Hi Steve,
no ideas here, I seem to have the same problem here...
About 90% of the time it works fine here. It is NOT after the link has been
established (running only 56k modem here). It can be somewhere 'in the
middle of web browsing', the link might be up for 20 or 30 minutes already.
I am running bind 8.2.2, ppp 2.3.8. I have setup 4 forwarding DNS servers in
my named.conf. It looks to me like a DNS problem.
I have no tcpdump yet but when it happens that the browser is not responding
normally, I can start a nslookup utility from the same client PC and a site
name like ozbytes.linuxberg.com is not resolved properly. You see that all
root servers get asked.
After a while, when it is back to normal, the nslookup for the same name
will be resolved fine.
Anybody who knows more about Bind/DNS:
Could it be that may forwarders are refusing requests now and then? Are they
allowed to do so?
Sorry Steve, even more questions ...
Michael
P.S.: I hope it's not an 'NZ only' problem ;-)
> I have a problem with diald and named (local zone and
> caching) (and maybe
> ppp??!)
>
> 75% of the time this works fine .... below is a description
> of when it does
> not work!!
>
> I have a fixed ip and I am using an ISDN modem. I am also
> running squid
>
> The following events take place (when it all goes wrong)
>
> 1 / Open browser to connect to the Internet (it is set to
> proxy through squid)
> 2 / In dialmon I watch the link begin to come up
> 3 / Immediately I get my first packet ... a dns lookup to my
> primary dns server
> 4 / The link is almost up and I get my second dns lookup to
> my secondary
> dns server
> 5 / The link is up (elapsed time around 8 secs) and I get my
> third dns
> query (it is stepping through the root servers now)
>
> 12:10:25.513439 210.55.xxx.xxx.4374 > 208.10.192.227.www: S
> 3659520171:3659520171
> (0) win 32120 <mss 1460,sackOK,timestamp 51750547[|tcp]> (DF)
> 12:10:28.503716 210.55.xxx.xxx.4374 > 208.10.192.227.www: S
> 3659520171:3659520171
> (0) win 32120 <mss 1460,sackOK,timestamp 51750847[|tcp]> (DF)
> 12:10:31.117948 210.55.xxx.xxx.1455 > 202.27.184.3.domain: 45154+ (45)
> 12:10:34.503725 210.55.xxx.xxx.4374 > 208.10.192.227.www: S
> 3659520171:3659520171
> (0) win 32120 <mss 1460,sackOK,timestamp 51751447[|tcp]> (DF)
> 12:10:35.004115 210.55.xxx.xxx.1455 > 202.27.184.5.domain: 45154+ (45)
> 12:10:39.004057 210.55.xxx.xxx.1455 > 208.10.192.67.domain: 45154 (45)
>
> eg above is tcpdump of tap0.
>
> 6 / Now everything goes to custard. Instead of a replay from
> a dns server
> coming back, and my http request going out, I get about 30
> secs of stepping
> through the root dns servers. Then all of a sudden (after
> around 30 secs)
> I get (i presume) a dns reply and connect to the site. From
> this point on
> the link is fine, dns is fine etc.
>
-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]
