We have loads of old IBM servers on maintenance. These are Pentium PRO 200
and below based boat anchors.
I just had to show two different IBM engineers how to replace a RAID disk in
a hot swap caddy. If you don't connect the SCSI select cable, or in the
second case, connect it backwards, it no workie.
Reminds me of the time a system manager at a place I was working ordered
a failed servoid off the site. He wanted to replace every PCB in some
expensive piece of equipment because they all failed diagnositcs. I'd
just turned up, slapped a meter on the 5V line and found it was sitting
at 4.2V....
Both engineers, after very little troubleshooting, wanted to order more
drives.
What is this world comming to?
On which planet have you been living for the last 20 years????
I've never met a field servoid who had any sort of clue whatsoever about
the operation and troubelshooting of computers -- they just replace
seemingly random parts until the fault goes away. Certainly I've never
seen oue use a multimeter, let alone a 'scope...
Needless to say I am not happy about this, this is no way to fix a
computer. I've given the arguments many times before, so suffice it to
say you'll never trace an intermittant fault that way. And you may not
even correctly diagnose a steady fault (if module A has gone out of
tolerance, then replacing module B (which is still good) with one that's
got a slightly different margin might get the machine working again --
that is untile A drifts some more. Seen it happen!).
I was once taught that the really good troubleshooter is the guy who
makes some measurements, thinks _a lot_ about the problem, and then
replacees exactly one part, and the machine works again for 10 more
years. I am not that good!
And don't give me the crap that proper troubleshooting is too expensive.
Firstly, the modern way flatly doesn't work!. I've lost count of the
number of times a machine has had to be 'repaired' several times for the
same fault because the idiot servoid hadn't a clue about what was really
wrong. All that downtime costs money!. And I am also not convinced that
replacing rather than repairing a PCB is the way to go. I've also lost
count of the number of times an expensive motherboard -- say in a
workstation -- has been replaced for a serial port problem. Anyone with
half a brain could find the dead buffer chip in 5 minutes, even without
schematics. And it would take not much longer to solder in a replacement.
Oh, what the heck... Here's a filk that might cheer you up...
This is dedicated to all those who called out DEC field service for a simple
problem, and wished you hadn't..........
It was on a Monday morning
The DEC man came to call,
My system wouldn't boot
There was no prompt at all
He pulled out all my SPC's
To try a new backplane
And I had to get the hardware guys
to put them back again
Oh it all makes work for field service men to do!
It was on a Tuesday morning
The hardware man came round
He soldered and he fiddled
And he said 'Look what I've found'
'Your ECOs are years behind'
'But I'll put it all to rights'
Then he shorted out the power supply
and out went all the lights
Oh it all makes work for field service men to do!
It was on a Wednesday morning
The power supply came
'It's newer and it's better'
'But it works just the same'
He could not fit the unit
without stripping half the rack
then he dropped my boot HDA
so He called Peripherals back
Oh it all makes work for field service men to do!
It was on a Thursday morning
The HDA came along
with a blocklist and a cable
and a list of what goes wrong
He put it into my drive
It took no time at all
But I had to get the software guys
to come and re-install
Oh it all makes work for field service men to do
It was on a Friday morning
That Software made a start
With BACKUP and SYSGEN
He configured every part
Every track and every sector
But I found when he was gone
He had overwritten the boot track
and I couldn't turn it on
On saturday and Sunday They do no work at all
So It was on a Monday morning that the DEC man came to call
-tony