[about updatedb]
> This was my first thought too, but after checking timestamps and logs,
> this was not it. In fact, I believe these were in very much idle
> situations -- which makes it even stranger. My hunch is something
> flaky in power management, but just wildly guessing.
I switche dover to QQ after I had an uptime of somewhere in the 30 days
range, with 4 seti's running. I also installed a new kernel (in fact,
the kernel was the reason I also flashed bios etc)
Then, I ran for several days and started using the 2.2.15pre series as I
saw CLOSE_WAIT hangs, Alan Cox suggested that the pre-series maybe could
affect this behaviour so I had some low uptimes. No sweat until a few
days ago. all over sudden with the QQ-2 bios, the system crashed. I
quickly found out that running X caused the locks. Then I narrowed it
down to the fact that I run 4xset (with ksetiwatch).
I stopped the seti clients and X stays alive.
If I start a single seti, it will crash in about 15 seconds to 15
minutes.
The only behaviour so far is that the fanspeeds increase when I start
seti. It does directly. I think that here it could be a power supply
thing, although I also must admit that make -j30 (load will ben 30+)
doesn't affect the stability at all, even if I run it several times.
when I start make oldconfig, make -j bzImage etc, the fanspeed also goes
a little up.
it's weird -- my system ran fine for months and all over sudden, it
starts to behave strange. Note that the line input to the power supply
is conditioned.
seti seems to be memory bandwidth aggressive afaik.
Roeland
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
