Fran�ois Pons ([EMAIL PROTECTED]) wrote:
> Han <[EMAIL PROTECTED]> writes:
> 
> > I just tried to update with a mirror that was down and I broke of
> > the process. After that I tried to get another mirror but that one
> > didn't work since the other process didn't die.
> > 
> > Could some urpmi/rpmdrake check for this via a pid-file or something
> > like that? With the option to kill the previous session?
> 
> If you use urpmi, just Ctrl-C.

But of course ;)

> For rpmdrake, It should not let you back until time out arrives ?
>
> How do you do exactly ?

Well for debugging purposes I started rpmdrake from the console. So I
pinged and found the host down and ctrl-c in the console after which
rpmdrake kept on working in the background. I looked again after 5
minutes and the three of them were still active. I had to kill them
manually. rpmdrake didn't notice other session were undead.

Mirrors should be big server so you can set very short times outs. Say
30 seconds? A ping in advance to test is the server is up would hurt at
all. I mean I know all the tricks with failing programms but every nono
should be able to work with it in a pleasant way. And waiting is
unpleasant.

Furthermore I noticed that during the downloading of packages rpmdrake
shows no progress.

And the warning not to use cooker is still very mild. I think it should
be a hidden feature of rpmdrake (rpmdrake --developer).

Something like that.


Groetjes, Han.
-- 
http://www.xs4all.nl/~hanb/software

Reply via email to