RE: [squid-users] problems upgrading from squid 2.4stable6 to 2.5stabel2

today we did some tests:

we wrote a small script that accesses pages from two sites.

1) squid 2.5.stable2, one client 18s and 38s
2) squid 2.5.stable2, each site 5 clients 2:11min and 3:40

now we dropped all our acls,

3) squid 2.5.stable2, one client 4s and 12s
4) squid 2.5.stable2, each site 5 clients 28s and 56s

the last test was with squid2.4stable7 AND ALL acls (should
be the same like in 1+2)

5) squid 2.4.stable7, one client 6s and 13s
6) squid 2.4.stable7, each site 5 clients 24s and 45s

so this looks like our acls were the "killer". but the question is
what causes squid to hang so badly. squid.conf for 2.4 and 2.5 are
the same (ok, apart from the minor changes for the version, but
acls are the same). so squid2.5 only performs the same as the old
version when we drop all our acls...

do you have any idea?

markus
      

-----Ursprüngliche Nachricht-----
Von: Marc Elsen [mailto:marc.elsen@imec.be]
Gesendet am: Donnerstag, 22. Mai 2003 08:56
An: markus.rietzler@rzf.fin-nrw.de
Betreff: Re: [squid-users] problems upgrading from squid 2.4stable6 to
2.5stabel2

B.11.00 A 9000/803
ram is 512 MB

where using aufs, cache_mem set to 32MB
cache_dir aufs /rzf/db/www/cache 512 16 256

we're using a large number of acls.

....

    What happens if the acl's are not being used ?

    M.

-- 
 'Love is truth without any future.
 (M.E. 1997)