Re: Re: yum.cron for testing & examination

On Mon, 2004-10-11 at 10:28, Ed L Cashin wrote:
> "David L. Parsley" <[email protected]> writes:
> 
> ...
> > Yeah, I found out about the 'read' timeout a few hours later, but
> > decided not to rewrite.  The current version looks pretty readable to
> > me, and since I use 'sleep', the timeout value is anything sleep
> > understands (like 12h).  Do you find the script readable, or should I
> > sprinkle some comments?
> 
> If it was any longer or more complicated comments might be nice, but
> I find it pretty readable as it is.
> 
> I wonder whether killing with SIGKILL is a bit harsh.  I don't know
> enough about yum to know whether it might have some cleanup it needs
> to do.  Maybe a SIGTERM; sleep 2; SIGKILL?

Yeah, that's a good idea.  I went with SIGKILL because on actual
machines I found hung in read(x,...), SIGTERM didn't cut it.  However,
on the off chance it times out while doing something else (something
interruptible?), might as well send TERM first.

If all goes well, I hope to roll this into the next Tao yum, and spin
new i386 isos later this week.

regards,
	David

_______________________________________________
Tao-discuss mailing list
[email protected]
http://mailman.taolinux.org/mailman/listinfo/tao-discuss

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQBBapsk4J4WoZuSf3URAkViAKDmOcejrDGDl+hKWL8va4e7ghryYQCffVw6
XCzyF2s12EzlYZm/MNr/4zQ=
=Auvt
-----END PGP SIGNATURE-----