[torqueusers] torque daemons can not be stopeed during uninstall

onceihave onceihave at 163.com
Tue Dec 9 17:59:59 MST 2008


ÔÚ2008-12-10£¬"Jonathan Billings" <jsbillin at umich.edu> wrote£º
>Garrick Staples wrote:
>> On Tue, Dec 09, 2008 at 10:08:03PM +0100, Bogdan Costescu alleged:
>>>> It doesn't matter if you think they should happen or not.
>>> Well, if such an argument is used as part of a discussion, there is 
>>> no discussion anymore. End of topic for me.
>> Don't be offended.
>> I brought up 2 possible scenerios to illustrate why I don't think the upstream
>> rpms shouldn't stop daemons.  You said admins shouldn't do those things.  You
>> may be right, but it doesn't stop the fact that they do.
>I tend to think that not stopping the daemons when a package is upgraded 
>falls under the principal of least surprise[1].  It's probably not a 
>great thing to do in production, but at least it isn't causing as much 
>problems as if the daemons were shut down.
Well. Would you point out the problems if we stop daemon during uninstall?

?>It makes sense when a package is *removed* to shut down the daemon, but 
>I think the problem we're seeing here is that the process of updating a 
>package involves running the removal script for the old package, thereby 
>stopping the daemon.
RPM script can distinguish uninstall/update by $1. Even in rpm update, restart 
(that means we still need stop daemon first) daemon is reasonable, because the
admin should want the new binaries take effect.


Steven Wang

>1. http://en.wikipedia.org/wiki/Principle_of_least_astonishment
>Jonathan Billings <jsbillin at umich.edu>
>The College of Language, Science, and the Arts
>LS&A IT - Research Systems and Support
>torqueusers mailing list
>torqueusers at supercluster.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20081210/c9561063/attachment.html

More information about the torqueusers mailing list