Start a Conversation

Unsolved

This post is more than 5 years old

1486

March 14th, 2012 12:00

Virtual AIX 6.1 host will not log into VMAXe

I have a virtual AIX 6.1 host that will not log into VMAXe on FA ports 2E:0, 4E:0.  I have checked zoning (Cisco 9513) and ensured that each of the four WWN's log into the switch. but when I run symaccess -sid 0554 list logins -dirport 2E:0 it shows the host a "Logged in" no  and "On Fabric" Yes.  I have done the following troubleshooting:

  1. EMC support logged in to the array and ran full health checks with no issues on the array or these specific ports.
  2. The system admin has run cfgmgr several times.
  3. The system admin has confirmed that the HBA's are set for "switched fabric"
  4. Ensured that the host WWN's log into the fabric where expected and with the correct WWN's.
  5. I have removed zoning, cleared the logins entries and added the zones back,
  6. Run an EMC grab on the host and it comes back with no issues.

ADDITIONAL INFO:

Host Name: LITIADBD001PF

Product Model                        : VMAXe
Symmetrix ID                         : 000195900554

Microcode Version (Number)           : 5875 (16F30000)
Microcode Registered Build           : 35
Microcode Date                       : 01.27.2012

Microcode Patch Date                 : 01.27.2012
Microcode Patch Level                : 231

There are two physical AIX 6.1 servers connected to these FA ports and functioning without issue.

I am also having this issue with two other AIX virtual nodes on FA ports 1F:0, 3F:0.

C:\Users\e1045156>symaccess -sid 0554 list logins -dirport 2E:0

Symmetrix ID            : 000195900554

Director Identification : FA-2E
Director Port           : 0

                            User-generated                      Logged On
Identifier       Type  Node Name        Port Name        FCID   In     Fabric
---------------- ----- --------------------------------- ------ ------ ------
10000000c9491f60 Fibre LILEVDBP12PF     HBA1             700a44 Yes    Yes
10000000c95547a4 Fibre LILEVDBP11PF     HBA1             700a45 Yes    Yes
c0507604590e001d Fibre LITIADBB001PF    HBA3             4f012e No     Yes


C:\Users\e1045156>symaccess -sid 0554 list logins -dirport 4E:0

Symmetrix ID            : 000195900554

Director Identification : FA-4E
Director Port           : 0

                            User-generated                      Logged On
Identifier       Type  Node Name        Port Name        FCID   In     Fabric
---------------- ----- --------------------------------- ------ ------ ------
10000000c94a0ad1 Fibre LILEVDBP12PF     HBA2             7a0b45 Yes    Yes
10000000c94b4506 Fibre LILEVDBP11PF     HBA2             7a0b46 Yes    Yes
c0507604590e001f Fibre LITIADBB001PF    HBA4             080131 No     Yes

1 Rookie

 • 

20.4K Posts

March 14th, 2012 15:00

don't have experience with VIO, sure sounds like something host related. Can't be physical media as you have two HBAs exhibiting the same behavior.

465 Posts

March 14th, 2012 15:00

Has the host established paths to the array?

  • Open SMIT. 
  • Select devices. 
  • Select MPIO management. 
  • Select MPIO path management. 
  • Select List paths... Check the WWN's. Are they there and are they to the VMAXe?

40 Posts

March 15th, 2012 05:00

Here is what that returns

Enabled hdisk0 0 vscsi0 810000000000

448 Posts

March 15th, 2012 06:00

I have had a decent number of issues getting virtuals on AIX to login to the SAN properly (we also boot from san).  Make sure that any HBA port not configured to an LPAR has the loopback connector in it (seems to wonk out scanning).  We do at times have to network boot the LPAR multiple times to get it to login correctly; the SAN switch will show green but as a U as it doesnt seem to complete a fabric login only the port login, this is at the physical VIO level.

Whats really odd is when the parent VIO will be logged in on all its paths but the underlying virtual is where the problem occurs.  We know physically its there as the VIO is connected up on the true WWPN of the HBA and the virtual has a generated WWPN; so far they all start with C0:00 on AIX. Again the virtual LPAR has to be network booted several times to get it to finally login all the way.

Once it is in the SAN we havent had too many issues getting the system to pick up the disks.  We have had  instances where a virtual will lose a path for no apparent reason.  All the other virtual LPAR's and the parent VIO's will be fine, again a reboot will fix it; scanning rarely does.

From the storage guy perspective I am not terribly thrilled with AIX and virtual systems.  It shouldnt take what it does to get these things working correctly in a SAN environment.  Dont even get me started on the caveats to support live partition mobility with a SAN environment.

40 Posts

March 15th, 2012 07:00

G

No Events found!

Top