Emulex FC HBA initialization "ERROR: 530: Mailbox timeout. (RUN_BIU

Asset ID:	1-72-1558092.1
Update Date:	2018-01-07
Keywords:

Solution Type Problem Resolution Sure

Solution 1558092.1 : Emulex FC HBA initialization "ERROR: 530: Mailbox timeout. (RUN_BIU_DIAG64: mb=10014f767980 tmo=30 Polled.)"

Applies to:

Solaris Operating System - Version 9 GA to 11.1 [Release 9.0 to 11.0]
Emulex FC HBA - Version All Versions and later
SPARC T5-2 - Version All Versions and later
Sun Hardware - Generic - Version All Versions and later
Information in this document applies to any platform.

Symptoms

The /var/adm/messages file will have the following events logged:

May 21 15:02:50 myhost emlxs: [ID 349649 kern.info] [13.0315]emlxs2: NOTICE: 200: Adapter initialization. (Firmware update not needed.)
May 21 15:03:20 myhost emlxs: [ID 349649 kern.info] [13.0D6F]emlxs2: ERROR: 530: Mailbox timeout. (RUN_BIU_DIAG64: mb=10014f767980 tmo=30 Polled.)
May 21 15:03:20 myhost emlxs: [ID 349649 kern.info] [13.03CF]emlxs2: ERROR: 201: Adapter initialization failed. (Unable to run BIU diag. Mailbox cmd=84 status=fe)
May 21 15:03:20 myhost emlxs: [ID 349649 kern.info] [ 5.075E]emlxs2: ERROR: 201: Adapter initialization failed. (status=5)
May 21 15:03:20 myhost emlxs: [ID 349649 kern.info] [ B.1CCA]emlxs2: ERROR: 101: Driver attach failed. (Unable to initialize adapter.)

It is also possible to see the following events:

May 11 10:17:27 myhost emlxs: [ID 349649 kern.info] [ 6.0B89]emlxs2: NOTICE: 200: Adapter initialization. (Firmware module unloaded.)
May 11 10:17:31 myhost emlxs: [ID 349649 kern.info] [13.033A]emlxs2: ERROR: 201: Adapter initialization failed. (Unable to configure port. Mailbox cmd=24 status=14 slimode=3 key=0)
May 11 10:17:37 myhost emlxs: [ID 349649 kern.info] [ 5.062D]emlxs2: NOTICE: 720: Link up. (8Gb, fabric, initiator)

And the FMA data shows the following output:

   Fault class : fault.io.pciex.device-interr
   Certainty   : 100%
   Affects     : dev:////pci@500/pci@2/pci@0/pci@a/SUNW,emlxs@0
   Status      : out of service, but associated components no longer faulty
   FRU
     Location         : "PCIE1"
     Chassis
        Name          : ORCL,SPARC-T4-2
        Part_Number   : 30473707+1+1
        Serial_Number : 1208BDY1D6
        Status        : repaired
Description : A problem was detected for a PCIEX device.

As a result of the above condition, the emlxs instance is not created and the controller/HBA path is not enumerated. The output of "luxadm -e port" will not show the HBA path, and "cfgadm -al" will not show the controller number. prtpicl -v output will show the controller's path, but will display a "-1" for the emlxs driver instance, which means the emlxs driver is not binding with the HBA.

Cause

The emlxs driver determines if the firmware needs to be downloaded to the HBA by comparing the revision that's installed, with the revision that's bundled with the driver. If it is the same, then it skips the download and proceeds with running the diagnostics. If a failure occurs during the diagnostics test, then the driver will log the above "Mailbox timeout. (RUN_BIU_DIAG64: mb=10014f767980 tmo=30 Polled.)" event in the /var/adm/messages file. The BIU_DIAG is another term for a power-on self-test. It has been determined this is an emlxs driver defect. There is a race condition between the DMA buffer allocation for the mailbox and when the emlxs driver checks the buffer for the diagnostic routine during adapter initialization. This defect has been documented in Bug 17331148.

Solution

A fix for this issue has been released with Solaris 11.1 SRU11.1.16.5.0 emlxs driver version: emlxs FCA v20131207-2.90.14 and Solaris 10 SPARC Patch <SunPatch:149173-04> (Solaris 10 x86 Patch <SunPatch:149174-04>) : SunOS 5.10: emlxs driver patch, emlxs driver version: emlxs FCA v20140122-2.90.15.0.

If the system can not be readily upgraded or patched, then use the following steps as a workaround:

a) Confirm the Solaris FMA event "PCIEX-8000-0A" and identify the UUID:

# fmadm faulty

b) Repair this FMA event:

# fmadm acquit "uuid"
# fmadm flush "uuid"

c) Reboot the node again.

To determine the physical path to the HBA, use the instance number of the emlxs driver (in the above trace, it is instance 2) and look it up in the /etc/path_to_inst file.

Use the following command to get the list of all emlxs instances and the correlating paths

# awk '$3 == "\"emlxs\"" {print}' /etc/path_to*

Note. In order to identify if its an Oracle HBA , its physical path, and the slot location on the server , see this other document:

How to identify Oracle[TM] Branded Fibre Channel (FC) HBA and CNA Cards and Their Slot Locations [Doc ID 1282491.1]

To discuss this information further with Oracle experts and industry peers, we encourage you to review, join or start a discussion in the My Oracle Support Community - Disk/Tape Storage Area Networks(SAN)

References

<BUG:17331148> - REPEATED PCIEX-8000-0A REPORTED EVEN AFTER PCIEX HBA HAS BEEN REPLACED TWICE.

Attachments

This solution has no attachment