Problems with fetching mail

[atomant]

Hi!

I am using rollernet.us as backup MX. Lately I noticed error in my log
when it do fetchmail. The line is:
----------
fetchmail: Query status=2 (SOCKET)
fetchmail: Node localhost not allowed: 459 <localhost>: service unavailable
fetchmail: Query status=3 (AUTHFAIL)
----------
Before it was working and now suddenly it doeasn't.
What am I missing here? Anything to change or what?

[RollerNetSupport]

How are you calling fetchmail? (What commands/options?) We only support ETRN at the moment.

[atomant]

fetchmail -p ETRN --fetchdomains my.domain mail.rollernet.us

[RollerNetSupport]

fetchmail -p ETRN --fetchdomains my.domain mail.rollernet.us

That's teh correct way to do it. The error message being returned means that whatever domain is being specified for the ETRN command resloves to (or somehow ends up being sent as) "localhost". That's the same as me telnetting to port 25 on mail.rollernet.us and doing this:

Code: Select all

Trying 67.118.43.92...
Connected to mail.rollernet.us.
Escape character is '^]'.
ETRN localhost
459 <localhost>: service unavailable
quit
221 Bye
Connection closed by foreign host.

If the domain is not a valid domain as far as the account manager is concerned, then it will return the "service unavailable" message. Since "localhost" is not valid, it returns the error message above. Verify that the domain is enabled in the MX section of the account manager, and also try http://dnsreport.com/ to run some various tests on the domain.

[atomant]

All domains are active.
Well, my domain is actually a free subdomain from dyndns.org.
I am just wondering why it suddenly stops working because it worked before without any problems.

[atomant]

I am desperate.
I am still unable to fetch mails from rollernet. Now I have in my logs following:

fetchmail: Queuing for localhost started
fetchmail: Polling lists.homelinux.net
fetchmail: Node localhost not allowed: 459 <localhost>: service unavailable
fetchmail: Query status=3 (AUTHFAIL)

Why it is telling that queuing started for localhost? Where does it get this information?What do I need to change on my server so it will work again?

[atomant]

More info on this problem:
if I run fetchmail from console I get normal response:

# fetchmail -v -p ETRN --fetchdomains lists.homelinux.net mail.rollernet.us
fetchmail: 6.2.5 querying mail.rollernet.us (protocol ETRN) at Tue 23 Nov 2004 09:27:18 PM CET: poll started
fetchmail: SMTP< 220 mail.rollernet.us ESMTP Postfix
fetchmail: SMTP> EHLO localhost
fetchmail: SMTP< 250-mail.rollernet.us
fetchmail: SMTP< 250-PIPELINING
fetchmail: SMTP< 250-SIZE 51200000
fetchmail: SMTP< 250-VRFY
fetchmail: SMTP< 250-ETRN
fetchmail: SMTP< 250 8BITMIME
fetchmail: ETRN> ETRN lists.homelinux.net
fetchmail: ETRN< 250 Queuing started
fetchmail: Queuing for lists.homelinux.net started
fetchmail: Polling mail.rollernet.us
fetchmail: ETRN> QUIT
fetchmail: SMTP< 221 Bye
fetchmail: 6.2.5 querying mail.rollernet.us (protocol ETRN) at Tue 23 Nov 2004 09:27:19 PM CET: poll completed
fetchmail: normal termination, status 0

but if I run it thru CRON then I get:

fetchmail: Queuing for localhost started
fetchmail: Polling lists.homelinux.net
fetchmail: Node localhost not allowed: 459 <localhost>: service unavailable
fetchmail: Query status=3 (AUTHFAIL)

CRON runs the fetchmail as root as I was when running directly from console.

Any idea what is wrong here?

[RollerNetSupport]

The only thing I could think of is the polling domain and polling server are reversed, or an option is missing in the cron command; perhaps a format error in the crontab line? What's happening is the same as if you had done:

fetchmail -p ETRN --fetchdomains localhost lists.homelinux.net

OR

fetchmail -p ETRN lists.homelinux.net

Throw the -v into the crontab based one and see what debug info it gives.

Even in the absence of fetchmail, queued message will still be delivered to you under the auto-retry cycle. Fetchmail just causes an immediate response rather than waiting for the mail server to retry on its own; it is not required to get the queue.

[atomant]

SOLVED.
Like you have said, the problem was in the cron command line - I forgot to add "--fetchdomains" in the line after ETRN.

Thanx for all the help.

Great service.

[RollerNetSupport]

No problem - glad it's fixed.