Jim> Hitting SHIFT+SCROLLOCK when this happens reveals that in every
Jim> case there is exactly *one* buffer that's locked... so I think
Jim> there's either a deadlock or some code path that's not
Jim> releasing a buffer when it should... I'm not sure this is an
Jim> Alpha-specific problem either...
Any chance you're running the syslogd that comes with sysklogd1.2?
There was a time where I saw the system "hang" because of it. It
turns out there was a very simple bug corrupting memory. Since I
fixed that, I haven't any hangs.
As we are speaking of hangs: since 1.3.36 there are occasions where
processes fail to terminate. The usually end up occupying no memory
(RSS=0) but using up all available CPU cycles. Unfortunately, they
can't be killed other than through a reboot. It's probably a simple
bug someplace, but I don't know yet how to reproduce it, so there
hasn't been a chance to hunt it down yet.
--david