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

[Amos Jeffries]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 13/02/2015 9:25 a.m., Yuri Voinov wrote:
> BTW,
> 
> now I have average 300-400 Mb total memory consumption over all 
> working instances of squidGuard. And see only one problem - idle 
> processes never dies after peak hours, and therefore idle=
> parameter is meaningless.

... and yet you just described in the other email how happy you were
not to have 100 helpers running constantly. Its the designed idle=N
behaviour which does that, the other parameters are just limits
bounding where the idle/standby initializer starts and stops having
effect.
Just set idle=100 and see how much meaningless it is.

It was created for a specific purpose, which it meets quite well. Its
documented as being what it is: the *minimum* amount of idle processes
to have running at any one time. And thats how it operates.

The helper itself is free to exit early if it needs to. Just not too
often or in big batches.

If you look at the numbers of objects requested when loading a page
like Facebook (240 objects) or CNN (170 objects) you will see that 60
is not that large a number of helpers anyway. The proxy can go from no
traffic to 60 concurrent requests in less than a second *at any time*.

Amos
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)

iQEcBAEBAgAGBQJU3TvYAAoJELJo5wb/XPRjLu8IAI+EvtMVaalfyUCCfjCeSwpd
o5idmVNF3NokuOZpysEeb0kU7LPedCxQJitbBVZbLLuRXxaTkPC+KnjkbMmH1YUs
FrzyBKznO7K7Onv5UxtJkZrNhvFseLDG3l643IW8KMdpVfEOcKNAib9QUcY4vbuR
w9yjx25VciSoSiEFpGA+MAr9aHkkgwr7b9tWHU7czEmlx60hdTkc1v+5/wbgNTSP
0B+BkRIH3aQCSqt/N+skW/5SFurS6iXl24JH3eO7SrRNNRVR8zuRVE4rRMIm14r5
k+6wrWUrbRfXKKMRG4OwTuCmFUJelw/U8Q6Bs6IDy7fI58J2Yswf6zcZKF+zkc0=
=3vU8
-----END PGP SIGNATURE-----