FYI All others who are running HP MirrorDisc/iX and might
still be planning to upgrade firmware on their HP 3000 add-on
FW-SCSI cards:

HP CE was out to do this for us yesterday.  We have two add-
on FW-SCSI cards in our 959-400. While there was apparently
nothing in the documentation the CE had along that said doing
a VSCLOSE on all effected volume sets was required, the CE
concurred that doing VSCLOSE first was probably "prudent"...
So did same before running FWSCSIPB to initiate download.

Firmware download to first FW card went fine (SIDEBAR:
bigger download than I would have guessed:  130K+, IIRC).
But then we had a slightly "exciting" interlude, because: Shortly
after first FW card download successfully completed, started
getting a bunch of console message indicating a bunch of
volume DISMOUNT / MOUNT requests (that by itself was not
necessarily a surprise).  Then checked DSTAT, and noticed
that all the MEMBERs in one bank for all effected user volumes
(but NOT the MASTERs) had gone into "PENDING-MD" state.

hmmm.....  realizing that I had forgotten to mention to the CE
before he started that we were running MirrorDisc/iX (I get
20 lashes), I asked if he thought we might have any unexpected
"events" if we went ahead and downloaded new firmware to the
second FW card with DSTAT showing the current state;  i.e.:
Might MirrorDisc have any problem "putting the mirrors back
together" after a firmware change ??....   CE said he didn't
think so, but wasn't sure (while clearly having some MPE
knowledge (encouraging), CE admitted he was primarily a
UNIX weenie).

Since this firmware upgrade effected both our production and
backup volume sets, said I'd like to be as sure as possible that
all would be well before running second download.  CE called
expert center;  had fairly extended discussion;  end result:
Recommendation was to proceed with second download;  rather
than try and VSOPEN all user volumes to see if mirrors would
resync with different firmware rev on the two FW cards (might
or might not work;  didn't get definitive answer on that).

Gritting my teeth a little, I told CE to go ahead with second
download......  and......   and......    All was well:  Shortly after
getting both FW cards back on the same firmware rev, DSTAT
showed all mirrored drives back in the LONER state (since we
were still in VSCLOSE state).

ReBOOTed the box;  all mirrored spindles came up normal;  no
REPLACEMIRRVOL was required;  end of story...


I pass this on to any and all fellow MirrorDisc users, to perhaps
help alleviate similar short interludes of elevated heart rates...
and to give Walter & other HP'ers responsible for MirrorDisc an
attaboy:  Apparently it is sophisticated enough to detect not
only that FW firmware has gotten out of sync (or, perhaps, it
detected that the card was busy with a download);  but also
smart enough to notice that when the firmware for both cards
was back in sync, it did NOT have to resync all mirrors....

Ken Sletten