I have a strange problem with a RaQ3i (CPU upgrade to 500 MHz, RaQ550 OS installed RAID1 with two new drives):

After having set up the server, it worked fine for about two days. Then (no live websites, it was just online for first configuration) it refused to work from one minute to the other.

Interestingly, since we added two new HDs to it, the server took one whole day before it synchronized the drives (shouldn't this happen after first startup immediately ?). The next day, it failed.

The server hang (not reachable by its IP even), showed up the clock symbol at the LCD display. It was just possible to turn power off "the hard way".

When turning power on again, it showed "Primary IP, no IP address set" (or similar, this happened last week, so I cannot remember exactly). The disk LED lights up and something seems to happen (synchronizing RAID ?). This takes around one hour. Then nothing happens. Buttons are dead, too.

Now when turning power off and on again, it is exactly the same. Unfortunately I had no chance to look at it by a console so far.

But maybe someone has an idea.

Can be anything. I had problems with my OS 550
rebuilding some database at every boot. Could
take from 30 minutes to 15 hours..

It stalled at 'Checking clock'.

The only way I found out what it did was through
a nullmodem cable and I suggest you do the same.

Yes, I guess our technician does this at the time we speak here :-)

Can we exclude hardware issues (I ask, because we got this as a used server) ?

You can't exclude the hardware unless you
see what it is over the console.

These machines are a couple of years old
and most of them has been turned on pretty
much since they were bought. Fans fail, PSU
can deliver an unsteady current, etc.

The last time something like this happened was because one of the PSU connection point was oxidised and it breaks the connection very few hours or minutes....

The other is overheating when there is high loads and maybe a faulty ram. But comparing my RaQ550 and my RaQ3/4(s) its the longest runnng one without any failures for years.

@galacnet: In this case I would believe to an oxidized soldering point or connection point. I'm unsure if this server was in a rack or a stand-alone office one, so humidity could be really a problem in the past. Thank you for the hint, I'll investigate this :-)

Our investigations so far showed up that one of the harddisks had a hardware problem and refused to work. We will now try the system again with two fresh disks seeing how anything performs for a few days.

However I can swear that Active Monitor hasn't showed up any disk integrity problems or hardware issues. Shouldn't it support SMART for harddisks ?

The used harddisk was a Maxtor Fireball 3 40 GB, which is at my knowledge a widely used one (we have used the Maxtor testing tool, but beside a numerical error message, it did not show up further details).