I seem not to understand what contract deems to be "fatal" and
"non-fatal" signal.
### Case 1: Start up sendmail, set hwerr and signal to be fatal event.
# ctrun -V -r 0 -o noorphan -f hwerr,signal
/lib/svc/method/smtp-sendmail start &
[1] 3558
# ctrun(3558): created contract id 326
# ctstat -vi 326
CTID ZONEID TYPE STATE HOLDER EVENTS QTIME NTIME
326 0 process owned 3558 0 - -
cookie: 0
informative event set: core
critical event set: signal hwerr empty
fatal event set: signal hwerr
parameter set: noorphan
member processes: 3563 3565
inherited contracts: none
Kill SIGTERM one sendmail process, one remains.
# kill 3563
# ctstat -vi 326
CTID ZONEID TYPE STATE HOLDER EVENTS QTIME NTIME
326 0 process owned 3558 0 - -
cookie: 0
informative event set: core
critical event set: signal hwerr empty
fatal event set: signal hwerr
parameter set: noorphan
member processes: 3565
inherited contracts: none
Kill SIGTERM that one too. Everybody dies. And stays dead, no restart.
Despite that critical event included "empty". Must I inform ctrun
somehow that I want it to restart on critical event as well, and not
just fatal event?
# kill 3565
# ctrun(3558): event from contract 326: contract empty
[1] Done ctrun -V -r 0 -o noorphan -f
hwerr,signal /lib/svc/method/smtp-sendmail start
### Case 2: (same as case 1) Start up sendmail, set hwerr and signal to
be fatal event.
# ctrun -V -r 0 -o noorphan -f hwerr,signal
/lib/svc/method/smtp-sendmail start &
[1] 3571
# ctrun(3571): created contract id 327
# ctstat -vi 327
CTID ZONEID TYPE STATE HOLDER EVENTS QTIME NTIME
327 0 process owned 3571 0 - -
cookie: 0
informative event set: core
critical event set: signal hwerr empty
fatal event set: signal hwerr
parameter set: noorphan
member processes: 3576 3578
inherited contracts: none
Kill SIGKILL one sendmail process. ctrun detect signal event, kills
what's left in the old contract, starts a new one.
# kill -9 3578
# ctrun(3571): event from contract 327: process 3578 received a fatal signal
signal: 9 (SIGKILL)
sender pid: 3042
sender ctid: 257
ctrun(3571): fatal "signal" event from contract 327
ctrun(3571): event from contract 327: contract empty
ctrun(3571): failure in contract 327, restarting command
ctrun(3571): created contract id 328
There is a distinction which the contract framework makes, which does
not matter to me. SIGKILL is fatal, but SIGTERM is not. I as a user
don't really care, the end result is the same to me, pids in the
contract eventually all exit. Why is this? Why this distinction, I mean.
And why won't ctrun restart for me on the critical event "empty"?
CT
