 Replies
Goodspd
Enthusiast
Enthusiast
Jump to solution

Hi,

Did you already force a rescan in the HBA with no luns?

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!
0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

Yes I have, several times. I apologize not mentioning that.

0 Kudos
Mark_Bradley
Enthusiast
Enthusiast
Jump to solution

Hi,

Did you apply all the latest patches? As a patch for QLogic HBA's was recently released... maybe this had an effect on the card that is a problem.

VMware KB1007505

This patch includes firmware for the HBA. When you boot up the server can you get into the BIOS for the HBA?

__________________________________________________________________________________________

Check out my blog at

__________________________________________________________________________________________ Check out my blog at http://www.ridethevirt.blogspot.com
0 Kudos
Texiwill
Leadership
Leadership
Jump to solution

Hello,

Verify the presentation and zoning of your LUNs to the HBAs.


Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education.

Blue Gears and other Blogs: http://www.astroarch.com/wiki/index.php/Blog_Roll

--
Edward L. Haletky
vExpert XIV: 2009-2023,
VMTN Community Moderator
vSphere Upgrade Saga: https://www.astroarch.com/blogs
GitHub Repo: https://github.com/Texiwill
0 Kudos
mike_laspina
Champion
Champion
Jump to solution

Hi,

Check if the HBA paths are ok using the following command on a console session.

esxcfg-mpath -l

If you are using ssh to connect then be sure to swtich to the root user with

su -

http://blog.laspina.ca/ vExpert 2009
0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

Last night I updated with all the latest updates, including the QLogic patch. Still same result. I have checked the bios settings of the 2 adapters, and they are the same but only the first adapter sees the luns.

0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

I will have to ask the SAN admins to check the presentation and zoning on Monday, but can updating ESX really mess with zoning? I thought all that kind of configuration was done on the SAN side and the client only received information and cannot change it.

0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

Here is the result of the esxcfg-mpath -l command. Correct me if I'm wrong, but as far as I can tell only the onboard disk controller and one of the HBA's is listed in the output.

Disk vmhba1:0:5 /dev/sdd (552960MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:5 On active preferred

Disk vmhba1:2:0 /dev/sdl (1171MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb710a vmhba1:2:0 On active preferred

Disk vmhba1:0:11 /dev/sdj (399360MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:11 On active preferred

Disk vmhba1:0:0 /dev/sda (460800MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:0 On active preferred

Disk vmhba1:0:8 /dev/sdg (552960MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:8 On active preferred

Disk vmhba1:0:6 /dev/sde (552960MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:6 On active preferred

Disk vmhba1:0:1 /dev/sdb (92160MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:1 On active preferred

Disk vmhba1:0:7 /dev/sdf (276480MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:7 On active preferred

Disk vmhba0:0:0 /dev/cciss/c0d0 (69973MB) has 1 paths and policy of Fixed

Local 8:0.0 vmhba0:0:0 On active preferred

Disk vmhba1:0:9 /dev/sdh (552960MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:9 On active preferred

Disk vmhba1:0:2 /dev/sdc (478940MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:2 On active preferred

Disk vmhba1:0:10 /dev/sdi (7318MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb6004 vmhba1:0:10 On active preferred

Disk vmhba1:1:0 /dev/sdk (1171MB) has 1 paths and policy of Fixed

FC 15:0.0 50060b00006bbb30<->50060e8004eb600a vmhba1:1:0 On active preferred

0 Kudos
Texiwill
Leadership
Leadership
Jump to solution

Hello,

You are correct. It looks like a presentation issue to me. Did you run this command at the previous version of ESX? Perhaps the SAN team did something at the same time?

I would verify using 'lspci' that you see both controllers? Or see both in the VMware Infrastructure Client.


Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education.

Blue Gears and SearchVMware Pro Blogs: http://www.astroarch.com/wiki/index.php/Blog_Roll

Top Virtualization Security Links: http://www.astroarch.com/wiki/index.php/Top_Virtualization_Security_Links

--
Edward L. Haletky
vExpert XIV: 2009-2023,
VMTN Community Moderator
vSphere Upgrade Saga: https://www.astroarch.com/blogs
GitHub Repo: https://github.com/Texiwill
0 Kudos
williambishop
Expert
Expert
Jump to solution

I had a similar thing happen twice. Once, I rebooted the server and everything was peachy (for some reason the device was not seeing what was presented). The second time, it turned out to be a device alias issue in the zoning (which was a puzzler, since I am the san administrator as well--). I deleted the device alias, re-added the alias and zone into the fabric and initialized it, and voila! I could see it again.

--"Non Temetis Messor."
0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

I did not run the command before the upgrade, but I do see both hba's in the Infrastructure client and with the lspci command. I doubt the san team did anything since one of them was on vacation and the other was sick when I upgraded. Hopefully they will find something. Thanks for trying to help.

0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Thank for the info, it is very much appreciated. I have forwarded your post to the san admins and hopefully they will figure it out.

0 Kudos
mike_laspina
Champion
Champion
Jump to solution

Hi,

The FC path zoning may be an issue so that still needs to be ruled out by the storage administrator otherwise your mappings are good and this could be a PCI enumeration issue or the FC driver has an IRQ issue

Are you able to do one or both of the following.

1 Revert back to update 2 to see if the issue goes away. (Driver)

2 Clear cmos or move the FC Adapter slot. (Enumeration)

Regards,

Mike

http://blog.laspina.ca/ vExpert 2009
0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

Yes downgrading to update 2 is of course an option. I will try that if the san admins don't come up with anything. This is a blade server and the fc adapter kan only be in one slot unfortunately.

Thanks for the suggestions.

GB

0 Kudos
depping
Leadership
Leadership
Jump to solution

try a "reset to factory defaults" on the HBA, it helped me fix an issue like this once.

Duncan

Blogging:

If you find this information useful, please award points for "correct" or "helpful".

0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Duncan,

Thanks for the suggestion. Unfortunately it did not help. I appreciate you trying to help.

GB

0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Hi,

An short update. The san admins cannot find anything wrong and downgrading to U2 did not help. Just for fun I'll try ESXI.

GB

0 Kudos
Mark_Bradley
Enthusiast
Enthusiast
Jump to solution

Hi,

Have you tried shutting down the blade, removing it from the chassis and putting it back in? Reseating blades can help with issues like this sometimes.

Also it could be that by some strange coincidence there was a hardware fault at the same time as the upgrade.

__________________________________________________________________________________________

Check out my blog at http://www.ridethevirt.blogspot.com

__________________________________________________________________________________________ Check out my blog at http://www.ridethevirt.blogspot.com
0 Kudos
Aftenposten
Enthusiast
Enthusiast
Jump to solution

Yes, I have tried that also with no luck. I have also had time to install a new server with u2, get everything working, upgrade to u3 and bang, no luns on one hba. Reinstalling u2 doesn't help either. But I have discovered something. It is only the path connected to one spesific fc switch (Brocade) that doesn't work. To me it seems like u3 (or maybe another update) does something to the hba that makes it fail to connect to the Brocade switch. I haven't had time yet to look through all the post u2 updates to see if one of them might explain this behaviour.

GB

0 Kudos