From: O'Brien, Pat (pobrien@mitidata.com)
Date: Wed Apr 17 2002 - 11:40:09 EDT
there were no answers from the list, however in a support call on this
matter, they stated in version 5.X there never is a emx1.
After upgrading a large alphaserver gs140 from 40f to 40g and then onto 5.1,
A storage issue occurred. This system has 4 "private sans", each with their
own hba. Prior to the upgrade the emx buses were referred as emx0, emx1,
emx2, emx3 while after the upgrade they were emx0, emx2, emx3, emx4. Notice
emx1 is no longer. The bad thing is that emx1 was munged with parts of emx2
to become the "new" emx2, the old emx2 was munged with the old emx3 to
become the "new" emx3, and the old emx3 was munged and strippedto become
emx4.
sample munging of emx2 and emx3:
emx2 SCSI target id assignments:
SCSI tgt id 0 : portname 5000-1FE1-000F-3841
nodename 5000-1FE1-000F-3840
SCSI tgt id 1 : portname 5000-1FE1-000F-3842
nodename 5000-1FE1-000F-3840
SCSI tgt id 7 : portname 1000-0000-C926-97A5 (emx3)
nodename 2000-0000-C926-97A5
Also the old emx1,2,3 has their scsi id's changed from 7 to 255 as seen
with emxmgr -t :
emx2 state information:
Link : connection is UP
Point to Point
Fabric attached
FC DID 0x011000
Link is SCSI bus 3 (e.g. scsi3)
SCSI target id 255
portname is 1000-0000-C922-3C01
nodename is 2000-0000-C922-3C01
Would ythis be normal behavior?
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:48:38 EDT