Receiving POP email is delayed for hours

wscreate

Active Member
Aug 14, 2001
36
0
306
Several customers have reported that their POP email is delayed for hours, sometimes 2, 3 and sometimes 8-10 hours after being sent.

I have duplicated this problem multiple times. I have verified that the server is not overloaded at the time of delivery. I have sent locally and it seems not to have a problem, only when the sender is not an account on the server.

Mail is NOT being bounced so I know it is not the /etc/localdomains problem.

Has anyone got any ideas about what could be wrong?

Thanks
 

wscreate

Active Member
Aug 14, 2001
36
0
306
Over a month and no replies. Some things we Cpanel customers have to jsut figure out on our own.
 

Annette

Well-Known Member
PartnerNOC
Aug 12, 2001
445
0
316
\"Duplicated this problem\" how? Did you watch the mail log as you sent mail to see if any errors were generated? Is there anything in /var/spool/exim/input hanging out, waiting for a retry? If there is, have you done a /usr/sbin/exim -qf to force an unfreeze and then watched the mail log to see what happens when the system tries to unfreeze and deliver the mail? What domain(s) is the mail going to? Are you able to do an nslookup on that domain? Have you tested one of the big dogs like Yahoo? You need to provide a bit more information about what steps you have taken to whittle down the possible causes of the problem before anyone can even begin to say definitively what the solution would be. I would not say it\'s \"pretty sad\" that DarkOrb doesn\'t have an answer. I would say that this type of thing should be presented with more information for those not using the Psychic Network as their backbone if any help is expected.
 

wscreate

Active Member
Aug 14, 2001
36
0
306
DNS lookup indicates the dns is fine. There is nothing in the exim log to indicate a problem concerning the email being undelivered. Email is \"simply\" not being delivered and there is not an indication in the logs as to why.

My question to you, Annette, is why should we have to unfreeze the email? Why would exim freeze email on a server that is not overloaded. The server load average of this server is consistently under 1.00. Is exim really that bad? My experience indicates that it is. Your suggestion is that this error needs to be studied more extensively. I have done this. My suggestion is that exim should be reconsidered as the main email software included with CPanel. It appears to be plagued with \"inexplicable\" problems.
 

rpmws

Well-Known Member
Aug 14, 2001
1,787
10
318
back woods of NC, USA
It seems I am having similiar problems with a server :-( no explanition just takes forever ..no loads.
 

bens

Member
Aug 15, 2001
24
0
301
Ditto to that.
It only happens to about one or two domains on the entire server. However, it\'s still weird and I can\'t figure out what is causing it..

[Edited on 10/30/01 by bens]
 

Annette

Well-Known Member
PartnerNOC
Aug 12, 2001
445
0
316
The load on a server has nothing to do with exim freezing or queueing mail, really, unless it\'s so high that exim has to wait its turn for a chance to get part of the processor time - but that would have to be pretty severe. We have had a single case of mail being delayed for hours because the remote gateway was closing the connection too quickly (right after the RCPT TO header, in fact), and there\'s not a lot that we can do about that except contact the postmaster of the recipient system to ask them what\'s going on. We haven\'t really had any problems with exim - or at least none more than we have with say, sendmail. Every MTA has its own issues, but it\'s hard to lay this problem on exim without seeing something that would indicate exim is rejecting or not even making the attempt to deliver the mail right off the bat.

Things I\'d check (that I did check, after finding that single domain with the delay issue):

Do you actually have mail queued up for the domain that is not receiving the mail? Or is the mail going out, according to the logs? If the former, did you take a look at the message in /var/spool/exim/msglog for that particular message id to see what exim says the problem is (usually the first line, since the remainder will be the standard \'retry time not reached\' stuff)? If the latter, then it doesn\'t really have anything to do with exim at all - if it\'s making it off your server, you really can\'t do anything about the routing once it leaves.

If you do have something in /var/spool/exim/msglog, what does it say for those messages? Have you attempted delivery to the problematic address while watching to log roll? What results are posted to the log when you try? Can you post a snip from the log (names obscured to protect the innocent, of course :))?

When you send mail through webmail on one of the affected domains, is the behavior different? When you send mail to the account on the system from somewhere else, is the mail likewise delayed? How about from the domain to which they are trying to send mail and finding it delayed? If it\'s delayed both ways, it isn\'t the server, it\'s the routing.

There\'s more, but that\'s a place to start. Believe me, although my initial message may have seemed harsh since I left off the little smiley after the Psychic Network piece, more information is really needed to properly troubleshoot things like this.