On Mon, 8 Jan 2007, Derrick J Brashear wrote: > On Mon, 8 Jan 2007, Russ Allbery wrote: > > > Derrick J Brashear <[EMAIL PROTECTED]> writes: > > > >> So basically, what *should* happen is that afsd -shutdown should cause > >> it to die, bt doesn't, because it's a new invocation of afsd. So it > >> can't remember rmtsysd's pid > > > >> I could do something where a child like the afsd handler holds in a > >> syscall until it gets a shutdown call, and then kills rmtsysd by pid. > > > >> That's the best idea I have. Other thoughts? > > > > That would work. We could also just write the PID to a file and have afsd > > -shutdown read it back from the file to kill the other process. > > Sort of wretched. Not really more wretched than the thing I proposed but I > considered and sort of made a face at it already
For better or worse, that's the way a lot of daemons handle it on Linux... have a look at the number of *.pid files in /var/run Steve Brown [EMAIL PROTECTED] _______________________________________________ OpenAFS-info mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-info
