The qpopper list archive ending on 27 Mar 2000
Topics covered in this issue include:
1. Re: "Unable to open Bulletin database" errors
Randall Gellens <randy at qualcomm dot com>
Fri, 17 Mar 2000 09:41:26 -0800
2. Same mail except "X-UIDL" header
Michio SHIZUTA <m-shizuta at fitech.co dot jp>
Mon, 20 Mar 2000 17:25:35 +0900
3. QPopper and Upgrade to AIX 4.3.3
Sherman Wong <Sherman_Wong at baruch.cuny dot edu>
Mon, 20 Mar 2000 08:52:04 -0500
4. Re: QPopper and Upgrade to AIX 4.3.3
Qpopper Support <qpopper at qualcomm dot com>
Mon, 20 Mar 2000 10:00:09 -0800
5.
"Lisa Casey" <lisa at jellico dot com>
Mon, 20 Mar 2000 13:11:54 -0500
6. Re: <no subject>
Qpopper Support <qpopper at qualcomm dot com>
Mon, 20 Mar 2000 12:02:57 -0800
7. envelope problem
Brent <bierblb at netins dot net>
Mon, 20 Mar 2000 19:18:44 -0600
8. envelope problem
Brent <bierblb at netins dot net>
Mon, 20 Mar 2000 19:39:35 -0600
9. Re: envelope problem
Raphael Maseko <ralph at zamnet dot zm>
Tue, 21 Mar 2000 07:57:35 +0200
10. Re: envelope problem
Rick Johnson <rsjohn at directcon dot net>
Mon, 20 Mar 2000 22:32:57 -0800 (PST)
11. Re: envelope problem
Brent <bierblb at netins dot net>
Tue, 21 Mar 2000 01:04:22 -0600
12. Re: envelope problem
"James Nelson" <james at digit.bloomnet dot com>
Tue, 21 Mar 2000 01:20:15 -0600
13. APOP's popauth - w/o prompting for PW?
Jason Englander <jason at interl dot net>
Tue, 21 Mar 2000 03:12:41 -0600 (CST)
14. Re: envelope problem
"Utku Er" <erutku at netone.net dot tr>
Tue, 21 Mar 2000 14:12:26 +0200
15. qpopper3.0 : hashed mailspool implies hashed drop directory
Eric Luyten <eluyten at vub.ac dot be>
Tue, 21 Mar 2000 16:02:15 +0100 (MET)
16. IP reported as localhost
"Jerry O'Brien" <jobrien at cuttingedge dot net>
Tue, 21 Mar 2000 10:29:42 -0800
17. Re: qpopper3.0 : hashed mailspool implies hashed drop directory
"Michael D. Sofka" <sofkam at rpi dot edu>
Tue, 21 Mar 2000 11:36:13 -0500
18. Re: IP reported as localhost
Gerhard Pfeiffer <gp at sun.bnbt dot de>
Tue, 21 Mar 2000 17:49:04 +0100 (MET)
19. RE: IP reported as localhost
Steven Fletcher <stevenf at shellnet.co dot uk>
Tue, 21 Mar 2000 16:58:59 -0000
20. RE: IP reported as localhost
Patrick Nixon <pnixon at somerville.ma dot us>
Tue, 21 Mar 2000 12:43:27 -0500
21. Re: IP reported as localhost
Raphael Maseko <ralph at zamnet dot zm>
Tue, 21 Mar 2000 20:17:38 +0200
22. Re: IP reported as localhost
"Jerry O'Brien" <jobrien at cuttingedge dot net>
Tue, 21 Mar 2000 12:45:03 -0800
23. Qpopper - LDAP
"Fabio Oliveira" <fabio.oliveira at nttb.com dot br>
Tue, 21 Mar 2000 16:02:07 -0300
24. Re: envelope problem
Randall Gellens <randy at qualcomm dot com>
Wed, 22 Mar 2000 17:58:23 +1030
25. Read timeout & maillock
Alexander Rusinov <boot at EuroCom.Od dot UA>
Wed, 22 Mar 2000 14:22:00 +0200
26. Quotas
P O O P I E <poopie at jeeves.poopie dot net>
Thu, 23 Mar 2000 10:55:27 +0000 (GMT)
27. Re: Quotas
Mihigo John <nmj at gorilla.nur.ac dot rw>
Thu, 23 Mar 2000 17:38:31 +0200 (SAST)
28. Re: Quotas
Joel Laing <joel at scripps dot edu>
Thu, 23 Mar 2000 09:18:26 -0800
29. RE: Quotas
roberto.bazan at inycom dot es
Thu, 23 Mar 2000 18:29:33 +0100
30. RE: Quotas
Forrest Aldrich <forrie at forrie dot com>
Thu, 23 Mar 2000 12:34:29 -0500
31. Re: Quotas
Wallace Nicoll <wallace at cityedin.demon.co dot uk>
Thu, 23 Mar 2000 18:38:13 +0000
32. Re: Quotas
Forrest Aldrich <forrie at forrie dot com>
Thu, 23 Mar 2000 13:58:47 -0500
33. RE: Quotas
"Kenneth Porter" <shiva at well dot com>
Thu, 23 Mar 2000 16:56:37 -0800
34. Re: Quotas
Brian Greene <brian at digilect dot com>
Thu, 23 Mar 2000 18:53:08 -0800
35. RE: Quotas
Randall Gellens <randy at qualcomm dot com>
Fri, 24 Mar 2000 17:39:57 +1030
36. RE: Quotas
Alan Brown <alan at manawatu.gen dot nz>
Fri, 24 Mar 2000 19:36:19 +1200 (NZST)
37. RE: Quotas
"Kenneth Porter" <shiva at well dot com>
Thu, 23 Mar 2000 23:51:35 -0800
38. Re: Quotas
Alan Brown <alan at manawatu.gen dot nz>
Fri, 24 Mar 2000 20:16:55 +1200 (NZST)
39. Pam & Radius & Qpopper
Kostas Zourkas <kzou at space dot gr>
Fri, 24 Mar 2000 15:18:16 +0200
40. Email Duplication
"Joseph Peterson" <jay at qtm dot net>
Fri, 24 Mar 2000 10:07:47 -0500
41. HELP! qpopper process up to 500 and growing
Dan Scoggins <dan.scoggins at gsfc.nasa dot gov>
Fri, 24 Mar 2000 13:50:26 -0500
42. Re: HELP! qpopper process up to 500 and growing
"James Nelson" <james at digit.bloomnet dot com>
Fri, 24 Mar 2000 14:13:37 -0600
43. Re: Quotas
"James Nelson" <james at digit.bloomnet dot com>
Fri, 24 Mar 2000 14:06:14 -0600
44. Re: Quotas
Randall Gellens <randy at qualcomm dot com>
Sat, 25 Mar 2000 10:00:30 +1030
45. Re: Pam & Radius & Qpopper
"Franco Catena" <catena at surson.com dot br>
Sat, 25 Mar 2000 11:21:41 -0300
46. Re: Pam & Radius & Qpopper
Qpopper Support <qpopper at qualcomm dot com>
Sun, 26 Mar 2000 22:36:11 +0930
47. Re: Pam & Radius & Qpopper
Qpopper Support <qpopper at qualcomm dot com>
Sun, 26 Mar 2000 22:40:37 +0930
48. Compiling qpopper 3 on IRIX
"Nordh, Valter" <valter at che.chalmers dot se>
Mon, 27 Mar 2000 17:58:47 +0200
49. EOF or I/O error
Brent <bierblb at netins dot net>
Mon, 27 Mar 2000 17:57:01 -0600
50. Re: Compiling qpopper 3 on IRIX
Georg Schwarz <schwarz at physik.tu-berlin dot de>
Tue, 28 Mar 2000 08:30:41 +0200 (MEST)
Date: Fri, 17 Mar 2000 09:41:26 -0800
From: Randall Gellens <randy at qualcomm dot com>
Subject: Re: "Unable to open Bulletin database" errors
At 10:36 AM +0100 3/17/00, Miss Parker wrote:
>they look like:
>
>Mar 13 14:38:16 server popper[25308]: gdbm_open failed: Resource temporarily
>unavailable (11)
>Mar 13 14:38:19 server popper[25308]: user at cable15.icpnet.pl
>(212.126.17.15): -ERR [SYS/TEMP] Unable to
>open Bulletin database; contact your administrator
It looks like the gdbm_open failed because another user had the db
open. (Qpopper probably opens the bulletins db too early, which
means it is kept open too long, but this behavior has been the case
since the bulletins database was added.)
Earlier versions of Qpopper didn't retry if the gdbm_open failed, but
since December 1999 they do.
>
>
>----- Original Message -----
>From: Qpopper Support <qpopper at qualcomm dot com>
>To: Miss Parker <miss.parker at dialcom.com dot pl>; Subscribers of Qpopper
><qpopper at lists.pensive dot org>
>Cc: <qpopper at qualcomm dot com>
>Sent: 16 marca 2000 18:34
>Subject: Re: "Unable to open Bulletin database" errors
>
>
>> At 12:34 PM +0100 3/16/00, Miss Parker wrote:
>>
>> >-ERR [SYS/TEMP] Unable to open Bulletin database; contact your
>administrator
>> >
>> >and then all the rest of the useres, trying to check their mail,
>> >get the same error, and no one can get their mail
>>
>> There should be a more detailed error message in the log. What does it
>say?
>>
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
I loathe housework. You make beds, wash dishes
-- and six months later you have to start all over again.
--Joan Rivers
Subject: Same mail except "X-UIDL" header
From: Michio SHIZUTA <m-shizuta at fitech.co dot jp>
Date: Mon, 20 Mar 2000 17:25:35 +0900
Recently I changed SMTP/POP server to another PC because
the harddisk of previouse mechine was bad.
But after that client mailer downloads the same email twice.
Mailbox of server doesn't contain two mail.
But mailer downloads the same email except "X-UIDL" header.
Why?
Pleas let me know if anybody has good ideas.
Thanks in advance.
Date: Mon, 20 Mar 2000 08:52:04 -0500
From: Sherman Wong <Sherman_Wong at baruch.cuny dot edu>
Subject: QPopper and Upgrade to AIX 4.3.3
I recently upgraded from AIX 3.2.5 to AIX 4.3.3. Before the upgrade,
the version of QPopper, 2.4, worked fine and the users could retrieve
their mail using Netscape Mail. However, since the upgrade, the users
have been unable to retrieve their mail using the server. The server is
active, but when one attempts to use it, the error message is "system
error: Unable to open temporary file. Do you own it?".
I have examined the FAQs for QPopper and found the question related to
exactly this message. I have checked the permissions for
/var/spool/mail/ and have verified them; yet and still, the problem
persists. Moreover, I have experimented with the permissions for
/var/spool/mqueue in hopes that those were incorrect.
What other actions should I try?
Thank you.
Sherman Wong
Sherman_Wong at baruch.cuny dot edu
Date: Mon, 20 Mar 2000 10:00:09 -0800
From: Qpopper Support <qpopper at qualcomm dot com>
Subject: Re: QPopper and Upgrade to AIX 4.3.3
At 8:52 AM -0500 3/20/00, Sherman Wong wrote:
>I recently upgraded from AIX 3.2.5 to AIX 4.3.3. Before the upgrade,
>the version of QPopper, 2.4, worked fine and the users could retrieve
>their mail using Netscape Mail. However, since the upgrade, the users
>have been unable to retrieve their mail using the server. The server is
>active, but when one attempts to use it, the error message is "system
>error: Unable to open temporary file. Do you own it?".
>
>I have examined the FAQs for QPopper and found the question related to
>exactly this message. I have checked the permissions for
>/var/spool/mail/ and have verified them; yet and still, the problem
>persists. Moreover, I have experimented with the permissions for
>/var/spool/mqueue in hopes that those were incorrect.
>
>What other actions should I try?
>
>Thank you.
>
>Sherman Wong
>Sherman_Wong at baruch.cuny dot edu
You should not run versions of Qpopper older than 2.5 because of
possible security vulnerabilities. You could use 2.53, which is the
latest released version, but I'd recommend 3.0fc1.
Also, check if there are existing .user.pop files, and if so, if the
user owns them and what the access is. You might also want to look
inside them to see if they are corrupted. .user.pop files should not
exist after the session ends, unless there is a problem.
From: "Lisa Casey" <lisa at jellico dot com>
Subject:
Date: Mon, 20 Mar 2000 13:11:54 -0500
Hi,
I sent these questions to the qpopper list on 03/15/00 and never did get any
responses so I'll ask again... :)
Got a couple of questions. I'm running qpopper 2.53 and am having a (rare)
problem of a user gettig a SIGHUP or SIGPIPE error message and having their
.pop spool file hang around after disconnecting. According to the FAQ, it's
recommended to upgrade to qpopper 3. Version 2.5.3 is the current release
version, 3 is still in public beta. We are an ISP, our customers rely on us
for their E-mail and I hate to run beta software in a production
environment. Anyone got any thoughts on this?
If I upgrading, should I just install the new version over the old? Is
there any danger in losing stuff (configurations, customers E-mail, etc.)?
Also, is there any future plans to make the archives of this list at
http://www.pensive.org/Mailing_Lists/Archives/qpopper/index.html
searchable, or is there a searchable archive somewhere else?
Thanks,
Lisa Casey, Webmaster
Interstate 2000, Inc.
lisa at jellico dot com
webmaster at jellico dot com
Date: Mon, 20 Mar 2000 12:02:57 -0800
From: Qpopper Support <qpopper at qualcomm dot com>
Subject: Re: <no subject>
At 1:11 PM -0500 3/20/00, Lisa Casey wrote:
>Hi,
>
>I sent these questions to the qpopper list on 03/15/00 and never did get any
>responses so I'll ask again... :)
>
>
>Got a couple of questions. I'm running qpopper 2.53 and am having a (rare)
>problem of a user gettig a SIGHUP or SIGPIPE error message and having their
>.pop spool file hang around after disconnecting. According to the FAQ, it's
>recommended to upgrade to qpopper 3. Version 2.5.3 is the current release
>version, 3 is still in public beta. We are an ISP, our customers rely on us
>for their E-mail and I hate to run beta software in a production
>environment. Anyone got any thoughts on this?
3.0 is now at Final Candidate 1. It is much more stable than 2.53.
The error messages may be caused by poor modem connections.
>
>If I upgrading, should I just install the new version over the old? Is
>there any danger in losing stuff (configurations, customers E-mail, etc.)?
No, but it is a good idea to review the INSTALL file, and create a
shell script to run ./configure, to make sure you have the options
you need.
Date: Mon, 20 Mar 2000 19:18:44 -0600
From: Brent <bierblb at netins dot net>
Subject: envelope problem
This is the second time that this "Unable to process From lines
(envelopes), change recognition modes. " error has occured. The only
time I fixed it last time was to delete all the messages for that
user. The head of the mail file looks like this. Can anyone see a
problem or know another reason this might be happening? (the file is
6 meg--is there a max?)
"From jmdavis at snyder-associates dot com Thu Mar 9 18:08:40 2000
Return-Path: <blah blah blah"
Thanks for everyone's time!
Brent Bierstedt
Date: Mon, 20 Mar 2000 19:39:35 -0600
From: Brent <bierblb at netins dot net>
Subject: envelope problem
Sorry about this second message...there are not "'s in the file, I
just put that there.
Brent
Date: Tue, 21 Mar 2000 07:57:35 +0200
From: Raphael Maseko <ralph at zamnet dot zm>
Subject: Re: envelope problem
Hi Brent,
I have experienced that on two occasions with my users and, like you, the
only way I could free the account was to delete all messages.
I found a FAQ on Eudola Light somewhere which mentioned that such an
error is a result of a malformed header in an incoming mail. It did not
give details and I would be interested to get more detail from anyone on
this list. ( I am on FreeBSD2 with Qpopper 2.53
Ralph
Brent wrote:
> This is the second time that this "Unable to process From lines
> (envelopes), change recognition modes. " error has occured. The only
> time I fixed it last time was to delete all the messages for that
> user. The head of the mail file looks like this. Can anyone see a
> problem or know another reason this might be happening? (the file is
> 6 meg--is there a max?)
>
> "From jmdavis at snyder-associates dot com Thu Mar 9 18:08:40 2000
> Return-Path: <blah blah blah"
>
> Thanks for everyone's time!
>
> Brent Bierstedt
--
Raphael Maseko
Communications Engineer
Zamnet Communications systems Ltd
Date: Mon, 20 Mar 2000 22:32:57 -0800 (PST)
From: Rick Johnson <rsjohn at directcon dot net>
Subject: Re: envelope problem
In my experience. This is caused by the first line of the mail file not
containign a valid From line. I'll usually go in by hand and forge one or
just delete that one message.
The FAQ explains it further.
http://www.eudora.com/qpopper/faq.html#from.lines
Rick
--
===========================================================
Rick Johnson <rsjohn at directcon dot net>
Director of Operations Phone: 530.677.1712
Direct Connect Internet Fax: 530.676.2664
http://www.directcon.net AIM: rsjohn01
===========================================================
PGPKey: http://www.directcon.net/rsjohn/pgpkey
If you pick up a starving dog and make him prosperous, he will not
bite you; that is the principal difference between a dog and a man.
- Mark Twain
On Tue, 21 Mar 2000, Raphael Maseko wrote:
:Hi Brent,
:I have experienced that on two occasions with my users and, like you, the
:only way I could free the account was to delete all messages.
:I found a FAQ on Eudola Light somewhere which mentioned that such an
:error is a result of a malformed header in an incoming mail. It did not
:give details and I would be interested to get more detail from anyone on
:this list. ( I am on FreeBSD2 with Qpopper 2.53
:Ralph
:Brent wrote:
:
:> This is the second time that this "Unable to process From lines
:> (envelopes), change recognition modes. " error has occured. The only
:> time I fixed it last time was to delete all the messages for that
:> user. The head of the mail file looks like this. Can anyone see a
:> problem or know another reason this might be happening? (the file is
:> 6 meg--is there a max?)
:>
:> "From jmdavis at snyder-associates dot com Thu Mar 9 18:08:40 2000
:> Return-Path: <blah blah blah"
:>
:> Thanks for everyone's time!
:>
:> Brent Bierstedt
:
:--
:Raphael Maseko
:Communications Engineer
:Zamnet Communications systems Ltd
:
:
Date: Tue, 21 Mar 2000 01:04:22 -0600
From: Brent <bierblb at netins dot net>
Subject: Re: envelope problem
The only problem is that pine and the unix mail can read it fine, and
if I do delete that message, it still does that until I delete the
whole file. I am too using FreeBSD with b30.
>In my experience. This is caused by the first line of the mail file not
>containign a valid From line. I'll usually go in by hand and forge one or
>just delete that one message.
>
>The FAQ explains it further.
>
>http://www.eudora.com/qpopper/faq.html#from.lines
>
>Rick
>
>--
>===========================================================
>Rick Johnson <rsjohn at directcon dot net>
>Director of Operations Phone: 530.677.1712
>Direct Connect Internet Fax: 530.676.2664
>http://www.directcon.net AIM: rsjohn01
>===========================================================
> PGPKey: http://www.directcon.net/rsjohn/pgpkey
>
>If you pick up a starving dog and make him prosperous, he will not
>bite you; that is the principal difference between a dog and a man.
> - Mark Twain
>
>On Tue, 21 Mar 2000, Raphael Maseko wrote:
>
>:Hi Brent,
>:I have experienced that on two occasions with my users and, like you, the
>:only way I could free the account was to delete all messages.
>:I found a FAQ on Eudola Light somewhere which mentioned that such an
>:error is a result of a malformed header in an incoming mail. It did not
>:give details and I would be interested to get more detail from anyone on
>:this list. ( I am on FreeBSD2 with Qpopper 2.53
>:Ralph
>:Brent wrote:
>:
>:> This is the second time that this "Unable to process From lines
>:> (envelopes), change recognition modes. " error has occured. The only
>:> time I fixed it last time was to delete all the messages for that
>:> user. The head of the mail file looks like this. Can anyone see a
>:> problem or know another reason this might be happening? (the file is
>:> 6 meg--is there a max?)
>:>
>:> "From jmdavis at snyder-associates dot com Thu Mar 9 18:08:40 2000
>:> Return-Path: <blah blah blah"
>:>
>:> Thanks for everyone's time!
>:>
>:> Brent Bierstedt
>:
>:--
>:Raphael Maseko
>:Communications Engineer
>:Zamnet Communications systems Ltd
>:
>:
From: "James Nelson" <james at digit.bloomnet dot com>
Subject: Re: envelope problem
Date: Tue, 21 Mar 2000 01:20:15 -0600
The root of the problem seems to be poor locking compatibility between
qpopper and procmail. I have heard of several reports of mail box
corruption like this and have seen it personally as well. When it
happened to me I upgraded my procmail and the problem went away
entirely. I am also aware that 3.0 has a greatly improved lock
mechanism. My suggestion would be to upgrade procmail and also
upgrade qpopper. 2.53 is a rather aging software these days-- there
are SEVERAL known bugs and to my knowledge, nobody has plans to fix
them. I can't think of a single instance where it would be
beneficial to run 2.53 instead of 3.0b.
----- Original Message -----
From: "Raphael Maseko" <ralph at zamnet dot zm>
To: "Brent" <bierblb at netins dot net>
Cc: "Subscribers of Qpopper" <qpopper at lists.pensive dot org>
Sent: Monday, March 20, 2000 11:57 PM
Subject: Re: envelope problem
| Hi Brent,
| I have experienced that on two occasions with my users and, like
you, the
| only way I could free the account was to delete all messages.
| I found a FAQ on Eudola Light somewhere which mentioned that such an
| error is a result of a malformed header in an incoming mail. It did
not
| give details and I would be interested to get more detail from
anyone on
| this list. ( I am on FreeBSD2 with Qpopper 2.53
| Ralph
| Brent wrote:
|
| > This is the second time that this "Unable to process From lines
| > (envelopes), change recognition modes. " error has occured. The
only
| > time I fixed it last time was to delete all the messages for that
| > user. The head of the mail file looks like this. Can anyone see
a
| > problem or know another reason this might be happening? (the file
is
| > 6 meg--is there a max?)
| >
| > "From jmdavis at snyder-associates dot com Thu Mar 9 18:08:40 2000
| > Return-Path: <blah blah blah"
| >
| > Thanks for everyone's time!
| >
| > Brent Bierstedt
|
| --
| Raphael Maseko
| Communications Engineer
| Zamnet Communications systems Ltd
|
|
Date: Tue, 21 Mar 2000 03:12:41 -0600 (CST)
From: Jason Englander <jason at interl dot net>
Subject: APOP's popauth - w/o prompting for PW?
I apologize if this is a FAQ...
Anyone have anything that'll do the equivalent of popauth for adding an
APOP user, but without prompting to enter the password?
For example, instead of passwd, for certain things I use:
echo "user:pass" | chpasswd
...which comes with Shadow on my Linux boxen. I ask because I'm more or
less c illiterate. If anyone can recommend a perl module that'll work,
I'd appreciate that too.
Thanks,
Jason
--
Jason Englander <jason at interl dot net>
Systems Administrator - InterLink L.C.
From: "Utku Er" <erutku at netone.net dot tr>
Subject: Re: envelope problem
Date: Tue, 21 Mar 2000 14:12:26 +0200
It's not only with procmail. there is a section in the FAQ:
http://www.eudora.com/qpopper/qpop_faq.html#from.lines
Utku Er.
----- Original Message -----
From: James Nelson <james at digit.bloomnet dot com>
To: Raphael Maseko <ralph at zamnet dot zm>; Brent <bierblb at netins dot net>
Cc: Subscribers of Qpopper <qpopper at lists.pensive dot org>
Sent: Tuesday, March 21, 2000 9:20 AM
Subject: Re: envelope problem
> The root of the problem seems to be poor locking compatibility between
> qpopper and procmail. I have heard of several reports of mail box
> corruption like this and have seen it personally as well. When it
> happened to me I upgraded my procmail and the problem went away
> entirely. I am also aware that 3.0 has a greatly improved lock
> mechanism. My suggestion would be to upgrade procmail and also
> upgrade qpopper. 2.53 is a rather aging software these days-- there
> are SEVERAL known bugs and to my knowledge, nobody has plans to fix
> them. I can't think of a single instance where it would be
> beneficial to run 2.53 instead of 3.0b.
>
> ----- Original Message -----
> From: "Raphael Maseko" <ralph at zamnet dot zm>
> To: "Brent" <bierblb at netins dot net>
> Cc: "Subscribers of Qpopper" <qpopper at lists.pensive dot org>
> Sent: Monday, March 20, 2000 11:57 PM
> Subject: Re: envelope problem
>
>
> | Hi Brent,
> | I have experienced that on two occasions with my users and, like
> you, the
> | only way I could free the account was to delete all messages.
> | I found a FAQ on Eudola Light somewhere which mentioned that such an
> | error is a result of a malformed header in an incoming mail. It did
> not
> | give details and I would be interested to get more detail from
> anyone on
> | this list. ( I am on FreeBSD2 with Qpopper 2.53
> | Ralph
> | Brent wrote:
> |
> | > This is the second time that this "Unable to process From lines
> | > (envelopes), change recognition modes. " error has occured. The
> only
> | > time I fixed it last time was to delete all the messages for that
> | > user. The head of the mail file looks like this. Can anyone see
> a
> | > problem or know another reason this might be happening? (the file
> is
> | > 6 meg--is there a max?)
> | >
> | > "From jmdavis at snyder-associates dot com Thu Mar 9 18:08:40 2000
> | > Return-Path: <blah blah blah"
> | >
> | > Thanks for everyone's time!
> | >
> | > Brent Bierstedt
> |
> | --
> | Raphael Maseko
> | Communications Engineer
> | Zamnet Communications systems Ltd
> |
> |
Subject: qpopper3.0 : hashed mailspool implies hashed drop directory
Date: Tue, 21 Mar 2000 16:02:15 +0100 (MET)
From: Eric Luyten <eluyten at vub.ac dot be>
Last night, when testing qpopper3.0b34 with hash spool method 2 enabled,
I discovered it needs a similarly hashed POP_DROP_DIR.
For performance reasons we redefine POP_DROP_DIR to /var/mail/poptemp
and allocate that to a completely separate device.
I had to rewrite four or five lines of code in common/genpath.c to have
Qpopper work with a hashed mailspool but a 'flat' temporary directory
for all transient/work files. I prefer the contents of that directory
to be completely visible using a single 'ls -la' command to check for
leftovers, forgotten locks, etc...
If your POP_DROP_DIR equals POP_MAILDIR (see the config.h created by
./configure) you will not run into this "problem".
Eric Luyten, Computing Centre VUB/ULB.
From: "Jerry O'Brien" <jobrien at cuttingedge dot net>
Subject: IP reported as localhost
Date: Tue, 21 Mar 2000 10:29:42 -0800
Hi all,
I just noticed that a few users' IP addresses are being logged as
"localhost" like this:
popper[25168]: Stats: <user> 0 0 6 18913 localhost 127.0.0.1
How can this happen? Can I force it to report actual dial-up IP instead?
Thanks,
Jerry O'Brien
Cutting Edge Systems
Date: Tue, 21 Mar 2000 11:36:13 -0500
From: "Michael D. Sofka" <sofkam at rpi dot edu>
Subject: Re: qpopper3.0 : hashed mailspool implies hashed drop directory
At 04:02 PM 3/21/00 +0100, Eric Luyten wrote:
>Last night, when testing qpopper3.0b34 with hash spool method 2 enabled,
>I discovered it needs a similarly hashed POP_DROP_DIR.
>
>For performance reasons we redefine POP_DROP_DIR to /var/mail/poptemp
>and allocate that to a completely separate device.
>
>I had to rewrite four or five lines of code in common/genpath.c to have
>Qpopper work with a hashed mailspool but a 'flat' temporary directory
>for all transient/work files. I prefer the contents of that directory
>to be completely visible using a single 'ls -la' command to check for
>leftovers, forgotten locks, etc...
>
>If your POP_DROP_DIR equals POP_MAILDIR (see the config.h created by
>./configure) you will not run into this "problem".
Interesting. With qpopper 2.53 I recently switched from a dedicated poptemp
device to put the lock and temp files in the mail spool. We had a separate
partition (disk, actually) and discovered it was a bottleneck. Here is what
happened:
We run in server mode. So, if all of the messages are deleted, or there was
no change to the mailbox, the lock and pop check were quick. But, if a
mailbox had any new messages, or some of the messages were kept on the
server, the mailbox would be copied from the spool to the lock directory,
and then back. When a couple of large mailboxes (say 50 or 100 meg)
were copied at the same time, this would delay the creation of pop locks
on the temporary device, driving up the total number of popper processes
running and eventually grinding the machine to a crawl until processes
started to clear. This would happen about once a week on average. (ps
would typically show about 350 popper processes as the machine was
recovering, and I was able to get a prompt. We have a gig of swap space
just to avoid running out of VM during these slowdows---under normal load,
about 10% of swap would be in use.)
The poplock partition was on a single SSA disk. The spool, however,
is on 6 mirrored SSA disks, with a 7th dedicated filesystem log disk. My
hope is that putting the lock and temp files on the spool will be less
likely to show problems. Since last week was spring break, this is the
first week of testing under full load. (I'll post a follow up when I have some
data, but yesterday went well.)
Some additional information. We aggressively preen mailboxes. Any read
messages over 21 days old are removed once a week. Despite that, some
mailboxes grow to larger sizes. There also seem to be clusters of departments
configured to keep mail on the server for a couple days---not unreasonable.
When those clusters share large documents (such as a word file), this
triggers a slowdown, since a number of large mailboxes now have to be
copied. This is more common just before lunch, for some reason.
BTW, finding locks and temp file is easy:
find /var/spool/mail \( -name ".[a-z]*.pop" -o -name "xmit*" \) -ls
does it. A couple of other scripts had to be rewritten.
I look at this as a temporary fix (and if it doesn't work, I have
some other disks to play with so that the pop pock file can be
interleaved with a dedicated log disk; but, that would take
some re-organization of disk space). The root problem is that large
objects are difficult to handle (anybody editing a large file on
their personal PC would recognize this) We have lots of storage, but as email
use changes, the size of mailboxes will increase. In addition,
more users will read email from multiple locations, which means keeping the mail on
the server for several days (since last year the average mailbox size
has doubled). So, one of the projects for the summer is
to look into hashed or indexed mail message schemes, such as those
used by WU-IMAP.
Mike
--
Michael D. Sofka sofkam at rpi dot edu
CIS/SSS Sr. Systems Programmer AFS/DFS, email, listproc, TeX, epistemology.
Rensselaer Polytechnic Institute, Troy, NY. http://www.rpi.edu/~sofkam/
Date: Tue, 21 Mar 2000 17:49:04 +0100 (MET)
From: Gerhard Pfeiffer <gp at sun.bnbt dot de>
Subject: Re: IP reported as localhost
Hallo,
On Tue, 21 Mar 2000, Jerry O'Brien wrote:
> I just noticed that a few users' IP addresses are being logged as
> "localhost" like this:
> popper[25168]: Stats: <user> 0 0 6 18913 localhost 127.0.0.1
> How can this happen? Can I force it to report actual dial-up IP instead?
Perhaps they are actually connecting from localhost?
A possible cause for this is, that they are using ssh's port-forwarding.
I don't think, that you can force qpopper to report the dialup-ip in this
cause.
Ciao,
Gerhard
From: Steven Fletcher <stevenf at shellnet.co dot uk>
Subject: RE: IP reported as localhost
Date: Tue, 21 Mar 2000 16:58:59 -0000
It shouldn't, it's only reporting what your OS is telling it to.
Are your users using, for example, a CGI based reader running on the local
machine, or are they using SSH to encrypt their sessions? That would explain
it.
Steven Fletcher
stevenf at shellnet.co dot uk
> -----Original Message-----
> From: Jerry O'Brien [mailto:jobrien at cuttingedge dot net]
> Sent: 21 March 2000 18:30
> To: Subscribers of Qpopper
> Subject: IP reported as localhost
>
>
> Hi all,
>
> I just noticed that a few users' IP addresses are being logged as
> "localhost" like this:
>
> popper[25168]: Stats: <user> 0 0 6 18913 localhost 127.0.0.1
>
> How can this happen? Can I force it to report actual dial-up
> IP instead?
>
> Thanks,
>
>
> Jerry O'Brien
> Cutting Edge Systems
>
>
>
From: Patrick Nixon <pnixon at somerville.ma dot us>
Subject: RE: IP reported as localhost
Date: Tue, 21 Mar 2000 12:43:27 -0500
If it's ssh based, you'll see log messages for the SSH traffic mapping the local port
-----Original Message-----
From: Gerhard Pfeiffer [SMTP:gp at sun.bnbt dot de]
Sent: Tuesday, March 21, 2000 11:49 AM
To: Jerry O'Brien
Cc: Subscribers of Qpopper
Subject: Re: IP reported as localhost
Hallo,
On Tue, 21 Mar 2000, Jerry O'Brien wrote:
> I just noticed that a few users' IP addresses are being logged as
> "localhost" like this:
> popper[25168]: Stats: <user> 0 0 6 18913 localhost 127.0.0.1
> How can this happen? Can I force it to report actual dial-up IP instead?
Perhaps they are actually connecting from localhost?
A possible cause for this is, that they are using ssh's port-forwarding.
I don't think, that you can force qpopper to report the dialup-ip in this
cause.
Ciao,
Gerhard
Date: Tue, 21 Mar 2000 20:17:38 +0200
From: Raphael Maseko <ralph at zamnet dot zm>
Subject: Re: IP reported as localhost
Do any of your users access their mail through a web interface installed on
the same machine as the POP3 server?
I have noticed that if you use an web interface ( eg mailman) it will show
that the user is "popping" mail from the address of the machine where the
software is installed.
-Ralph
Jerry O'Brien wrote:
> Hi all,
>
> I just noticed that a few users' IP addresses are being logged as
> "localhost" like this:
>
> popper[25168]: Stats: <user> 0 0 6 18913 localhost 127.0.0.1
>
> How can this happen? Can I force it to report actual dial-up IP instead?
>
> Thanks,
>
> Jerry O'Brien
> Cutting Edge Systems
--
Raphael Maseko
Communications Engineer
Zamnet Communications systems Ltd
From: "Jerry O'Brien" <jobrien at cuttingedge dot net>
Subject: Re: IP reported as localhost
Date: Tue, 21 Mar 2000 12:45:03 -0800
It turns out they are coming from a web interface (acmemail) that I had
overlooked. Thanks for all your help.
Jerry
----- Original Message -----
From: Raphael Maseko <ralph at zamnet dot zm>
To: Jerry O'Brien <jobrien at cuttingedge dot net>
Cc: Subscribers of Qpopper <qpopper at lists.pensive dot org>
Sent: Tuesday, March 21, 2000 10:17 AM
Subject: Re: IP reported as localhost
> Do any of your users access their mail through a web interface installed
on
> the same machine as the POP3 server?
> I have noticed that if you use an web interface ( eg mailman) it will
show
> that the user is "popping" mail from the address of the machine where the
> software is installed.
> -Ralph
>
> Jerry O'Brien wrote:
>
> > Hi all,
> >
> > I just noticed that a few users' IP addresses are being logged as
> > "localhost" like this:
> >
> > popper[25168]: Stats: <user> 0 0 6 18913 localhost 127.0.0.1
> >
> > How can this happen? Can I force it to report actual dial-up IP instead?
> >
> > Thanks,
> >
> > Jerry O'Brien
> > Cutting Edge Systems
>
> --
> Raphael Maseko
> Communications Engineer
> Zamnet Communications systems Ltd
>
>
>
From: "Fabio Oliveira" <fabio.oliveira at nttb.com dot br>
Subject: Qpopper - LDAP
Date: Tue, 21 Mar 2000 16:02:07 -0300
Hi All,
I would like to know about integration of LDAP (Directory Server) w/ QPOP
and Sendmail. I dont like that use dbm file on my system.
How can I configure it?
Thanks for any comments and suggest.
Regards,
-------------------------------------------------
Fabio de Oliveira
Network System Engineer
------------------------------------------------
Date: Wed, 22 Mar 2000 17:58:23 +1030
From: Randall Gellens <randy at qualcomm dot com>
Subject: Re: envelope problem
At 1:20 AM -0600 3/21/00, James Nelson wrote:
> When it
>happened to me I upgraded my procmail and the problem went away
>entirely. I am also aware that 3.0 has a greatly improved lock
>mechanism. My suggestion would be to upgrade procmail and also
>upgrade qpopper. 2.53 is a rather aging software these days-- there
>are SEVERAL known bugs and to my knowledge, nobody has plans to fix
>them. I can't think of a single instance where it would be
>beneficial to run 2.53 instead of 3.0b.
I agree.
Also, check if the user has a .user.pop file, and if so, if it
appears corrupted. If so, try deleting this file instead of the
spool. See the FAQ for more details, as several people already
posted.
(Also, 3.0 is now fc instead of b).
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
If I knew what was going on, I'd be indignant!
Date: Wed, 22 Mar 2000 14:22:00 +0200
From: Alexander Rusinov <boot at EuroCom.Od dot UA>
Subject: Read timeout & maillock
Hi,
I'm running qpopper on Linux (Kernel 2.2.13, Slackware 7.0). Very
often the mailboxes stay locked for about 20 minutes after abnormal
disconnection of dial users. I've tried to vary read timeout with -T
option (from 30 to 300 seconds), but that didn't help.
I've used almost all betas of qpopper 3.0, now I'm trying 3.0fc1 and
still occur the problem.
While making qpopper I've used the following config:
./configure --enable-servermode --enable-specialauth
In inetd.conf I have:
pop3 stream tcp nowait.100 root /usr/sbin/tcpd /usr/local/sbin/popper
-T 60 -sc
I've also tried -T60 (without whitespace).
What am I doing wrong?
--
Alexander Rusinov
Date: Thu, 23 Mar 2000 10:55:27 +0000 (GMT)
From: P O O P I E <poopie at jeeves.poopie dot net>
Subject: Quotas
I know this isn't a popper problem, but I figure some people
run popper with sendmail and might know...
I want to use quotas so that people can't send large
attachements. How do I do this?
Is this a sendmail configuration? or is it based on the OS?
For instance, if it's a FreeBSD machine, would I have to enable
quotas on /var?
thanx for the help
Date: Thu, 23 Mar 2000 17:38:31 +0200 (SAST)
From: Mihigo John <nmj at gorilla.nur.ac dot rw>
Subject: Re: Quotas
Quotas are based in the OS.
You can try man quotaon and make values in /usr/etc/quotaon.
In FreeBSD it should be very easy, because quotas are setted by default
way.
Some one can add more explaination.
Thanks
*****************************
Nzabonimana Mihigo Jean
Universite Nationale du Rwanda
Centre Informatique
Administrateur Reseau Internet
Phone: 250 530666 (bureau)
250 530074 (habitation)
08511917 (mobile)
Fax : 250 530105 / 530210
URL: www.nur.ac.rw
BP: 56 Butare, Rwanda
On Thu, 23 Mar 2000, P O O P I E wrote:
> I know this isn't a popper problem, but I figure some people
> run popper with sendmail and might know...
>
> I want to use quotas so that people can't send large
> attachements. How do I do this?
> Is this a sendmail configuration? or is it based on the OS?
> For instance, if it's a FreeBSD machine, would I have to enable
> quotas on /var?
>
> thanx for the help
>
>
>
Date: Thu, 23 Mar 2000 09:18:26 -0800
From: Joel Laing <joel at scripps dot edu>
Subject: Re: Quotas
A line like this in /etc/sendmail.cf on the relay machine will limit
relayed messages to 7 megs.
# maximum message size
O MaxMessageSize=7168000
-Joel
P O O P I E wrote:
>
> I know this isn't a popper problem, but I figure some people
> run popper with sendmail and might know...
>
> I want to use quotas so that people can't send large
> attachements. How do I do this?
> Is this a sendmail configuration? or is it based on the OS?
> For instance, if it's a FreeBSD machine, would I have to enable
> quotas on /var?
>
> thanx for the help
From: roberto.bazan at inycom dot es
Subject: RE: Quotas
Date: Thu, 23 Mar 2000 18:29:33 +0100
And if the message is for example 8 Megs. What does sendmail do with this
message?
Thanks
> A line like this in /etc/sendmail.cf on the relay machine will limit
> relayed messages to 7 megs.
>
> # maximum message size
> O MaxMessageSize=7168000
>
> -Joel
>
>
>
> P O O P I E wrote:
>
Date: Thu, 23 Mar 2000 12:34:29 -0500
From: Forrest Aldrich <forrie at forrie dot com>
Subject: RE: Quotas
Sendmail will reject it (bounce back) to the sender, with an error message
indicating that it's too big.
_F
At 06:29 PM 3/23/00 +0100, roberto.bazan at inycom dot es wrote:
>And if the message is for example 8 Megs. What does sendmail do with this
>message?
>Thanks
>
>
> > A line like this in /etc/sendmail.cf on the relay machine will limit
> > relayed messages to 7 megs.
> >
> > # maximum message size
> > O MaxMessageSize=7168000
> >
> > -Joel
> >
> >
> >
> > P O O P I E wrote:
> >
Date: Thu, 23 Mar 2000 18:38:13 +0000
From: Wallace Nicoll <wallace at cityedin.demon.co dot uk>
Subject: Re: Quotas
Forrest Aldrich wrote:
> Sendmail will reject it (bounce back) to the sender, with an error message
> indicating that it's too big.
And when it's too big for that user's quota, or there's insufficient disk
space ...
Wallace.
(having experienced this situation for several days ...)
--
======================================================================
Wallace Nicoll wallace at cityedin.demon.co dot uk
(Internet Services Team)
City of Edinburgh Council IT Services,
Chesser House, 500 Gorgie Road, Phone : 0131 469 5343
Edinburgh, EH11 3YJ, Scotland Fax : 0131 469 5335
[From overseas [P]+441314695343 [F]+441314695335 ]
======================================================================
Date: Thu, 23 Mar 2000 13:58:47 -0500
From: Forrest Aldrich <forrie at forrie dot com>
Subject: Re: Quotas
Quota... for Cyrus or System Quota, or ?
If sendmail accepts the message, it will act accordingly... what happens
next is reliant upon the application that controls the underlying
quota. For example, on some OS's, you'll not be able to write to the
drive... in that case, I've seen before messages bounced back to me that
says "user is over quota" or some ambiguous message like that.
What are you seeing ?
_F
At 06:38 PM 3/23/00 +0000, Wallace Nicoll wrote:
>Forrest Aldrich wrote:
>
> > Sendmail will reject it (bounce back) to the sender, with an error message
> > indicating that it's too big.
>
>And when it's too big for that user's quota, or there's insufficient disk
>space ...
>
>Wallace.
>
>(having experienced this situation for several days ...)
>
>--
>======================================================================
> Wallace Nicoll wallace at cityedin.demon.co dot uk
> (Internet Services Team)
> City of Edinburgh Council IT Services,
> Chesser House, 500 Gorgie Road, Phone : 0131 469 5343
> Edinburgh, EH11 3YJ, Scotland Fax : 0131 469 5335
> [From overseas [P]+441314695343 [F]+441314695335 ]
>======================================================================
>
From: "Kenneth Porter" <shiva at well dot com>
Date: Thu, 23 Mar 2000 16:56:37 -0800
Subject: RE: Quotas
On Thu, 23 Mar 2000 18:29:33 +0100, roberto.bazan at inycom dot es wrote:
>And if the message is for example 8 Megs. What does sendmail do with this
>message?
If the message has no header indicating the size, I believe it
truncates the message once the size limit is exceeded. If sendmail
knows up front how big the message will be, it can refuse the message
with an appropriate error response.
>> A line like this in /etc/sendmail.cf on the relay machine will limit
>> relayed messages to 7 megs.
>>
>> # maximum message size
>> O MaxMessageSize=7168000
Most people should use the m4 variable (see confMAX_MESSAGE_SIZE in
http://www.sendmail.org/m4/tweakingoptions.html), and not directly edit
the sendmail.cf file. The m4 file is much easier to upgrade from one
version of sendmail to the next than is the cf file.
The best place to get answers to sendmail questions is the
comp.mail.sendmail newsgroup.
Ken
mailto:shiva at well dot com
http://www.sewingwitch.com/ken/
Date: Thu, 23 Mar 2000 18:53:08 -0800
From: Brian Greene <brian at digilect dot com>
Subject: Re: Quotas
You have an excellent point! the MAXMESSAGESIZE directive in the sendmail.cf
file can be a security risk.
Kenneth Porter wrote:
> On Thu, 23 Mar 2000 18:29:33 +0100, roberto.bazan at inycom dot es wrote:
>
> >And if the message is for example 8 Megs. What does sendmail do with this
> >message?
>
> If the message has no header indicating the size, I believe it
> truncates the message once the size limit is exceeded. If sendmail
> knows up front how big the message will be, it can refuse the message
> with an appropriate error response.
>
> >> A line like this in /etc/sendmail.cf on the relay machine will limit
> >> relayed messages to 7 megs.
> >>
> >> # maximum message size
> >> O MaxMessageSize=7168000
>
> Most people should use the m4 variable (see confMAX_MESSAGE_SIZE in
> http://www.sendmail.org/m4/tweakingoptions.html), and not directly edit
> the sendmail.cf file. The m4 file is much easier to upgrade from one
> version of sendmail to the next than is the cf file.
>
> The best place to get answers to sendmail questions is the
> comp.mail.sendmail newsgroup.
>
> Ken
> mailto:shiva at well dot com
> http://www.sewingwitch.com/ken/
Date: Fri, 24 Mar 2000 17:39:57 +1030
From: Randall Gellens <randy at qualcomm dot com>
Subject: RE: Quotas
At 4:56 PM -0800 3/23/00, Kenneth Porter wrote:
>If the message has no header indicating the size, I believe it
>truncates the message once the size limit is exceeded. If sendmail
>knows up front how big the message will be, it can refuse the message
>with an appropriate error response.
This doesn't sound right to me. A server should never truncate an
SMTP message; it should be accepted or rejected.
There is a SIZE extension to SMTP which is supported by some clients
and some servers. It lets a server announce the maximum size of a
message it will ever accept, and lets a client indicate the size of
the message before the message is sent, so the server can reject it
right away (with a permanent or temporary error).
At 6:38 PM +0000 3/23/00, Wallace Nicoll wrote:
> > Sendmail will reject it (bounce back) to the sender, with an error message
>> indicating that it's too big.
>
>And when it's too big for that user's quota, or there's insufficient disk
>space ...
If a bounce is itself undeliverable, it is supposed to be either
discarded or sent to the postmaster. Otherwise there would be mail
loops. Bounces are only to be sent to the envelope MAIL FROM, and
must themselves be sent with a null MAIL FROM, so they cannot bounce.
If you are seeing loops in this situation, something is very broken.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
Q: Why do mountain climbers rope themselves together?
A: To prevent the sensible ones from going home.
Date: Fri, 24 Mar 2000 19:36:19 +1200 (NZST)
From: Alan Brown <alan at manawatu.gen dot nz>
Subject: RE: Quotas
On Thu, 23 Mar 2000 roberto.bazan at inycom dot es wrote:
> And if the message is for example 8 Megs. What does sendmail do with this
> message?
If it's ESMTP and the sending host uses the SIZE parameter, the
receiving host will reject on the spot as too big.
If the SIZE parameter isn't used, the receiving host will bounce it
_after_ receiving the entire message. :-(
AB
From: "Kenneth Porter" <shiva at well dot com>
Date: Thu, 23 Mar 2000 23:51:35 -0800
Subject: RE: Quotas
On Fri, 24 Mar 2000 17:39:57 +1030, Randall Gellens wrote:
>>If the message has no header indicating the size, I believe it
>>truncates the message once the size limit is exceeded. If sendmail
>>knows up front how big the message will be, it can refuse the message
>>with an appropriate error response.
>
>This doesn't sound right to me. A server should never truncate an
>SMTP message; it should be accepted or rejected.
See the message loop in sendmail/collect.c. When reading a message
body, if the size is exceeded, all remaining characters are discarded.
(I'm compiling a new source RPM of 8.10.0 right now.)
>There is a SIZE extension to SMTP which is supported by some clients
>and some servers. It lets a server announce the maximum size of a
>message it will ever accept, and lets a client indicate the size of
>the message before the message is sent, so the server can reject it
>right away (with a permanent or temporary error).
Ok, I couldn't remember what the actual mechanism was. I just
remembered that the handling was different based on whether sendmail
knew the message was too big up front or not.
Ken
mailto:shiva at well dot com
http://www.sewingwitch.com/ken/
Date: Fri, 24 Mar 2000 20:16:55 +1200 (NZST)
From: Alan Brown <alan at manawatu.gen dot nz>
Subject: Re: Quotas
On Thu, 23 Mar 2000, Wallace Nicoll wrote:
> Forrest Aldrich wrote:
>
> > Sendmail will reject it (bounce back) to the sender, with an error message
> > indicating that it's too big.
>
> And when it's too big for that user's quota, or there's insufficient disk
> space ...
Then it double bounces to the postmaster.
Life's a bitch and then you die.
AB
Date: Fri, 24 Mar 2000 15:18:16 +0200
From: Kostas Zourkas <kzou at space dot gr>
Subject: Pam & Radius & Qpopper
--------------4BD0D2E70B560421221B8490
Content-Type: text/plain; charset=iso-8859-7
Content-Transfer-Encoding: 7bit
Hi all,
i have installed the IcRadius server with mysql support and qpopper with
pam support. Now even though the radius server authenticates correctly
the request by the qpopper server (checked by a sniffer on the net) the
server rejects the request made by the client by saying that there is a
problem in the pam module, which i am sure that is working fine since it
is authenticating other services.
Any ideas commments will be appreciated.
Thank you for your help
--
=========================
======================
Kostas Zourkas
Network Administrator
Space Net S.A.
Mesogion 302
Holargos
Tel: +301 6547400
http://www.space.gr
=====================
========================
--------------4BD0D2E70B560421221B8490
Content-Type: text/html; charset=iso-8859-7
Content-Transfer-Encoding: 7bit
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Hi all,
<p>i have installed the IcRadius server with mysql support and qpopper
with pam support. Now even though the radius server authenticates correctly
the request by the qpopper server (checked by a sniffer on the net)
the server rejects the request made by the client by saying that there
is a problem in the pam module, which i am sure that is working fine since
it is authenticating other services.
<p>Any ideas commments will be appreciated.
<pre>Thank you for your help</pre>
<pre>--
=========================
======================
Kostas Zourkas
Network Administrator
Space Net S.A.
Mesogion 302
Holargos
Tel: +301 6547400
<A HREF="http://www.space.gr">http://www.space.gr</A>
=====================
========================</pre>
</html>
--------------4BD0D2E70B560421221B8490--
From: "Joseph Peterson" <jay at qtm dot net>
Subject: Email Duplication
Date: Fri, 24 Mar 2000 10:07:47 -0500
We use FreeBSD 3.3-STABLE and QPOP (version 3.0b36)
The only time we're running into this mailbox duplication problem is when
people retrieve their email with our web-based mail program called 'mailman'
(by Endymion Corporation, www.endymion.com) Any input would be greatly
appreciated! Thanks =)
Date: Fri, 24 Mar 2000 13:50:26 -0500
From: Dan Scoggins <dan.scoggins at gsfc.nasa dot gov>
Subject: HELP! qpopper process up to 500 and growing
I have a problem which I can not get ahead of enough to figure out what in
the heck is going on.
We are running solaris 2.6/qpopper 2.53 on a sparc 20 hypersparc with
approx 700 user mailboxes most of which use Eurdora. The qpopper is
spawning process upward to 500 in less then an hour. Most of which are
orphaned and do not appear to go away. As I understand, from this list,
those processes are orphaned because the Eurdora client connection timesout.
I can not seem to get the system stable. Any suggestions are welcome!
Thanx,
DanS
--
From: "James Nelson" <james at digit.bloomnet dot com>
Subject: Re: HELP! qpopper process up to 500 and growing
Date: Fri, 24 Mar 2000 14:13:37 -0600
Use the latest release candidate for 3.0.
Also, tweak your inetd config for some more reasonable timeout values
and allowed session requests per second. (The original RFC for popper
is not very reasonable, and the default values for inetd are not the
greatest either. The inetd values worked for our ISP until we got up
to the 700 concurrent user mark)
If you are not behind a good firewall (IE something like a Sidewinder
that insists on a 3 way handshake) TCP wrapper the thing with some
paranoid wrappers to rule out the possibility that someone is hitting
you with a DOS attack.
----- Original Message -----
From: "Dan Scoggins" <dan.scoggins at gsfc.nasa dot gov>
To: "Subscribers of Qpopper" <qpopper at lists.pensive dot org>
Sent: Friday, March 24, 2000 12:50 PM
Subject: HELP! qpopper process up to 500 and growing
| I have a problem which I can not get ahead of enough to figure out
what in
| the heck is going on.
|
| We are running solaris 2.6/qpopper 2.53 on a sparc 20 hypersparc
with
| approx 700 user mailboxes most of which use Eurdora. The qpopper is
| spawning process upward to 500 in less then an hour. Most of which
are
| orphaned and do not appear to go away. As I understand, from this
list,
| those processes are orphaned because the Eurdora client connection
timesout.
|
| I can not seem to get the system stable. Any suggestions are
welcome!
|
| Thanx,
|
| DanS
| --
|
From: "James Nelson" <james at digit.bloomnet dot com>
Subject: Re: Quotas
Date: Fri, 24 Mar 2000 14:06:14 -0600
The size declaration is a EHLO function which can be used to gain some
rather sensitive info about your SMTP server.
EHLO in combination with ETRN support is one to really watch out for.
----- Original Message -----
From: "Randall Gellens" <randy at qualcomm dot com>
To: "Kenneth Porter" <shiva at well dot com>; "Wallace Nicoll"
<wallace at cityedin.demon.co dot uk>; "Subscribers of Qpopper"
<qpopper at lists.pensive dot org>
Sent: Friday, March 24, 2000 1:09 AM
Subject: RE: Quotas
| At 4:56 PM -0800 3/23/00, Kenneth Porter wrote:
|
| >If the message has no header indicating the size, I believe it
| >truncates the message once the size limit is exceeded. If sendmail
| >knows up front how big the message will be, it can refuse the
message
| >with an appropriate error response.
|
| This doesn't sound right to me. A server should never truncate an
| SMTP message; it should be accepted or rejected.
|
| There is a SIZE extension to SMTP which is supported by some clients
| and some servers. It lets a server announce the maximum size of a
| message it will ever accept, and lets a client indicate the size of
| the message before the message is sent, so the server can reject it
| right away (with a permanent or temporary error).
|
| At 6:38 PM +0000 3/23/00, Wallace Nicoll wrote:
|
| > > Sendmail will reject it (bounce back) to the sender, with an
error message
| >> indicating that it's too big.
| >
| >And when it's too big for that user's quota, or there's
insufficient disk
| >space ...
|
| If a bounce is itself undeliverable, it is supposed to be either
| discarded or sent to the postmaster. Otherwise there would be mail
| loops. Bounces are only to be sent to the envelope MAIL FROM, and
| must themselves be sent with a null MAIL FROM, so they cannot
bounce.
|
| If you are seeing loops in this situation, something is very broken.
| --
| Randall Gellens
| Opinions are personal; facts are suspect; I speak for myself
only
| -------------- Randomly-selected tag: ---------------
| Q: Why do mountain climbers rope themselves together?
| A: To prevent the sensible ones from going home.
|
Date: Sat, 25 Mar 2000 10:00:30 +1030
From: Randall Gellens <randy at qualcomm dot com>
Subject: Re: Quotas
At 2:06 PM -0600 3/24/00, James Nelson wrote:
>The size declaration is a EHLO function which can be used to gain some
>rather sensitive info about your SMTP server.
>
>EHLO in combination with ETRN support is one to really watch out for.
Can you provide more detail? I don't see the problem. What is the
harm in advertising a maximum message size? And ETRN can't be used
without EHLO, so I really don't follow your point here.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly-selected tag: ---------------
Every solution breeds new problems.
From: "Franco Catena" <catena at surson.com dot br>
Subject: Re: Pam & Radius & Qpopper
Date: Sat, 25 Mar 2000 11:21:41 -0300
This is a multi-part message in MIME format.
------=_NextPart_000_00E9_01BF964C.4B241500
Content-Type: text/plain;
charset="iso-8859-7"
Content-Transfer-Encoding: quoted-printable
Hi fellows
where to find qpopper with PAM support?????
thanks
-----Mensagem Original-----
De: Kostas Zourkas
Para: Subscribers of Qpopper
Enviada em: Sexta-feira, 24 de Março de 2000 10:18
Assunto: Pam & Radius & Qpopper
Hi all,
i have installed the IcRadius server with mysql support and qpopper =
with pam support. Now even though the radius server authenticates =
correctly the request by the qpopper server (checked by a sniffer on =
the net) the server rejects the request made by the client by saying =
that there is a problem in the pam module, which i am sure that is =
working fine since it is authenticating other services.
Any ideas commments will be appreciated.
Thank you for your help
--
==========================
======================
Kostas Zourkas
Network Administrator
Space Net S.A.
Mesogion 302
Holargos
Tel: +301 6547400
http://www.space.gr
=====================
========================
------=_NextPart_000_00E9_01BF964C.4B241500
Content-Type: text/html;
charset="iso-8859-7"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content="text/html; charset=iso-8859-7" =
http-equiv=Content-Type>
<META content="MSHTML 5.00.2314.1000" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT size=2>Hi fellows</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>where to find qpopper with PAM =
support?????</FONT></DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV><FONT size=2>thanks</FONT></DIV>
<DIV> </DIV>
<BLOCKQUOTE
style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: =
0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px">
<DIV style="FONT: 10pt arial">-----Mensagem Original----- </DIV>
<DIV
style="BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: =
black"><B>De:</B> <A
href="mailto:kzou@space dot gr" title=kzou at space dot gr>Kostas Zourkas</A> =
</DIV>
<DIV style="FONT: 10pt arial"><B>Para:</B> <A
href="mailto:qpopper at lists dot pensive dot org"
title=qpopper at lists.pensive dot org>Subscribers of Qpopper</A> </DIV>
<DIV style="FONT: 10pt arial"><B>Enviada em:</B> Sexta-feira, 24 de =
Março de
2000 10:18</DIV>
<DIV style="FONT: 10pt arial"><B>Assunto:</B> Pam & Radius & =
Qpopper</DIV>
<DIV><BR></DIV>Hi all,
<P>i have installed the IcRadius server with mysql support and qpopper =
with
pam support. Now even though the radius server authenticates correctly =
the
request by the qpopper server (checked by a sniffer on the net) =
the
server rejects the request made by the client by saying that there is =
a
problem in the pam module, which i am sure that is working fine since =
it is
authenticating other services.
<P>Any ideas commments will be appreciated. <PRE>Thank you for your =
help</PRE><PRE>--
==========================
======================
Kostas Zourkas
Network Administrator
Space Net S.A.
Mesogion 302
Holargos
Tel: +301 6547400
<A href="http://www.space.gr">http://www.space.gr</A>
=====================
========================<=
/PRE> </BLOCKQUOTE></BODY></HTML>
------=_NextPart_000_00E9_01BF964C.4B241500--
Date: Sun, 26 Mar 2000 22:36:11 +0930
From: Qpopper Support <qpopper at qualcomm dot com>
Subject: Re: Pam & Radius & Qpopper
At 11:21 AM -0300 3/25/00, Franco Catena wrote:
>where to find qpopper with PAM support?????
<http://www.qpopper.org>. It's in Qpopper 3.0.
Date: Sun, 26 Mar 2000 22:40:37 +0930
From: Qpopper Support <qpopper at qualcomm dot com>
Subject: Re: Pam & Radius & Qpopper
At 3:18 PM +0200 3/24/00, Kostas Zourkas wrote:
>the server rejects the request made by the client by saying that
>there is a problem in the pam module, which i am sure that is
>working fine since it is authenticating other services.
Try enabling debug tracing (do a 'make clean', run ./configure adding
'--enable-debugging', run make, install, add '-t tracefile' to the
inetd.conf line). It should indicate the exact error returned by the
pam module.
Also, make sure you have an entry for the pop3 service (or whatever
service you specified when you ran ./configure) in the pam.d
directory.
Date: Mon, 27 Mar 2000 17:58:47 +0200
From: "Nordh, Valter" <valter at che.chalmers dot se>
Subject: Compiling qpopper 3 on IRIX
--=====================_58813203==_.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed
When I have tried to compile qpopper 3, latest version I get ALOT of
warnings. I am new to this, but I ran into a solution that might intrest
som people:
From comp.sys.sgi.admin thread with subject "Compile Warning - Help!!!"
I haven't had time to try, but it may help some people...
/Valter
>
> When I compiled Qpopper3.0b, I met the message below, and compiling was
> finished.
>
> ----------------------------------------------------------------------------
> ---------------------------------
> ld32: WARNING 84: /usr/lib32/libsocket.so is not used for resolving
> any symbol.
> ld32: WARNING 134: weak definition of ffs in /usr/lib32/libnsl.so
> preempts that weak definition in /usr/lib32/mips3/libc.so.
>
> This is a part of Error message.
It isn't an error message (that's why it says "WARNING"). See man 3
intro. Apparently, the code uses SVR4 networking, as a result compiles
with -lnsl, and the warnings are telling you that you are superseding
network calls from libc by using the nsl library.
(3N) There are actually 3 types of networking in IRIX.
1) BSD sockets implemented in the kernel, along with SUN RPC and
NIS (YP). The functions that implement these calls are in libc
[see (3C) section above].
2) SVR4-style STREAMS/TLI networking (not sockets), along with
SVR4-style RPC. SVR4-style RPC doesn't work with sockets, only
with the TLI. Also, SVR4-style networking does not support NIS
[see intro(3N)]. The functions that implement these calls are in
libnsl, and to use them, the code must be compiled with :
cc -D_SVR4_TIRPC prog.c -lnsl
The messages tell you "you are using 2), not 1)". Not more, not less.
> whatever you do, whenever you are You've got to be "HAPPY"!!! ^^
I am, but I can assure you that is not an obligation ;).
--
<standard disclaimer: these are my personal views, not SGI's>
Alexis Cousein al at brussels.sgi dot com
Systems Engineer SGI Belgium
--=====================_58813203==_.ALT
Content-Type: text/html; charset="us-ascii"
<html>
<font size=3>When I have tried to compile qpopper 3, latest version I get
ALOT of warnings. I am new to this, but I ran into a solution that might
intrest som people:<br>
From comp.sys.sgi.admin thread with subject "Compile Warning -
Help!!!"<br>
<br>
I haven't had time to try, but it may help some people...<br>
<br>
/Valter<br>
<br>
> <br>
> When I compiled Qpopper3.0b, I met the message below, and compiling
was<br>
> finished.<br>
> <br>
>
----------------------------------------------------------------------------<br>
> ---------------------------------<br>
> ld32: WARNING 84: /usr/lib32/libsocket.so is not used for
resolving<br>
> any symbol.<br>
> ld32: WARNING 134: weak definition of ffs in
/usr/lib32/libnsl.so<br>
> preempts that weak definition in /usr/lib32/mips3/libc.so.<br>
<br>
> <br>
> This is a part of Error message.<br>
<br>
It isn't an error message (that's why it says "WARNING"). See
man 3<br>
intro. Apparently, the code uses SVR4 networking, as a result
compiles<br>
with -lnsl, and the warnings are telling you that you are
superseding<br>
network calls from libc by using the nsl library.<br>
<br>
(3N) There are actually 3 types of networking in IRIX.<br>
<br>
1) BSD sockets implemented in the kernel, along with SUN RPC and<br>
NIS (YP). The functions that implement these calls are in libc<br>
[see (3C) section above].<br>
<br>
2) SVR4-style STREAMS/TLI networking (not sockets), along with<br>
SVR4-style RPC. SVR4-style RPC doesn't work with sockets, only<br>
with the TLI. Also, SVR4-style networking does not support NIS<br>
[see intro(3N)]. The functions that implement these calls are in<br>
libnsl, and to use them, the code must be compiled with :<br>
<br>
cc -D_SVR4_TIRPC prog.c -lnsl<br>
<br>
The messages tell you "you are using 2), not 1)". Not more, not
less.<br>
<br>
> whatever you do, whenever you are You've got to be
"HAPPY"!!! ^^<br>
<br>
I am, but I can assure you that is not an obligation ;).<br>
<br>
-- <br>
<standard disclaimer: these are my personal views, not SGI's><br>
Alexis Cousein
</font><font size=3 color="#0000FF"><u>al at brussels.sgi dot com</u></font><font size=3>
<br>
Systems Engineer SGI Belgium<br>
</font><br>
</html>
--=====================_58813203==_.ALT--
Date: Mon, 27 Mar 2000 17:57:01 -0600
From: Brent <bierblb at netins dot net>
Subject: EOF or I/O error
I'm using mailman (a web email client) and it keeps giving this error
in my logs for every logoff. I have mailed the makers of mailman,
but they said that this only happens once in a blue moon, ie they
have no idea. Could anyone help me to stop this from flooding my
logs? Thanks in advance!
Mar 27 14:47:42 athena popper[3874]: gregm at athena.unlimitedweb dot net:
-ERR POP EOF or I/O error (0)
Brent
Subject: Re: Compiling qpopper 3 on IRIX
Date: Tue, 28 Mar 2000 08:30:41 +0200 (MEST)
From: Georg Schwarz <schwarz at physik.tu-berlin dot de>
> > ld32: WARNING 84: /usr/lib32/libsocket.so is not used for resolving
> > any symbol.
> > ld32: WARNING 134: weak definition of ffs in /usr/lib32/libnsl.so
> > preempts that weak definition in /usr/lib32/mips3/libc.so.
the first line means that you can (and should) probably leave out
-lsocket in the Makefile. Just try it, if it still links without.
--
Georg Schwarz schwarz at physik.tu-berlin dot de
Institut für Theoretische Physik Tel. +49 30 314-24254, FAX -21130
Technische Universität Berlin http://home.pages.de/~schwarz/
Sekr. PN 7-1, Hardenbergstr. 36, D-10623 Berlin IRC: kuroi