HBA Failover to itself

[klr]

hi,
I have a production system HP DL380 G3 with Dual HBA card. Red Hat Enterprise Linux v.3 Update 4 is installed on that.
I have observed some weird messages on the dmesg & /var/log/message. It seems it tries to failover by itself.

scsi(0): qla2x00_src: MBA_PORT_UPDATE ignored
qla2x00: FAILOVER DEVICE 1 from 50060e80035a7607 -> 50060e80035a7607 - LUN 00, reason=0x1
qla2x00: FROM HBA 0 to HBA 0
qla2x00: FAILOVER DEVICE 1 from 50060e80035a7607 -> 50060e80035a7607 - LUN 00, reason=0x1
qla2x00: FROM HBA 0 to HBA 0

[Don Vanco - Linux Ninja]

RHEL3 Update 4 is fairly ancient. Are you on current / supported drivers?

I assume your servers are in fact multipathed? Everything correct in modules.conf? Sounds like a config issue.

[Dan Am]

I had this one occur with SLES9 SP3 on the same hardware connected to MSA 1500. The OS would report everything ok, the MSA did not see the second path to the server. I had everything checked from OS through HBA and Switches to the MSA Controller and it would not budge. Support had it escalated to Nirvana. Finally our MSA guy pulled out the active controller in mid-air, waited 5 Minutes, pushed it back in and it was handled.
Just a story from the trenches...
HTH
Dan

[Don Vanco - Linux Ninja]

Well, you're not very forthcoming with required information, so here's some common issues:

Your switches are in loop mode when they should be in fabric mode

Your HBAs are not 2200s, they're 2300s and you need to make some config changes

You're using the kernel-based driver that was known to be faulty when you should be using H-P's (or QLogic's)

...but until you cut loose with a little more information on what you've got (hardware, driver, topology details) I don't know that anyone can offer you much help.