[Asterisk-Users] * box hangs after a couple of days...

[Michael Loftis]

--On Tuesday, October 12, 2004 20:32 +0200 Patrick 
<asterisk at puzzled.xs4all.nl> wrote:

> On Tue, 2004-10-12 at 20:23, Michael Loftis wrote:
>> Yup it's official, the box is now a brick thanks to the zaptel drivers.
>> What madness they're doing I have no idea, but the filesystems are all
>> pretty much corrupted, with the drives being totally fine, both sides of
>> the mirrored set which pretty much rules out the drives themselves.
>
> I seriously doubt that the zaptel drivers caused this. Seems like an
> ext3 bug or some RAID hardware-kernel interaction related issue. I agree
> with Steven Critchfield on this one. Where is the proof accompanying
> your bold statement?

Lets see, the fact that this exact box, minus the Zaptel gear, has ran 
flawless for a year.  I've used it as a log analyzer box in the recent past 
which means it did nothing but hammer on the drives for hours on end 
without any corruption.  All the other code paths in this box are well 
tested.  The fact that this same kernel runs on 15+ other boxes here, I've 
got two that are identical except the Zap stuff.

Now it could be an interaction, but whatever it is the root or atleast 
cause is pretty surely in the Zap drivers or in the hardware (hard to say 
which since the hardware isn't active without the drivers).

I've been trying to work this, went through I don't know how many 
motherboards before finally finding one that the finicky.

I'm going to work with what I've got and we do have a Supermicro P4SBR 
based system (just arrived today) but the experience here says that there 
is something of a gremlin lurking under the hood.  Everyone here it seems 
wants to pass it off on 'get better hardware' but in my experience it's 
more likely that there's just something not quite right with the drivers or 
the hardware being installed when you get a lot of that going on.  Or pass 
it off on anything but the Zap or Digium gear.

I don't want to use a new box for this, mostly because I have no idea if 
the box is stable or not on it's own.  So far I've found two motherboards 
that the Zap hardware works in, one I can't use (it's my desktop) and the 
other I'm using for testing, every system it's been tried in is known true 
and has seen atleast a year of good service and been well burned in.

I'm just really concerned that what's been pawned off as 'cheap hardware 
problems' is really just a bad driver problem.

Anyway noone cares to listen to my clanging to I'll shut up.