Re: [squid-users] is anyone listening to my queries summary o f events

From: khiz code <khizcode@dont-contact.us>
Date: Mon, 20 Aug 2001 23:31:21 -0700 (PDT)

Hi all
sorry for disturbing u all once more but there r certain doubts which
need to be put at rest

1> getting the "Resource Temporarily unavailable" message is NoRMAL as
per Henrick
   " You only get this message because you have enabled more detailed
debugging.
The message is expected, and is printed at debug level 2."

2> as per adrian's and robert's suggestions i did install squid2.4S1 on
a plain vanilla redhat6.2(zoot) kernel 2.2.14-5.0
  compile optyions
--enable-snmp, --enable-storeio=ufs,diskd
 this particular machine is not directy connected to the internet so i
used a parent proxy using the foll config
both are 2.4S1

cache_peer 10.1.11.18 parent 3128 0 no-query default
never-redirect allow all
child ip = 10.1.11.225
parent ip 10.1.11.18

browser used child proxy as its proxy
 cache.log reports
a > of child proxy
   after almost every request we get the same
 2001/08/21 11:05:56| clientReadRequest: FD 17: no data to process
((11) Resource
 temporarily unavailable
b> cache.log of parent proxy
 NO SUCH "Resource unavail Messages"
not even once !!!!
  
===== now isnt this strange ======
so user clients like IE do hv problems but majority of ppl use MS iE
my particular version is Ie5.5

3> case 2 is some what similar to my earlier posting rgd web-polygraph
test when i didnt get these "resource unavailable errors" when
subjecting squid to polygraph test
demo.pg
  Robot R = {
        kind = "R101";
        public_interest = 50%;
      req_rate=7/sec;
        pop_model = { pop_distr = popUnif(); };
        addresses = ['10.1.11.225' ** 200];
        origins = S.addresses;
};
ploygraph showed ne abt 130 -150 replies/sec which was correctly shown
in mrtg!!!
4> to enable transparentrouting i had used ipchains on my machine
ipchains -A input -s localhost/32 -d localhost/32 -j ACCEPt
ipchains -A input -p tcp -d 0/0 80 -j REDIRECT 3128
ipchains -A input -s 0/0 -d 0/0 -j ACCEPT

under load i cpatured netstat -an
then did grep SYN_RECV | wc -l
the o/p was abt 1024 to 1030
this is at putting the machine atload for not more than 10 mins
does this indicate a SYN flood
do i need to increase tcp_max_syn_backlog in /proc
or do i need to change tcp_syncookies

so to summarize it all i just say that browsers give me reosurce
unavailable problem

 so joe .. i hv followed adrians advice.. (tried out stable 2.4 and not
daily snaphot..sorry ...)
tried using 2.4S1 without tranparent proxying now what remains to be
done is remove raid 0 and see if that makes any difference
btw i hv the tlan driver module

waiting for any further guidance

still optimistic,
khiz

--- Joe Cooper <joe@swelltech.com> wrote:
> khiz code wrote:
>
> > i am using squid 2.2 becoz it is the fastest squid on the block
>
>
> It's not the only option anymore, for performance...(And my article
> that
> made the assertion that 2.2 was the fastest now has a "Note" in
> bright
> red letters attached to it, explaining that 2.4 is now just about
> equal.)
>
> Besides, if someone (Adrian and Robert) who know Squid better than
> anybody suggest you try the new version and report back about what
> happens...it is /wise/ to follow their advice. Even if you revert
> back
> to 2.2 after figuring out your problem.
>
> No one is going to be able to help you if you don't follow through
> with
> the steps people are giving to help track down the problem. (And no
> one
> is going to continue to try, if your only response to these
> suggestions
> is to repost the same questions over and over, without trying the
> advice.)
> --
> Joe Cooper <joe@swelltech.com>
> Affordable Web Caching Proxy Appliances
> http://www.swelltech.com
>

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/
Received on Tue Aug 21 2001 - 00:31:23 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:01:51 MST