From: MURPHY, DEAN (SBCSI) (dm6757@sbc.com)
Date: Tue Jul 01 2003 - 14:09:06 EDT
Just wondering if anyone out there has ever seen any similar errors or
problems
using an JNI FCE1473 HBA direct loop attached to an EMC Storage Array.
Host is an Enterprise 4500 running Solaris 8
Direct loop attached to 2 separate EMC syms for storage.
Dual pathed using VXVM for volume management.
Basically I'm running into an issue with only one of the 2 attached storage
syms.
On a reboot or power on, I am periodically receiving the following errors on
link
initialization:
OUTPUT from host console:
07:10:01: jnic146x4: Hba: fcr Model: FCE-1473-N Type: scsi
07:10:01: jnic146x4: FCode: Version 3.8.9 [668] XCode: 2100b [106a]
07:10:01: jnic146x4: Slot: 1 IRQ: 3 Mode: 64 Channel: 0
07:10:01: jnic146x4: WWNN: 1000000173004E9C WWPN: 2000000173004E9C
07:10:02: jnic146x4: Forcing link speed to 1 Gig
07:10:02: jnic146x4: Configured as Private Loop port
07:10:03: jnic146x4: JNIC v5.1.1 (02051300)
07:10:03: jnic146x4: Copyright(c) 1995-2001 JNI Corp, All Rights Reserved.
07:10:13: jnic146x4: ulm_event: RESET_REQUIRED
07:10:14: jnic146x4: Forcing link speed to 1 Gig
07:10:14: jnic146x4: Configured as Private Loop port
07:10:24: jnic146x4: ulm_event: RESET_REQUIRED
07:10:24: jnic146x4: Forcing link speed to 1 Gig
07:10:24: jnic146x4: Configured as Private Loop port
Finally after several attempts, the host completes initialization and
continues booting,
minus the erroring path. Since we are using DMP under Veritas, the host does
come up with all of the volumes intact. The link continues to reset in the
background and eventually does regain link, and the devices come up. I can
then manually run a vxdctl enable to regain my downed path.
On one boot, both paths to the sym failed with the above errors, which
caused
the host to dump into single user. We had to reset-all the host several
times
in order to get at least one link up to complete the boot.
At EMC's suggestion the driver was updated to 5.2.1 and firmware updated to
latest available.
Even with this we are still experiencing the problem.
I have searched all over the net, and have yet to find a reference to what a
"ulm_event" is and possibly its cause.
If anyone out there has ever seen anything even similar or possibly has some
suggestions,
that would be great. I can supply more information if necessary.
Thanks for everyone's time!
Dean
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:41 EDT