[squid-users] Is Squid can shutdown unused idle redirector's children?

[Yuri Voinov]

Also, gents.

ufdbGuard is cool, but:

- Where is good documentation? I found only one connon PDF. No 
performance recommendations, no administrator's guide - this good piece 
of software not so trivial as squidGuard, i.e., I don't know, how to 
support only used blocking categories databases without rebuilding them 
all, no concepts guide - the architecture of solution is not obvious. 
May be I need glasses, but Reference manual + man pages is not enough 
for average SA's. Not at all will read sources.

- AFAIK, it uses daemon-centric architecture. Well, but different OS 
uses different startup facilities. I want to have possibility to tune it 
up by myself or installation must correct do it on target OS. And please 
note, that not only Linux existing in the world. ;) SystemV init was 
deprecated in some systems years ago. ;) And will be good to document 
all of this in installation guide.

Did you agree?

17.02.15 17:28, Antony Stone пишет:
> On Tuesday 17 Feb 2015 at 11:00, Marcus Kool wrote:
>
>> On 02/16/2015 11:43 PM, Amos Jeffries wrote:
>>> PS. Marcus, perhaps you should go on search around to find distro
>>> maintainers who are publishing SG and convince them to replace the
>>> defaults with ufdbguard. I have to do that periodically to clear up old
>>> Squid versions being forced on users. It helps to find out what bugs
>>> they are patching or struggling with silently as well.
>> For reasons unknown I am not very good in convincing other people what
>> they should do :-)
>> Perhaps it is better to wait for the time that maintainers and admins
>> see for themselves what is best.
> Hm, the problem I see with that is that package maintainers are not always
> admins of the systems using those packages, therefore they're not the people
> who run into the problems caused by outdated packages.
>
> Lots of system admins (the ones who don't appear on this list, for example)
> just assume "it's the current software being provided by my distro, therefore
> it must be the one recommended by the developers", therefore they never find
> out that the developers of the software recommend doing something new, while
> the package maintainers have never moved on from old versions or old
> techniques.
>
> I do think we need to inform package maintainers that what they're doing is no
> longer what the developers recommend, to try to close the gap between the
> people who start from "here's what I get from my distro" and "but this is the
> right way to do things" (which we see often enough on this and similar lists).
>
> When the package maintainer is a developer, this situation generally sorts
> itself out pretty quickly, but this is not often (AFAIK) the case, so
> announcing the current recommendations to the people runnning the distros can
> make a big difference.
>
>
> Regards,
>
>
> Antony.
>