"Server not responding" error when checking e-mail


#1

A few hours ago, all of the people at my domain who use Eudora as their e-mail client (Macintosh and Windows both) began having problems checking e-mail. When they try to POP mail down, they get the following error:

“Server not responding. 706; Lost connection to server.”

However, they are still able to send mail.

I thought it might be some kind of authentication problem. I’ve tried with and without SSL (default is without.) I’ve tried overlapping POP3 commands and not overlapping POP3 commands. Even tried using IMAP instead of POP. Nothing works.

Other mail clients are not affected.

There have been no changes made to the Eudora client configurations on this end, so I figure something changed on Dreamhost’s end of things. I submitted a request to DH Support about five hours ago, and I know from experience they usually take much longer than that to reply, but I have annoyed users starting to breathe down my neck, including the President of my organization, the CFO, and at least one Program Director. Ulp! (At least WebMail is still working!)

Hoping someone here might have an idea as to what’s wrong. Anyone?

Thanks,
David


#2

Can you compare the session logs from Eudora with that of other clients?

:cool: Perl / MySQL / HTML+CSS


#3

Same here - Apart from all users trying to connect pop/imap using several variations of Outlook & Outlook Express.
Able to send, but not receive anything on all addresses belonging to a single domain.
Have contacted support a while ago with no response, as yet.
I too have many irate clients and superiors who are frantically requesting I get it fixed !!!
As per the OP - Nothing has changed on any of the client systems.
The only thing that was noticed, is the errors started just after receiving the latest dreamhost announcement below !
Coincidence or not - Help anyone please !!!

[quote]
-Upgrade to Debian 3.1 (Sarge) on your server is complete-
In an effort to provide you with better support for new software, we have upgraded
the Operating System Software on your server, larry, to the new stable Debian
release today. Total downtime was only a couple of minutes and affected shell, ftp,
jabber, and web services. Mail was not affected during the upgrade.

The preceding was a Upgrade announcement, sent 2005-10-05 15:13:51. “DreamHost Announcement Team” do.not.reply@dreamhost.com


#4

OK, everything seems to work this morning, but I went ahead and reviewed session logs from both Eudora and Outlook Express anyway. I’m not techie enough to know whether looking at them might provide a clue as to what could have been going wrong yesterday, but I submit them nonetheless, in case those more knowledgeable than myself can glean some useful information.

However, before I paste that stuff, I have to share the reply that arrived in my inbox from DH Support last night at 9:53 PM:

[quote]Well, if webmail is working then the accounts are fine on this end - I
would suggest having them log into webmail and remove all junkmail
(that usually removes the message causing the pop client to choke -
if that doesn’t do it just have them move all messages from the inbox
or delete them and that should also clear things up).

[/quote]

Is it just me, or is this completely inane? There might be some junkmail “causing the POP client to choke?” Then advising me to delete the inboxes of my senior staff!

As it turns out, I keep a dummy, non-published e-mail account at my domain that I sometimes use for troubleshooting things like this. It had (and generally has) no e-mail in it’s inbox, so I never followed DH’s suggestion, not that I think I ever would have. Does what they wrote actually make sense to anyone – that a piece of junkmail could cause one specific POP client, in this case Eudora, to “choke” in such a way that generates a server-specific error like I described? Technically speaking, what the heck is “choking” supposed to mean?

Am I being uncharitable?

Anyway, rant over. Here are the session logs.

EUDORA:
Thu Oct 06 09:39:32 2005
Version 6.2.3.4
LogLevel 127 (0x7F)
Mode 0
5932 16:0.0 Open 66.33.213.158:110
5932 64:0.0 Rcvd: "+OK Hello there.\r\n"
5932 32:0.0 Sent: "CAPA\r\n"
5932 64:0.0 Rcvd: "+OK Here’s what I can do:\r\n"
5932 64:0.0 Rcvd: "TOP\r\n"
5932 64:0.0 Rcvd: "USER\r\n"
5932 64:0.0 Rcvd: "LOGIN-DELAY 10\r\n"
5932 64:0.0 Rcvd: "PIPELINING\r\n"
5932 64:0.0 Rcvd: "UIDL\r\n"
5932 64:0.0 Rcvd: "IMPLEMENTATION Courier Mail Server\r\n"
5932 64:0.0 Rcvd: ".\r\n"
5932 32:0.0 Sent: "USER mXXXXXXX\r\n"
5932 64:0.0 Rcvd: "+OK Password required.\r\n"
5932 32:0.0 Sent: "PASS *******\r\n"
5932 64:0.0 Rcvd: "+OK logged in.\r\n"
5932 32:0.0 Sent: "STAT\r\n"
5932 64:0.1 Rcvd: "+OK 0 0\r\n"
5932 32:0.1 Sent: "QUIT\r\n"
5932 64:0.1 Rcvd: "+OK Bye-bye.\r\n"
MAIN 8:0.2 Dialog: "Sorry, you don’t have any new mail."
MAIN 8:0.4 Dialog: “Dismissed with 1”

OUTLOOK EXPRESS:
Outlook Express 6.00.2800.1506
POP3 Log started at 10/06/2005 09:56:18
POP3: 09:56:18 [rx] +OK Hello there.
POP3: 09:56:18 [tx] USER mXXXXXXX
POP3: 09:56:18 [rx] +OK Password required.
POP3: 09:56:18 [tx] PASS *******
POP3: 09:56:18 [rx] +OK logged in.
POP3: 09:56:18 [tx] STAT
POP3: 09:56:18 [rx] +OK 0 0
POP3: 09:56:18 [tx] QUIT
POP3: 09:56:18 [rx] +OK Bye-bye.


#5

Does what they wrote actually make sense to anyone – that a piece of junkmail could cause one specific POP client, in this case Eudora, to “choke” in such a way that generates a server-specific error like I described?

Yes, malformed messages (typically spam) can cause POP3 clients to hang. I experienced this myself (well, my GF, actually) just a couple weeks ago. Logging in via webmail and deleting all her spam did fix it.

You don’t have to delete legimate mail. Clear out all the obvious spam first. If that doesn’t fix it, move all the mail in their inbox into a temp folder. If it works now, move the mail back a little at a time until you determine which message is causing the problem.

Also, you think that Eudora message is unhelpful? Apple Mail doesn’t throw an error, it just stops, while continuing to check other accounts just fine. It took me half an hour to track that one down.


If you want useful replies, ask smart questions.


#6

A couple of releases of Eudora ago, it had a serious problem where the app would crash on messages with long subject lines using non-ASCII characters (or something like that). At a previous job working on small business mail servers, we ran into issues like this with other mail clients as well. So it’s definitely possible for a single email to cause a problem like that.

I think the comment from support is reasonable in that it helps determine where the problem lies. A couple weeks ago DH had a brief problem where the POP/IMAP servers would abruptly close the connection after receiving a correct password. In Eudora, this could show up as a Server not responding error. I believe Webmail relies on IMAP internally, so if Webmail it working, it would provide evidence that DH was not the issue.

If you find the problem is with a single message, then you can use a different mail client (or telnet to the POP server) to fetch/delete that message.


#7

Thanks to all for the perspectives on Dreamhost’s reply. Good stuff to know. I apologize to DH Support for being so dubious.

However, I think that one thing I clearly didn’t convey in my most recent post is that I maintain a non-public account at my domain only for testing or troubleshooting mail related issues (literally "bogus@mydomain.org"), and it normally has NO messages at all in the inbox. And it was definitely empty yesterday. Even so, the “lost connection to server” error still occurred when I checked that mailbox with Eudora, though not so with other clients.

So, given that the problem occurred even when trying to check an empty inbox, it would appear there’s no way it could have stemmed from a malformed spam message or a non-ASCII character in a subject line. Maybe I’m missing something?

I still hold that it was some kind of problem with authentication, perhaps some incompatibility between the manner or order in which Eudora transmits user and password information, and the way Dreamhost’s server expects to receive that info. But with that, I’m just making guesses beyond the limits of my expertise.

Like I said before though, the problem was fixed when I came in this morning. Whether or not someone at DH did something, or the problem “just fixed itself,” I don’t know. The question, I suppose, is moot.

Thanks again,
–d.


#8

[quote]I have to share the reply that arrived in my inbox from DH Support last night at 9:53 PM:
Well, if webmail is working then the accounts are fine on this end

[/quote]

This of course is rubbish. Your failure is on POP3 - DH webmail does not use POP3.

[quote]I would suggest having them log into webmail and remove all junkmail
(that usually removes the message causing the pop client to choke

[/quote]

Good grief. One wonders if Support even bothered reading your report properly.

[quote]Is it just me, or is this completely inane?

[/quote]

It is. I guess to avoid DH dumping the blame on your POP client, you’ll next time have to try it with a few different ones. www.mail2web.com and www.pop2web.com shoud do.