Bug#515727: exim4-config: running queue on ppp/ip-up hardly delays other ip-up scripts

[Sergey Dorofeev]

On 07.05.2010 21:01, Andreas Metzler wrote:
> On 2010-05-06 Sergey Dorofeev<sergey at prodo.ru>  wrote:
>    
>> Package: exim4-config
>> Followup-For: Bug #515727
>>      
>
>    
>> When there are messages in queue addressed to domains with
>> unresponsible MX servers, runq may take VERY long time.  So others
>> scripts in ip-up.d must wait (e.g. adding routes), and taking up
>> interface may become unacceptably long.  Easy solution - manually
>> adding "&" to end of line which starts runq in /etc/ip-up.d/exim4 so
>> queue starts running in background, and ip-up.d continues.  May be
>> it contradicts some policy, I don't know, but it solves the problem.
>>      
> [...]
>
> It is quite some time since /I/ was relying on ip-up/down scripts,
> before geting a flatrate. But at that point of time being sure that
> exim had finished its job before the ip-up script exited was very
> important for me. I kept onlime time minimal, but immediately droping
> the connection again after the ip-up scripts (fetching/sending
> mail/news) had finished. Just as another point of data.
>
> On a sidenote directly delivering to MXes from a dialup system will
> often fail, due to them blocking dialup users.
> cu andreas
>    

In my situation ppp is used only for pptpd/pptp-linux.
So mail delivery is unneeded at all on the connection.
May be exim4-config should ask users, put exim4 script to ip-up.d or not.
Or may be script must distinguish connections, on which it must start 
runq and on which not.