Koozali.org: home of the SME Server
Legacy Forums => Experienced User Forum => Topic started by: Stefan on May 02, 2002, 07:08:50 PM
-
Hi,
I've got a problem fetching multiple email accounts on dial-up. On my ME Server 5.1.2 is Vincent's fetchmail tool running. When I start-up the server he tries to fetch mail but nothing happens. It's the same when the server starts the job on every hour.
But when starting the 'fetchmail' script manually (the line is down at that time) everything works fine and mail is coming in.
Changing the ISP didn't change the behaviour.
I've searched the whole forum finding a solution for that problem. But every workaround I've tried could not fix the problem.
Is there any solution out there?
Regards,
Stefan
-
Hi,
Are you using ISDN? If so, please send your logfiles to bugs@e-smith.com
with description of the problem.
In the mean time, change your /etc/crontab file and change every 'startmail' with 'fetchmail'
Regards,
guestHH
-
Hi RequestedDeletion,
yes i'm using an ISDN-Modem contolled by AT commands. In e-smith version 5.0 everything was okay. Since using 5.12 (on the same hardware) i've got this problem.
Regards,
Stefan
-
Hi Stefan,
Took me a while to catch up. My advise is to send your problems AND logs to bugs@e-smith.com.
I can't help any further for i just don't know. I do know that ISDN is still experimental and connections may vary from country to country, so it's hard for Mitel to keep up with all different curcomstances.
Regards,
guestHH
-
I have a problem with fetchmail.... I enter all the information the server requires to collect from a pop mail domain.
Select multi-drop, enter user name and password. I check /etc/fetchmail to make sure it iis polling the correct server. When I type fetchmail is tells me no mail server has been configured.
Please tell me what i have missed here.
Paul
-
I have a problem with fetchmail.... I enter all the information the server requires to collect from a pop mail domain.
Select multi-drop, enter user name and password. I check /etc/fetchmail to make sure it iis polling the correct server. When I type fetchmail is tells me no mail server has been configured.
Please tell me what i have missed here.
Paul
-
Hi there
After so interesting work hours and loads of help we have discovered the following bug to the startmail script
if /sbin/ifconfig | /bin/grep -q '^ippp' not if /sbin/ifconfig | /bin/grep -q '^ppp' for internal ISDN cards
the crontab was not loaded with roots cron file
run crontab /etc/crontab
We have also found that changes to the /etc/hosts to included the pop domain and IP address.
We hope this helps
-
Please, file a bug report to bugs@e-smith.com.
-
I don't think that modifing ^ppp in ^ippp would work.
With Sync ISDN, ippp is always present, while with standard ppp and diald, ppp appears only when connected.
The purpose of that lines in startmail is to force a connection if not connected.
But ippp0 will be present even if not connected.
A solution could be to look for "P-t-P:0.0.0.0" or /var/run/dialdseq.default containing a 0.
-
thanks for your input with regards to this problem. We have been struggling with ISDN Internal dial up.
The problem from start to finish is this:
We have installed the box as per the the manual. We configured it to collect multidrop mail. The box dials connects and thats it.....nothing else happens. On the portmaster where is connects the only request passed is for DNS.
So any and all input on this problem would be of great help.
Thanks
-
There have been reports of no traffic at all, but I'm not sure.
Do pinging the point2point address work ?
Sync ISDN comes with debug enabled, any hint from syslog ?
Post some excerpts from /var/log/messages.
-
Hi all,
I'm glad this thread has come to life again because I have similar (probably the same) problems with 5.1.2 as the rest of you.
In my case I use a bog standard 56k external dial-up modem connection with mail pick-ups scheduled several times a day via cron job. This is the part which seems to fail because I found (by coincidence) that if I was on line at the same time as a scheduled pick-up occurred then mail was downloaded normally but failed every time from a cold start when it had to dial the connection.
Funny thing is the connection goes quite normally
Dials, Connects, Dynamic IP change, NTPsync, but no fetchmail just an error message saying "no packets received within xx seconds closing link", I've tried extending the fetchmail timeout well beyond 30 secs but no joy.
Anyone else found this ? or
Anyone got an ideas what it does differently from a cold start cron dial-up job ,as opposed to the link already up (like a permenant connection almost) ?
Cheers Rob Thomson
-
I had similar problems about a moonth ago I managed to fix the problem by doing a number of things , although I cannot say for certain which of these in fact solved the problem as I had no time to just experiment.
1. I craeted a template /etc/ppp/ip-up.local/ip-up.local
and placed a line in there to send my ip address to an external account
" /sbin/ifconfig | grep ppp mail -s IP-ADDRESS user@xxx.xxx "
2. I diabled dns look ups by fetchmail by editing the 50multipop file in /etc/fetchmail template
" under the protocal pop3 line I created an exact line saying " no dns"
3. I created my own checkfetchmailscript that checks to see if the link is up and if fetchmail is running , i dont have the scipt with me but if anyone wants it just email me and i will send it.
remeber to expand the templates after you have edited them.
I found that forcing an email out at every dialup helped with the no traffic error
Good Luck
Dean
-
Dean
Thanks for your response.
I have already used the no dns line without success, but not your other points.
I'll give them a bash over the weekend if you could kindly send me a copy of your script.
I have to reconfigure 5.1.2 as it was relegated to a test box in server only mode. 4.1.2 is doing the grunt work as it has faithfully for nearly a year.
Maybe this is was a case of "If it ain't broke"!!!
Cheers Rob,
-
i'd also like to hear what the solution to this is. I have a customer whose dial up line only collects email when it is already online, and it is driving them (and me) mad.
-
brian read wrote:
>
> i'd also like to hear what the solution to this is. I have a
> customer whose dial up line only collects email when it is
> already online, and it is driving them (and me) mad.
The answer to this is usually to bypass DNS. Put the mail server address is as its IP number.
Ed Form
-
Ed, no I'm afraid it is not that problem, I have already set the IP address for the maiI server.
We can only pick up mail if we make sure that the dial up modem is already on (by surfing the web).
Brian
-
Brian wrote:
> We can only pick up mail if we make sure that the dial up modem is already
> on (by surfing the web).
Thats right ! And I've remembered a couple of other things from when I was testing 5.1.2 about three weeks ago now and no notes here so this is from memory.
I noticed after one of many reboots, that it did pickup mail on the first connection but not subsequently.
On comparing logs with the 4.1.2 setup (same config) I noticed that it (4.1.2) starts every connection by calling IP 0.0.0.0
5.1.2 uses IP 0.0.0.0 also the first time but after that it calls the IP received during the dynamic IP change and this is usually different each time it connects.
I also noted from the log comparison, just after the connection is made 4.1.2 does KERNEL ppp whereas 5.1.2 has an error for MOD CHAR_MAJOR xx (sorry can't remember exactly now)
Brian perhaps you could comfirm this if your setup is still running ? Mine is not yet. (this weekend I hope.)
Cheers Rob
-
Rob
I can't get to the logs until Monday or Tues, so perhaps if you can see over the weekend, I'd be grateful
Brian
-
This is a very old thread to revitalise but I trawled and trawled it for answers and it did not solve my little problem. In the end I had to resort to commercial support but looking back the answer is above but somewhat confused....
Internal PCI ISDN modem connecting to a dialup isp Onetel (UK) on a 5.5 box with update 2
When someone (ie support) was entering the e-smith box via PPTP the mail would collect and then when they broke the connection it stopped...... Take my hat off to Paul at LinuxIT - he solved it... down to the ^PPP command in /etc/startmail. An ISDN modem starts up an IPPP so this is not found by ^PPP because the ^ means that it must start with PPP not contain the letters PPP.
Remove the ^ and bingo - you have new mail.
Want to see what is going on.. swap the --silent --syslog in fetchmail for --verbose and the admin postbox should get an email of each successful connection.
Sam
Iam learning slowly, bear with me, stick with it, this one took me about 8 days of messin !