I have a test machine running onnv-gate:2007-05-15 that occasionally
appears to hang after rebooting. It prints the OS banner and associated
stuff, but then it stops doing anything else. It happened again last
night, this time while booted under kmdb, and it's clear that there's
not much going on:
[1]> ::ps
S PID PPID PGID SID UID FLAGS ADDR NAME
R 0 0 0 0 0 0x00000001 fffffffffbc25670 sched
R 3 0 0 0 0 0x00020001 fffffffec9a163a8 fsflush
R 2 0 0 0 0 0x00020001 fffffffec9a123a8 pageout
R 1 0 0 0 0 0x4a004000 fffffffec9a103a8 init
R 100006 1 100006 100006 0 0x42000000 fffffffeca82e3a8 svc.configd
R 100004 1 100004 100004 0 0x42000000 fffffffeca7763a8 svc.startd
It's been like this for ten hours.
26 of svc.startd's threads are executing lwp_park(), i.e. they're
waiting on some user-level synchronization primitive. I couldn't figure
out what type of primitive it is, glancing at the kernel state, and as
far as I can tell ::context doesn't work in kmdb. How can I gather more
data about what's going on? If someone inside of Sun wants to take a
look at the problem, I can give you the console connection.
Blake
