Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

2932

September 6th, 2012 08:00

Host only sees 2 out of 4 paths

I have a host which is only seeing 2 paths for all luns, while all other hosts are seeing 4 paths.

A sample lun is shown below:

Pseudo name=emcpowerq

CLARiiON ID=APM00071602762 [hp1]

Logical device ID=6006016018C21C00F44A96B635E2E111 [TST_C1S0U01]

state=alive; policy=CLAROpt; priority=0; queued-IOs=0;

Owner: default=SP A, current=SP A       Array failover mode: 1

==============================================================================

--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---

###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors

==============================================================================

   0 lpfc                     sddx      SP A5     active  alive       0      0

   1 lpfc                     sdeo      SP A4     active  alive       0      0

Ive noted that some luns only show "A" ports and others only show "B" ports, but no luns show both "A" and "B" ports.

On other hosts, I see that each and every luns has both "A" and "B" ports for a total of 4.

Ive chekced the zoning to make sure that both host HBA's can see both SPs.

Ive also checked the "host connectivity" to ensure that both host HBA's are defined and assigned to the host in the array itself.

Unsure why this is happening, or how to diagnose further.

Suggestions welcome.

2 Intern

 • 

5.7K Posts

September 7th, 2012 02:00

Can you open the properties of the storage group and go to the hosts tab in engineering mode? Click on the advanced button and check to see if all paths are active.

Option 2: reboot.

89 Posts

September 6th, 2012 11:00

Hello,

Check if the HBAs are seeing the Clariion ports down every path.  Use the software utility that comes from your HBA manufacturer.  PowerPath will only display what the HBAs are able to detect.

If the HBAs detects all four ports on the Clariion then check Connectivity status on the Clariion.  Verify the HBA WWNs (initiators) are listed, one for every path, and all are logged in, registered and connected to the storage group. 

2 Intern

 • 

222 Posts

September 6th, 2012 12:00

2 Intern

 • 

222 Posts

September 6th, 2012 12:00

Not sure what brand the HBA's physically are. The WWN of the HBAs are

0x10000000c9711572

0x10000000c9711573

Im assuming they are either Emulex or Qlogic, but not sure how to tell without taking the box offline.

I didn't get any software, but im assuming that if I can figure out the card, then I could get it from the MFR website.

Yes, the Clariion sees both controllers in the connectivity status. The storage group also shows both HBA's under the "Hosts" tab. What is interesting is that for luns owned by SPA, I only see the A4/A5 paths and for luns owned by SPB I only see the B4/B5 paths.

Not sure why this is happening, other hosts see all 4 paths from A4/A5/B4/B5 for each and every lun.

Is there a way to tell the HBA MFR from some data in the Linux /proc FS or somthing under /sys/class/fc_host tree?

2 Intern

 • 

222 Posts

September 6th, 2012 12:00

1 Attachment

2 Intern

 • 

222 Posts

September 7th, 2012 07:00

The host ports are all active. See attached pix. Are you talking about rebooting host or CX array?

2 Intern

 • 

222 Posts

September 10th, 2012 07:00

Did a host reboot over the weekend, and that seems to have resolved the issue. I have 4 paths to each lun now.

Thanks

2 Intern

 • 

5.7K Posts

September 12th, 2012 06:00

I knew it.

2 Intern

 • 

222 Posts

September 12th, 2012 07:00

Im glad it resolved the immediate issue.

After the reboot I got 4 paths to all the luns.

However, after a short time, it went back to 2 paths, so there is still something going on with the host which is causing it to drop half the paths…

2 Intern

 • 

5.7K Posts

September 12th, 2012 07:00

Like you said: on the host.

I suggest you run EMCGRAB and have it checked by HEAT (or create a SR and have EMC help you out). Perhaps it’s a firmware mismatch, a wrong VMware HBA setting, maybe you’re in need of some sort of OS patch?

No Events found!

Top