Greetings pluggers,
So I got on the broadbandwagon. Running a Fedora Core 1 SOHO server for over an year. Never bothered to upgrade to FC2 or FC3. It aint broke yet.
There are a few PCs behind this box on a private LAN through eth0. eth1 goes to Smartax MT800. rp-pppoe. Running a caching BIND too rather than going through bsnl dns. AND running sendmail for direct delivery. (To my knowledge bsnl dont have an smtp server yet!) This brings us to the convenience of hanging on to a "good" dynamic IP for a long time. Many mail servers dont accept mails from tainted IPs that make it to spamhaus' sbl/xbl lists or are deemed to be open relays. You can find out if your current IP is blacklisted from spamhaus.org's webpage. If it is not listed, it would be great if we get to keep it forever. Otherwise mails to vsnl.net, yahoo.com, aol.com etc are rejected. Needless to say smtp.vsnl.net refuses relay connections from dataone.
I presumed that if you stayed connected your dynamic IP will stay put. NOT! If there is little activity, it reconnects after 30 minutes or so. I found that MT800 configuration has an "idle timeout" setting. Apparently bsnl guys changed it to 30 minutes (from the default 0 that is supposedly for no timeout). Fine. I changed it to 0. Doesnt work. Then I changed it to an arbitrary high value. Nada.
OK. A simple trick then. Set up a cron job to ping somebody every 5 minutes. Nyet. Time for sterner measures. Set up a cron job doing
wget -O - http://www.google.com
every 20 minutes. No cigar. I run SQUID too. Thought wget might be fetching from squid cache. But apparently wget bypasses squid. So bsnl's "inactivity" flags should have been reset by this.
With this I have exhausted all possibilities. So I am back to square one. PLUG Gurus, how do you keep a dataone IP address longer than 30 minutes?
Additional info:
1. During daytime, with frequent web access, the IP stays with me longer. Several hours in some cases.
2. In /var/log/messages pppd says the reason for disconnecting is as follows:
Sometimes LCP terminated by peer. Modem hangup.
Other times LCP terminated by peer. ioctl(PPPIOCSASYNCMAP): Inappropriate ioctl for device(25)
Apparently the probability of these two is 50-50.
3. I have also tried to change CONNECT_TIMEOUT and PPPOE_TIMEOUT values in /etc/sysconfig/network-scripts/ifcfg-ppp0. Makes no difference.
I would appreciate any suggestions.
Uday
-- ______________________________________________________________________ Pune GNU/Linux Users Group Mailing List: ([email protected]) List Information: http://plug.org.in/mailing-list/listinfo/plug-mail Send 'help' to [EMAIL PROTECTED] for mailing instructions.
