All,
Those of you who are thinking of applying the FLASH PROM patch 103346-04
to your Ultra Ex000's please be aware that it has a serious bug (in my
opinion) in it.
I applied this patch first time 2 months ago on an E4000 and this knocked
out the I/O board and serial cards on the system. It wasn't aparent, at the
time, to the engineer and ended up changing the components. We could not
say for sure that it was caused by the patch as at the same time we had applied
other patches and added more memory. ( I know, things should have been done
one at a time but a downtime of 15mins a month is not enough to have this
luxury). It was also friday the 13th - not that I am superstitious (sp?).
It took 6 hours to bring the system back up.
Anyway, last friday, acquiring a longer downtime on the system (1.5-2hrs),
applying patches one at a time, the reboot of the machine failed to see
disks other then the system disk. My nightmare had come back and this time
it was Valentines day and not friday the 13th.
probe-scsi and probe-scsi-all would hang the system. In the end, the SUN
engineer changed the I/O board. Then the ePOST reported Serial Card checksum
errors. Comparing the firmware revisions of the cards with an okay system
and the FE manual indicated a possible firmware rev problem. He then called
the SunSolve in states to find out if this was the case as their FE manuals are
not uptodate ;-). They said that this is a known bug BugID: 4024654. The only
was to avoid the checksum error is to turn off all the diags and reboot the
system with minimum diag. This rebooted the systems and the cards were
initialised. However, I am still experiencing manual X.25 start/stop problems.
I works from rc files though. This time the downtime was almost 9 hrs. :-(
Looking through the Special Instruction once again in the README file of the
patch does not mention anything of such nature.
Since this is a known bug they why SUN has not released a fix for it? I reported
the problem first 2 months ago. I spoke with the H/W manager in SUN and he
said that he'll get the s/w people to update the README of the patch and
apologised.
Anyway, the moral of the story is beaware of this patch and avoid extended
downtime on the system.
-- __________________________________________________________________________ Umesh Vaghela e-mail: umesh.vaghela@one2one.co.uk Senior Unix Systems Analyst tel: +44 181 214 3357 Network Services fax: +44 181 214 2278Mercury Personal Communications Ltd. (one2one) Elstree Tower, Elstree Way, Borehamwood, Herts., WD6 1DT, UK
Disclaimer: I only speak for myself hence the opinions are mine only...and the rest of the text.... __________________________________________________________________________
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:46 CDT