pcp
[Top] [All Lists]

Re: PCP start/stop script regression

To: nscott@xxxxxxxxxx
Subject: Re: PCP start/stop script regression
From: Max Matveev <makc@xxxxxxx>
Date: Fri, 25 May 2007 13:45:12 +1000
Cc: markgw@xxxxxxx, kimbrr@xxxxxxx, pcp@xxxxxxxxxxx
In-reply-to: <1180064585.6273.582.camel@edge>
References: <1180062348.6273.575.camel@edge> <4656577B.4020105@sgi.com> <1180064585.6273.582.camel@edge>
Sender: pcp-bounce@xxxxxxxxxxx
>>>>> "nscott" == Nathan Scott <nscott@xxxxxxxxxx> writes:

 nscott> On Fri, 2007-05-25 at 13:26 +1000, Mark Goodwin wrote:
 >> 
 >> Hi Nathan, yes that's definitely a regression; we'll get it

 nscott> The patch I'm using atm is attached - its a minimal change.  From
 nscott> what I can tell, Max had a bunch of other PMDA kill logic in there
 nscott> which has gone in 2.7.x - but I don't know any of the history here,
 nscott> so can't help much more.
Killing logic was reworked to deal with pmcd being killed when pcp was
installed into chroot build environment, aka "serial killer" problem.
It just wasn't reworked sufficiently to deal with pecularies of rpm
and backward compatibility.

max

PS. Your patch is going to re-introduce the serial killer problem.

<Prev in Thread] Current Thread [Next in Thread>