Re: Sieve notify implementation
It seems to be an older protocol not used much anymore. Ideally, I would like to be able to send notifications via other IMs (eg. IRC, ICQ, AIM, etc.). Would this require a new implementation or are there zephyr gateways/bridges which could be used. I've always thought it would be nifty to do a Jabber notifier, but tis one of those things I'll probably never get around to. That, and the possibility of Sieve scripts to non-user folders, perhaps via folder annotations, would be rather cool IMHO Amos --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Problems with 2.2.10 on HPUX 11i
Connie Starr Fensky wrote: *I've just upgraded to imap 2.2.10 (with sasl 2.1.10) on my 11i mail server. I did this on a test box, but could not simulate our load (several thousand connections, ~8 messages delivered a day), and had no troubles. But, of course, putting on the mail server and starting up, performance is abhorrent. I'm running 100% cpu utilization AND memory utilization all of the time, which make the whole system sluggish and the users grumpy. All my fault, and I need to fix it yesterday. * * I have a pretty basic system--same as can be to my old 2.0.16 one that ran like a dream. Does anyone have any ideas for me--here is configure script: # csf 12/13/04 This is the command I used for configure CC=/opt/ansic/bin/cc CFLAGS='-g +DAportable +z' LDFLAGS='-L/usr/local5/db-3.3.11/lib -L/usr/local5/lib -L/usr/local/lib' LIBS=-lpthread ./configure -with-auth=unix --with-cyrus-prefix=/usr/local5 --prefix=/usr/local5 --with-sasl=/usr/local5/cyrus-sasl-2.1.20/ --with-dbdir=/usr/local5/db-3.3.11 --with-bdb-libdir=/usr/local5/db-3.3.11/lib --with-bdb-incdir=/usr/local5/db-3.3.11/include --with-lock=fcntl --with-mboxlist-db=flat Holy Moly. A flat mboxlist? Yeah, that'll be slow. Any chance you can upgrade that to skiplist? I don't know if that'll resolve all issues, but I should think that would help considerably Amos --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: 64bit capability bug
Alex Deiter wrote: Henrique de Moraes Holschuh wrote: Cyrus IMAP on 64bit arch incorrectly interprets defaults numerical parameters of a imapd.conf: all of them are equal to zero! Can't confirm on alpha, gcc 3.3.5: And while I haven't tried with 2.2, Debian has 2.1 compiled for at least ia64, alpha and amd64. It is also possible that the hppa and sparc ports are 64bit nowadays, too. Alpha, IA64, AMD64 is (like i386) a little endian machine. Sparc and HPPA is big endian. Cyrus-IMAP 2.1.x and 2.2.x (up to 2.2.3) work fine on sparc64. After 2.2.4 (on 2.2.10) i got error: Dec 27 21:23:19 satira nntp[24755]: DBERROR db4: PANIC: fatal region error detected; run recovery So do you compile Berkeley DB for 64-bit? Haven't tried it yet myself. At some point will need to do 64-bit though because getting some large quota needs for some shared folders. Amos --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Question about lmtp.lock
David G Mcmurtrie wrote: On Mon, 13 Dec 2004, Michael Loftis wrote: There are a number of different lockfiles that cyrus uses during it's lifecycle. Though one thing you need to know is that it is not designed to run like you're attempting to run it. Each cyrus instance is meant to be standalone, and running them all from a shared disk could very likely cause mailbox corruption, as well as skiplist corruption for the seen/subs entries and such. The only part of cyrus that may be multi-node from same storage safe is the BerkeleyDB stuff used for the mailboxes and deliver DBs. The cluster filesystem honors POSIX locking semantics across all nodes in the cluster. Take a look at MURDER, that's what you'll want if you want to cluster Cyrus. MURDER+heartbeat with a shared storage medium provides for failover (heartbeat is a separate, generic failover package). We chose a clustered environment over murder because we wanted scalability and redundancy, not just scalability. You're using the Cyrus code from CVS (3.0 branch?), right? Based on what I've picked up in the past, you'll need that for such a clustered environment. I know Ken had done some work on such a layout, but I don't know what the present state is. (I think in that situation there were some issues with Sun's QFS, but not sure.) Amos --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: how to post to shared folders
Ken Murchison wrote: Keep in mind, that unless jstaerk and chanser use SMTP AUTH and Postfix passes this info to Cyrus, that they still won't be able to post to the mailbox. Most folks end up giving the 'p' right to user 'anonymous'. I've been too distracted with other things to follow it closely, but I think I noticed in the announcement for a recent Postfix snapshot that AUTH credentials are now being cached in the queue file. This means that eventually it may be possible to use AUTH creds for posting like you can currently do with Sendmail. Amos --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: db4 backend alternative
Incidentally, any experiences with Berkeley DB 4.3? Michael Sims wrote: Mark Hannessen wrote: just wondering... are there alternatives to the db4 backend for storing mail and such? I am thinking in the direction of mysql. (I have openldap as well, but that doesn't seem to be the right place to store mail) I am really wondering about this because I just can't seem to get db4 to run stable. Nov 22 22:07:51 xp2600c imap[15437]: DBERROR db4: PANIC: fatal region error detected; run recovery See: http://acs-wiki.andrew.cmu.edu/twiki/bin/view/Cyrus/WhatDatabaseBackend --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Question on ACL's
On Fri, 30 Jan 2004, Jason Williams wrote: > Hmm. From the sound of it, that will require some further research and > testing. As it stands right now, I have smtp-auth setup through postfix in > a chrooted environment. Now, I need to figure out how to set postfix to > either pass AUTH or authenticate to lmtpd... On Mon, 2 Feb 2004, Jason Williams wrote: > So the trick that I am presented with is to figure out how to get my MTA > (Postfix) to authenticate to the LMTP process in order to have acl's work > correctly on shared mail boxes, correct? > > Has anyone done this with postfix by chance? Sadly, it won't work, at least not with Postfix. There's no way to save the auth state accepted via SMTP so that it can later be used by LMTP for authenticated posting. To be honest, I'm skeptical that'll ever be a feature because it would require changing the Postfix internal queue file format, and a big catch with that is some feeling of uneasiness about passing such credentials. At least that was my impression a couple of years ago. If passing such credentials is really important to you, you may have to use Sendmail. It know Sendmail handles it. -- Amos
Re: nntp fiddling
> On Tue, 30 Dec 2003 15:59:16 -0500, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> I added support for the Xref header to CVS last week. Cool. Yup, now even tin happy. -- Amos
Re: nntp fiddling
> On Tue, 30 Dec 2003 15:12:30 -0500, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> I specifically do not serve user.* via NNTP, mainly because I km> figured somebody might screw something up. But since I now have km> relatively good access controls, I can probably remove this km> restriction. km> I also thought it might be disconcerting for users to see their km> personal folders (e.g. user.ken.*) as "newsgroups". I'm inclined to agree. Besides, I think NNTP of read/write folders that generally won't be expired, like an INBOX, probably wouldn't work out so well. Maybe a brief comment in imapd.conf would be helpful, though. However, I'm realizing that a lot of our shared folders tend to be used more like NNTP groups than IMAP read/write folders, so this NNTP access is quite intriguing. Especially when the access permissions also apply. Very cool. Being able to set squat (that's so funny to say) and expire via annotations is so cool. km> How are things running? Any issues with NNTP clients or servers? So far, things are looking pretty good. The only oddity I've seen is from a reader called tin that at one time seemed to be pretty popular here: Your server does not have Xref: in its XOVER information. Tin will try to use XHDR XREF instead (slows down things a bit). I haven't tested it with other servers, mainly because our ancient news server has had terrible feeds for a long time, and has suffered serious hardware problems. -- Amos
nntp fiddling
So, if I leave this imapd.conf setting blank, does that mean all folders that I have access to via IMAP should appear as NNTP news groups (authenticated login)? newsprefix: Prefix to be prepended to newsgroup names to make the corresponding IMAP mailbox names. Just fiddling around it seems as if I see everything except those folders under "user.". Is that correct? I noticed some old posts that indicated any folder should be viewable, but so far in my test environment I'm not seeing that Mainly just curious. It's funny. Ages ago I remember a friend going on about how cool it was to read mail via nn. Personally I couldn't get my head around it, but now I'm playing with NNTP access to an IMAP server. Funny how things go around like that, given enough time. (Of course I'm posting this via Gnus, another news reader! ;-) -- Amos
Re: 2.2.2-beta new cyradm features?
>>>>> On Tue, 23 Dec 2003 22:33:02 -0600, >>>>> Amos Gouaux <[EMAIL PROTECTED]> (ag) writes: ag> I notice in the 'help' command for 2.2.2-beta cyradm there's ag> the following: ag> mboxcfg, mboxconfig configure mailbox ag> setinfo set server metadata ag> What do these do? Is this related to annotations? Ah yes. Pretty slick. localhost> mboxcfg news.utd.test expire 7 localhost> info news.utd.test {news.utd.test}: expire: 7 lastupdate: 22-Dec-2003 02:31:30 -0600 partition: news2 size: 135530 -- Amos
Re: ldap ptloader
> On Tue, 23 Dec 2003 16:21:02 -0500 (EST), > Igor Brezac <[EMAIL PROTECTED]> (ib) writes: >> I notice the imapd.conf man page mentions the 'memberOf' attribute. >> Unless I'm mistaken, that's a bit of a controversial thing, huh? ib> Why is that? Oh, when googling around and digging through various forums I was getting the impression that the 'memberOf' approach wasn't too well supported by the OpenLDAP community, or at least at one point. ib> You lose the group functionality with this approach, although you ib> get better performance. ...and this was the same argument for the roles mechanism in SunONE. -- Amos
2.2.2-beta new cyradm features?
I notice in the 'help' command for 2.2.2-beta cyradm there's the following: mboxcfg, mboxconfig configure mailbox setinfo set server metadata What do these do? Is this related to annotations? -- Amos
ldap ptloader
Just curious if anybody out there has been fiddling with the ldap ptloader experimental code. Thoughts? Comments? Observations? I'm just now starting to look into it myself, and it doesn't seem like there's been much chatter about it -- Amos
Re: ctl_cyrusdb -r tries to malloc 3GB, and fails.
> On Fri, 19 Dec 2003 16:55:29 +0100, > Paul Boven <[EMAIL PROTECTED]> (pb) writes: pb> The system in question is Solaris 9 Sparc, cyrus-imapd-2.2.2-BETA, pb> cyrus-sasl-2.1.17. Berkely is db-4.1.25. I don't know if this will alter things any, but since you're already using latest of the other software, why not use latest of db too? (Not, "DB2". HA! Sorry, that's pretty lame.) -- Amos
Re: imap-2.1.16 and berkeley db 4.2.52?
> On Mon, 15 Dec 2003 10:55:34 -0200, > Andreas <[EMAIL PROTECTED]> (a) writes: a> Has anybody tried to compile cyrus-imapd-2.1.16 with berkeley db 4.2.52? a> Should it work? So, what are folks' impressions of BDB 4.2.52 so far? Does it seem to be pretty solid? -- Amos
Re: Is there a limit on mailbox name length in Cyrus ?
> On Fri, 21 Nov 2003 18:35:12 +, > Patrick Welche <[EMAIL PROTECTED]> (pw) writes: pw> While talking limits, is there a limit to the number of UIDs that cyrus pw> can take, say in a SEARCH or UID STORE command? (While trying to sort out pw> a mailbox with thousands of message, I saw an error along the lines of pw> "word too long"?) I've seen that from time to time when manipulating some ridiculously large (~40K messages) folders using GNUS. I started looking into it, but didn't get far as there are way too many other things I'm supposed to be doing. -- Amos
Re: Highly available and scalable Cyrus configurations
> On Tue, 7 Oct 2003 18:43:21 -0400 (EDT), > Rob Siemborski <[EMAIL PROTECTED]> (rs) writes: rs> The key about the usability of the filesystem is that file locks need to rs> be obeyed throughout the entire cluster, and mmap() needs to be efficient rs> (and able to deal with read() and write() being called on that file at rs> the same time). My efforts in this regard keep getting pushed back due to too many other things flying around, but I've been told that Sun Cluster *should* respect full local filesystem semantics. How close to reality that is, I have not been able to confirm. I'm hoping we might be able to do some experimenting later this term, but then I've been saying that for a while. Sigh. -- Amos
OpenGroupware.org (OGo)
A bit off topic, but I know this is something that has come up on this list from time to time over the years, and besides, they mention Cyrus! ;-) http://www.infoworld.com/article/03/07/10/HNogo_1.html http://opengroupware.org/en/index.html -- Amos
Re: [PROPOSAL] Sieve for shared mailboxes
> On Tue, 08 Jul 2003 14:33:53 -0400, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> IIRC, the original IMAP client written by Mark Crispin was written in LISP. Hey, some of us still use (e)LISP mail clients, not that I'm encouraging a client/language debate. Really. ;-) -- Amos
Re: [PROPOSAL] Sieve for shared mailboxes
> On Tue, 08 Jul 2003 09:59:20 -0400, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> Now that Cyrus 2.2 has support for mailbox annotations, I believe km> that we can provide the functionality that people desire. I just have to say: WOOHOO!!! :-) OK, perhaps a little premature, but this sounds really cool and terribly useful. Now I need to read up on this annotate stuff. -- Amos
Re: Logs filling with strange error
> On 07 Jul 2003 18:17:29 -0700, > Scott Bronson <[EMAIL PROTECTED]> (sb) writes: sb> As long as they don't indicate something amiss, I can live with them. sb> Thanks. Incidentally, they're gone in 2.1.14. -- Amos
Re: sieve script not running?
> On Mon, 7 Jul 2003 21:34:37 -0400 (EDT), > Igor Brezac <[EMAIL PROTECTED]> (ib) writes: >> > gcc -L/usr/local/lib -R/usr/local/lib -L/usr/local/BerkeleyDB.4.0/lib >> > -R/usr/local/BerkeleyDB.4.0/lib -L/usr/local/lib -R/usr/local/lib >> > -L/usr/local/lib -L/usr/local/lib -o test \ >> > test.o libsieve.a ../et/libcom_err.a ../lib/libcyrus.a >> > Undefined first referenced >> > symbol in file >> > shutdown../lib/libcyrus.a(util.o) >> > ld: fatal: Symbol referencing errors. No output written to test >> > collect2: ld returned 1 exit status >> > gmake: *** [test] Error 1 >> > ib> You are missing -lsocket -lnsl in your link statement. Maybe that library changed between 2.1.12 and 2.1.13 to require those libraries? I notice the compile line is the same for both, but since 2.1.13 the error above appears. Yeah, as you say, adding those two fixes link problem. Looks like configure.in is missing something that the others have. -- Amos
Re: sieve script not running?
> On Mon, 07 Jul 2003 21:13:22 -0400, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> Which version of Cyrus gives the problem below? Reviewing my logs from my build script, it appears to have started with 2.1.13. Yeah, 2.1.12 compiled just fine. Sorry for not mentioning it sooner. Kinda forgot about it. -- Amos
Re: sieve script not running?
> On Mon, 07 Jul 2003 15:56:17 -0400, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> Test your script/message combo by running the them through the km> sieve/test.c program in the distro. Oh, that reminds me of something. I generally try to compile that because it's handy for manually testing scripts (will this work with 2.2?). However, last time I tried, got this: + cd sieve + gmake test gcc -c -I. -I.. -I. -I./../lib -I./../et -I/usr/local/BerkeleyDB.4.0/include -I/usr/local/include -I/usr/local/include -I/usr/local/include -DHAVE_CONFIG_H -I. -I. -Wall -g -O2 \ test.c test.c: In function `hashheader': test.c:70: warning: subscript has type `char' test.c: In function `parseheader': test.c:108: warning: subscript has type `char' test.c:121: warning: subscript has type `char' gcc -L/usr/local/lib -R/usr/local/lib -L/usr/local/BerkeleyDB.4.0/lib -R/usr/local/BerkeleyDB.4.0/lib -L/usr/local/lib -R/usr/local/lib -L/usr/local/lib -L/usr/local/lib -o test \ test.o libsieve.a ../et/libcom_err.a ../lib/libcyrus.a Undefined first referenced symbol in file shutdown../lib/libcyrus.a(util.o) ld: fatal: Symbol referencing errors. No output written to test collect2: ld returned 1 exit status gmake: *** [test] Error 1 This is on Solaris 8 with gcc-3.2.3. -- Amos
Re: Larry Greenfield leaving CMU
I'll always remember his help in the early days of the skiplist db. I can't imagine getting through that rough spot without his help. -- Amos
Re: cyrus-imapd-2.1.13 sieve curiosity
> On Mon, 9 Jun 2003 12:50:16 -0400 (EDT), > Rob Siemborski <[EMAIL PROTECTED]> (rs) writes: rs> I don't think so, I'm pretty sure its only caused by errors during rs> sieve_script_parse, not during sieve_execute_script. I'm not sure yet, but I think I might have found the culprit. The other day my boss was playing with uploading Sieve scripts using Mulberry. It would seem that a string of binary characters was included in this script after the comment '# Generated from text'. I've commented out that line. We'll see if things resume back to abby-normal. When I learn more just what he did, I'll contact Cyrusoft. -- Amos
Re: cyrus-imapd-2.1.13 sieve curiosity
> On Mon, 9 Jun 2003 12:14:15 -0400 (EDT), > Rob Siemborski <[EMAIL PROTECTED]> (rs) writes: rs> The sieve parser state doesn't get reset properly after an error. Thus rs> once a script fails to parse, all scripts will fail to parse (try to rs> submit a good script after a bad one in a timsieved session to see this in rs> action). Interesting. Would an error like this also apply? sieve runtime error for X id <[EMAIL PROTECTED]>: Fileinto: Mailbox does not exist (The "X" is one of our users, but not myself.) -- Amos
cyrus-imapd-2.1.13 sieve curiosity
I and a few others here have begun noticing that periodically Sieve won't filter a message properly. Sometimes I've even resent the message again to find that on that occasion it did get filtered just as expected. I wonder, could that BerkeleyDB leak for which Rob just posted those two patches have any bearing on this? Sadly, so far I haven't been able to detect a pattern to make this report more meaningful, but I'll keep an eye out. -- Amos
Re: need organizational hint
> On Sat, 12 Apr 2003 09:41:12 -0500, > Phil Howard <[EMAIL PROTECTED]> (ph) writes: ph> Not all users want, or do anything with, a separate spam folder. We've been experimenting with this very approach this semester. That is: between the gateway MTA and the Cyrus server, mail gets processed by SpamAssassin and is "tagged" appropriately. If folks wish it, they can then use Sieve to save spam into a JunkMail folder. We've then got a weekly cron job (probably could use Cyrus events) to perform the following: ipurge -f -s -d 60 "user.*.JunkMail" So far we haven't made a big deal about this, but have set up a number of folks. Overall, I think it is safe to say the response has been very positive. While I imagine some folks do forget to check their JunkMail folder, based upon some of the comments I've heard so far, it would seem that a fair number don't. The other day I was about to get agitated because a user forwarded a bunch of spam (with full headers) and I thought they might do that for all the mail in their JunkMail folder. Turns out they were either messages that SpamAssassin didn't tag, or the score was too low. Upon further research I learned that this individual does indeed check the JunkMail folder on about a weekly basis. To be honest, no system is going to be perfect. That's just fact of life, especially with such tenacious foes to contend with. While I imagine some would enjoy the ability to customize some of this tagging behavior, and I do plan on somewhat following developments along those lines, based upon what we've observed so far, I strongly suspect very, very few would opt to immediately reject the mail instead of utilizing the JunkMail folder. At least in an EDU environment, folks get really touchy when it comes to systematically blocking things. ;-) Anyway, that's my contribution. -- Amos
Re: MTAs that pass SMTP AUTH?
> On Mon, 31 Mar 2003 08:42:04 -0500, > Scott Balmos <[EMAIL PROTECTED]> (sb) writes: sb> Okay, fine. This is what I have also. The crux seems to be getting the MTA to sb> pass along the AUTH info. So far I guess only Sendmail and Exim do such a sb> thing, right? Has anyone *possibly* come up with a patch for Postfix about sb> this? I remember a few days ago some mumblings on the list that to record sb> such AUTH info to pass along with the message would be somewhat irritating. I think I was volunteered to raise this issue on the postfix-users list, but life is still a bit chaotic for me (still unpacking from a move). It's also been a long, long time since I've studied Postfix source at all, and much has changed since my meager tinkerings. So I was trying to re-acquaint myself with such things, and how something like this might be arranged. As has been mentioned before, one thing is for sure: it would require altering the format of the queue file. I *think* that if this was strictly controlled by some variables, which by default were off, *maybe* this would be accepted. It certainly doesn't harm that an increasing number of MTAs start supporting this. The more the merrier! :-) At any rate, this isn't likely to be something that appears overnight, alas. sb> This is *EXACTLY* what I have right now, Kevin. I've always thought that since sb> there is no password, and the "user" to authenticate is in the message sb> itself, such that anyone reading that message sees the full address along sb> with a username that has posting credentials to that folder, it was sb> completely insecure. I guess it's just a risk, you only hope that the users sb> (in my case, only about 300, so it's not that big a deal) don't abuse it, and sb> you just make sure the folder admins are quick to delete. sb> Well that makes me feel at least a little more comfortable knowing that at sb> least one other person does this convoluted user+folder "authentication" sb> setup like I was thinking of using. :) Well, what we did in some of these cases is a pretty gross, convoluted mess. I guess it works OK, but certainly could be better. You see, we got into a bit of a pinch when we tried to rely exclusively on shared folders ("BBs") for important campus discussions. The best we could do that was acceptable to all parties was to continue to use the long-time existing campus mailing lists AND some shared folders. The shared folder membership is derived from the mailing list membership. This shared folder is then listed as a special member on that list. If an individual wants to only refer to the shared folder instead of getting inundated by stuff in their inbox, then they can set the VACATION flag on their subscription for that list. This means they're still on the list, hence still a member of the shared folder, but don't get the mail in their inbox. Though not entirely fool-proof, this crazy arrangement seems to have worked out OK for the last couple or so years (forget when it was started). With some MLMs you can remove some headers and do a little masquerading, thereby tightening things up a tad. At least some of the IMAP diehards can enjoy the shared folders. However, LMTP-AUTH tied to SMTP-AUTH sure would be nice to play with. In the back of my mind I've been wondering if/how the new NNTP support coming with the new Cyrus now in beta might be employed to deal with these sort of discussions. However, I haven't pictured just how yet. -- Amos
Re: delivering to bb folders
> On Thu, 27 Mar 2003 15:43:17 +0200, > Nikos Voutsinas <[EMAIL PROTECTED]> (nv) writes: nv> This is not a cyrus topic I agree, but unless I'm mistaken, one of the Cyrus developers contributed the SMTP-AUTH code to Sendmail. Anyway, I have just one more question. nv> From RFC 2554: nv> http://www.ietf.org/rfc/rfc2554.txt nv> The optional AUTH parameter to the MAIL FROM command allows nv> cooperating agents in a trusted environment to communicate the nv> authentication of individual messages. But what is meant by "trusted environment", and how does one establish such a "trusted environment". -- Amos
Re: delivering to bb folders
> On Thu, 27 Mar 2003 08:50:57 -0300, > Henrique de Moraes Holschuh <[EMAIL PROTECTED]> (hdmh) writes: >> I think section 5 of RFC 2554 covers it fairly well. hdmh> It does. Amos, would you kindly post this little bit on the postfix list, Indeed. hdmh> please? I dare not :) Stinker. ;-) hdmh> Propagating auth info in postfix requires internal changes to the queue hdmh> format, though, so it will take its sweet time... Yeah, and as I recall from previous discourse, Wietse seemed to be fundamentally displeased by the notion of passing auth credentials through a chain of possibly unknown SMTP servers. I think the part he was picking up on is: "If the server trusts the authenticated identity of the client" Well, how is this trust established? I can see his concern when going from a remote SMTP server to another possibly remote SMTP server. Though, if you accept it at the SMTP server, surely it should still be valid for LMTP, the final end point, right? I don't know. I'll think about it while running some errands today. -- Amos
Re: delivering to bb folders
> On Wed, 26 Mar 2003 17:05:22 +0200, > Nikos Voutsinas <[EMAIL PROTECTED]> (nv) writes: nv> Lawrence Greenfield wrote: >> You might have to add -D_FFR_AUTH_PASSING=1 to your site.config.m4, nv> You realise that this feature of sendmail (was already available in nv> early releases of 8.12) combined with an MSA server and cyrus ACLs nv> can really boost cyrus functionality. (The "anyone p", mentioned in nv> cyrus documentation, has been always inaccurate, underestimating nv> cyrus capabilities). Alas, not all MTAs are capable of passing SMTP-AUTH onto the LMTP session, Postfix being at least one that I know of. In fact, I am skeptical that this -AUTH communication will ever be incorporated into Postfix without some sort of RFC to back it up. (Unless I'm mistaken, this continuity of -AUTH credentials isn't covered too well, but please correct me if I'm wrong.) -- Amos
Re: Why are only admins allowed to AUTH to lmtpd?
> On Fri, 3 Jan 2003 14:55:10 -0500 (EST), > Rob Siemborski <[EMAIL PROTECTED]> (rs) writes: rs> There's no reason regular users should be submitting to the LMTP server, rs> they should be submitting using SMTP to an SMTP server, and then the LMTP rs> server trusts the SMTP server. This (admitedly marginaly) simplifies the rs> authorization code in lmtpd. How does the posting ("p") right fit in? -- Amos
Re: [PATCH][CVS IMAPd 2.1] lmtp_downcase_rcpt implementation (Re:Case Sensitivity)
> On Wed, 25 Dec 2002 00:37:19 -0600 (CST), > Scott Smith <[EMAIL PROTECTED]> (ss) writes: ss> Damn, I fouind it earlier. Do a search for canonical, and/or lowercase ss> and/or regex. Really, it applies to any map lookup, which is why we haven't run into any problems... ss> It's actually funny: Back in ~2000 Weitse sent several E-mails stating ss> that Postfix lowercased everything because he felt it was simply ss> unacceptable to require end users to know what CaPiTaLiZaTiOn (he even ss> typed it that way!) a user name was, even though the RFC specified that ss> the MTA be case sensitive. ss> Cut to earlier this year, and he decides to stop lowercasing everything... Yeah, for one that professes displeasure at inconsistency, it seems ironic. When I saw the latest discussion I had a twitch in my gut, but life has been so hectic lately that I didn't dive in. Oh well -- Amos
Re: Case Sensitivity
> On Mon, 23 Dec 2002 19:12:42 -0200, > Henrique de Moraes Holschuh <[EMAIL PROTECTED]> (hdmh) writes: hdmh> On Mon, 23 Dec 2002, Vernon A. Fort wrote: >> >> Cyrus-2.0.16 to cyrus-2.1.11 using the same version of postfix. >> hdmh> Well, as a workaround, use pipe delivery on postfix (and thus, use the hdmh> deliver Cyrus program, which is much slower than LMTP), and tell postfix to hdmh> lowercase the recipient. I guess this is why the original Postfix lmtp client downcased everything. I remember noticing it at the time, but it didn't cause me problems, so I didn't think much of it. When the discussion came up on postfix-users after some folks found this problematic, it was decided that it was wrong of the lmtp client to forcibly downcase everything. Now we run into this fun. I wonder, perhaps there could be a switch to the Cyrus lmtp server to downcase stuff? Maybe I'm wrong, but it seems like this is logic more appropriate for the LDA, not the conduit to that LDA. Thoughts? What does Sendmail do, anyway? -- Amos
Re: cyrus 2.2 status
> On Tue, 17 Dec 2002 20:28:57 -0600, > archive info-cyrus <[EMAIL PROTECTED]> (ai) writes: ai> --On Tuesday, December 17, 2002 4:24 PM -0800 Jonathan Marsden ai> <[EMAIL PROTECTED]> wrote: ai> | So the question becomes: what, if anything can non-CMU people do that ai> | would help cause a release of 2.2 (or 2.1 with virtdomains in it??) to ai> | happen sooner rather than later? ai> Just a guess... send Ken some money. Actually, perhaps the CMU developers ought to think of doing what the SpamAssassin developers are doing: create Amazon (or other vendors) wish lists so that folks eager for enhancements can demonstrate their appreciation. ;-) Ho. Ho. -- Amos
Re: Compiling cyrus-imapd for Solaris 2.5 ( Respost )
> On Wed, 23 Oct 2002 10:13:18 +0200, > Francesc Guasch <[EMAIL PROTECTED]> (fg) writes: fg> anything else I could try ? Upgrading Solaris? You're using a hugely buggy release -- Amos
Re: [STATUS] NNTP support (10/9/02)
Interesting discussion. At times we have created local usenet groups for class discussions. The ability to utilize ACLs offers interesting possibilities, it would seem. I have to admit, initially I was a bit dubious about this. Maybe it's because local interest in usenet has been dropping off in recent years. However, the more I see this progress, the more I am beginning to be intrigued as to how this might be put to use. -- Amos
Re: posting to shared mailboxes
> On 23 Sep 2002 18:26:16 -0400, > Matt Vanderveer <[EMAIL PROTECTED]> (mv) writes: mv> I'm having problems getting lmtpd authentication working, at least as I mv> think I need it to. Posting directly to the shared box from the users's mv> mua works fine. However for various reasons, I need to let users mail mv> to a shared box, not just post from the mua, and still limit who can mv> mail to the shared box. This leads to several questions: mv> Am I correct in thinking that my issue is on the Postfix side in getting mv> Postfix to pass the smtp-auth info through its lmtp client to cyrus' mv> lmtpd? Correct. In Postfix, credentials passed via SMTP-AUTH will *not* be conveyed to the LMTP session for LMTP-AUTH. I really don't see a solution to this for two reasons: 1) Postfix is not a monolithic application, and so the queue file format would have to be altered to retain the SMTP-AUTH info so that it can be used for LMTP-AUTH, and 2) past discussions revealed that Wietse was not comfortable with the lack of verification of the SMTP-AUTH credentials to blindly pass them onto other services. If this is an important requirement, you'll have to use Sendmail. -- Amos
Re: Cyrus IMAP Presentation
> On Tue, 17 Sep 2002 21:56:07 -0500, > Schmehl, Paul L <[EMAIL PROTECTED]> (spl) writes: spl> This simply isn't true. At UTD we are using altnamespace *and* spl> we allow our users to have a PEA (Personalized Email Address) spl> which can include firstname.lastname@. This is done by the MTA (Postfix in our case, but easily done with Sendmail and others) before delivery to Cyrus. At least from what I've seen/heard of in the past, this tends to be the typical approach instead of having folks login with a "first.last" login. However, the "[EMAIL PROTECTED]" inbox trick has been used in the past to try to emulate multi-domain support. Maybe with the stuff that is now being worked on this will no longer be necessary. (Sorry, I haven't followed some of those threads too closely since it isn't something I'm really needing at present.) -- Amos
Re: Cyrus IMAP Presentation
> On 17 Sep 2002 13:17:06 -0400, > Brian <[EMAIL PROTECTED]> (b) writes: b> I knew someone would mention that book ;-) ... but it really doesn't b> count. It was pretty much out of date by the time it came out and, b> while it did cover Cyrus, it was pretty broad in scope. Well, truth be told, I was a little frustrated that it made absolutely no mention of Postfix. I tried, but as I saw, putting something together like that is hell. b> While it was a good book, I really want one that is nothing but Cyrus b> version 2 written for admins (ie with no "What is IMAP?" chapter). Cyrus b> is probably one of the few really great and useful programs with a large b> user-base that has no real O'Reilly book to indoctrinate the unwashed b> masses still using those other IMAP servers. Yeah, and the new virtual server stuff that is being developed I suspect could be a rather meaty chapter on its own. -- Amos
Re: Cyrus IMAP Presentation
> On 17 Sep 2002 12:05:34 -0400, > Brian <[EMAIL PROTECTED]> (b) writes: b> I've often wondered why O'Reilly has never commissioned a book about b> Cyrus. With co-lo's getting cheaper and cheaper, I always recommend b> Cyrus whenever one of my sysadmin friends says they're going to do b> POP/IMAP, but it's always met with "Where can I read up on it?" Well, they kinda did in the Managing IMAP book. Alas, 2.x was just coming out of the shoot when that book was written, and the authors have been too strapped for time to do an update. I haven't talked to them in a while so don't know what their schedule is for an update. -- Amos
Re: Cyrus IMAP Presentation
> On Tue, 17 Sep 2002 16:52:28 +0200, > Luca Olivetti <[EMAIL PROTECTED]> (lo) writes: lo> I don't understand: this is a socket opened *by* cyrus, and postfix lo> is instructed to talk lmtp to that unix socket. lo> A local user could use the socket to inject mail to cyrus, but (s)he lo> could also use port 25. To quote <[EMAIL PROTECTED]>: The "lmtp" delivery agent should be listening on "private/lmtp", so it will work, but... This said, the public and private directories are for internal Postfix IPC. The LMTP server should be using a different subdirectory of /var/spool/postfix. http://marc.theaimsgroup.com/?l=postfix-users&m=102489182913823&w=2 -- Amos
Re: Cyrus IMAP Presentation
> On Tue, 17 Sep 2002 08:55:53 +0200, > Luca Olivetti <[EMAIL PROTECTED]> (lo) writes: lo> edit /etc/cyrus.conf and replace the existing lmtpunix line with lo> lmtpunix cmd="lmtpd" listen="/var/spool/postfix/public/lmtp" prefork=1 While this will work, as has been discussed on postfix-users, those sockets really are not for third party software such as Cyrus. I'm not sure it is wise to instruct folks to go down that path. -- Amos
Re: FYI: pop-before-smtp works with cyrus-imapd-2.1.4
> On 19 May 2002 14:38:52 -0700, > Ron Kuris <[EMAIL PROTECTED]> (rk) writes: rk> While this script works most of the time, it wasn't very reliable during rk> log rollovers. Try continuously rolling over the log to reproduce the rk> problem. Also, parsing log entries takes a lot more CPU than the patch rk> I provided recently. Precisely why we use DRAC. rk> My recent patch just updates access.db directly. No separate process is rk> required. While a separate process is required for DRAC, the nice thing about it is that it will clear out entries after some configurable amount of time. -- Amos
Re: Patch for SMTP after IMAP
> On 18 May 2002 18:10:34 -0700, > Ron Kuris <[EMAIL PROTECTED]> (rk) writes: rk> Then, whenever someone successfully logs into POP or IMAP, their IP rk> address is added as a RELAY into the sendmail access db, using the db3 rk> database routines. What about the DRAC patch that's already in the contrib directory? http://mail.cc.umanitoba.ca/drac/ This works for both IMAP and POP. -- Amos
Re: I Had attempted to do LMTP with Postfix ...
> On Thu, 03 Jan 2002 17:27:10 -0700, > Scott M Likens <[EMAIL PROTECTED]> (sml) writes: sml> Ahh... sml> Since the LMTP is well quite out of date i'm more or less trying to sml> figure out the password maps and everything i would need to setup. sml> I guess i'll wait for someone to update it and take another crack at sml> it. As far as the LMTP_README goes, it's not *that* out of date. I think the biggest one was the "lmtp -a" as a possibility. The lmtp_sasl_* stuff is the LMTP-AUTH I was referring to. If you compile Postfix with SMTP-AUTH then you get LMTP-AUTH as well. Though, the LMTP-AUTH in Postfix isn't quite as useful as that in Sendmail. (Because any auth specified via SMTP-AUTH is not conveyed over to the LMTP-AUTH sequence, but I digress.) To be honest, there hasn't been that much change in the Postfix LMTP client and Cyrus LMTP server since that readme was slapped together, so it is still valid. -- Amos
Re: I Had attempted to do LMTP with Postfix ...
>>>>> On Wed, 15 May 2002 20:04:27 -0300, >>>>> Henrique de Moraes Holschuh <[EMAIL PROTECTED]> (hdmh) writes: hdmh> On Wed, 15 May 2002, Amos Gouaux wrote: >> to take advantage of single instance delivery. (The local delivery >> agent in Postfix can only deliver one recipient at a time.) hdmh> Well, use the lmtp transport, then. It can deliver through an unix socket hdmh> just fine. If you use mailbox_transport, you'll still only get one recipient at a time. I don't recall if that's also true with local_transport. I think it is. However, if there's not a high probability that you'll see lots of mail addressed to many recipients on the same Cyrus partition, then it's a mute point. -- Amos
Re: I Had attempted to do LMTP with Postfix ...
> On Thu, 03 Jan 2002 13:23:55 -0700, > Scott M Likens <[EMAIL PROTECTED]> (sml) writes: sml> So then i can assume it would be more proper to do LMTP over Unix sml> Socket? I can do either, i just am trying to find the best method to sml> do this. The properness depends on your environment. TCP is handy if using separate boxes for Cyrus and MTA. Using TCP also makes it easier to take advantage of single instance delivery. (The local delivery agent in Postfix can only deliver one recipient at a time.) -- Amos
Re: I Had attempted to do LMTP with Postfix ...
> On Tue, 14 May 2002 12:09:14 -0600, > Scott M Likens <[EMAIL PROTECTED]> (sml) writes: sml> May 13 15:52:32 shell postfix/lmtp[21546]: [ID 197553 mail.info] sml> 614D48A84E: to=<[EMAIL PROTECTED]>, sml> relay=shell.bourg.net[207.229.76.2], delay=1, status=deferred (host sml> shell.bourg.net[207.229.76.2] said: 430 Authentication required) Ugh, README_FILES/LMTP_README really does need updating. I should try to do that during our inter-session. Anyway, if you're doing LMTP over a TCP connection, you'll either need to use LMTP-AUTH (like SMTP-AUTH), or use the "-a" flag as in: SERVICES { ... lmtp cmd="lmtpd -a" listen="[127.0.0.1]:lmtp" prefork=0 ... } Though, if you use the "-a" flag, be sure to restrict access to this LMTP server. This can be done by either binding to a specific IP address as done above and/or by using tcp_wrappers. sml> Also has anyone seen this with the new postfix 1.1.9-Experimental? sml> May 13 15:39:38 shell postfix/lmtp[17534]: [ID 947731 mail.warning] sml> warning: spurious attribute sender in input from lmtp socket Not yet. I was waiting for the dust to settle a bit before trying latest Postfix snapshot/release. -- Amos
Re: Duplicates triplicates ...
> On Mon, 06 May 2002 11:37:26 -0600, > Scott M Likens <[EMAIL PROTECTED]> (sml) writes: sml> Well i just wanted to say quite simply, i haven't gotten one sml> Duplicate yet. Either Postfix is removing them, or something. sml> Because as i use Mulberry client and enjoy it very much, i see not sml> one duplicate ever. Well, not to take Mulberry down, but I suspect this is a difference between those sites using the duplicate delivery trap in Cyrus, and those that aren't. We have duplicate delivery enabled and as I might expect, I haven't seen these dups either. Anyway, just speculation since I haven't actually looked into this matter myself. -- Amos
Re: sendmail vs postfix
> On Fri, 19 Apr 2002 18:50:48 +0300, > Nick Ustinov <[EMAIL PROTECTED]> (nu) writes: nu> Hey everyone, nu> Finally, I have upgraded to 2.1.3. I use skiplist for mailboxes and nu> duplicate db now. Cyrus works together with sendmail and the server is nu> pretty heavily loaded. mailboxes.db file is over 50Mb. The system receives nu> about 150-350 emails per minute, sendmail is set up it deliverymode=q with 1 nu> minute interval. So with Sendmail you've got everything getting queued, then later delivered. nu> At peak times load average was going to 15-20, now it's a bit less in nu> average, however sometimes goes even higher. Anyway -- isn't postfix nu> supposed to be better for such high loads? I've tried to install postfix nu> 1.1.7 (lmtp as mailbox_transport) however when load avg went to 50 I stopped nu> it and returned to sendmail. I know disk speed is essential here -- well, it nu> is fast. But with Postfix it tries to deliver as fast as possible. No wonder you're seeing higher load peaks. Try throttling Postfix to only a certain number of lmtp processes. The request will get queued until load drops off. In the maxproc column for lmtp in master.cf experiment with the numbers until a reasonable balance is struck. On our SMTP gateway we've currently got it set to 4. There's another way to keep all incoming mail queued and then let another process initiate deliveries, but I forget how that's set up. Besides, you may still want to throttle lmtp some. nu> Any comments? Maybe Exim will be better? I guess I will have to put MTA on nu> another machine, otherwise it keeps getting even worse. That's what we've been doing for quite some time now. We do lmtp over private network between the machines. Has worked out well. -- Amos
Re: Postfix/Cyrus reports "temporary failure"
> On Tue, 09 Apr 2002 22:49:25 -0500, > Dustin Puryear <[EMAIL PROTECTED]> (dp) writes: dp> Apr 9 22:37:41 mars postfix/pickup[23810]: 5521660E06: uid=0 from= dp> Apr 9 22:37:41 mars postfix/cleanup[26079]: 5521660E06: dp> message-id=<[EMAIL PROTECTED]> dp> Apr 9 22:37:41 mars postfix/qmgr[68938]: 5521660E06: dp> from=<[EMAIL PROTECTED]>, size=367, nrcpt=1 (queue active) dp> Apr 9 22:37:41 mars postfix/pipe[26082]: 5521660E06: dp> to=<[EMAIL PROTECTED]>, relay=cyrus, delay=0, status=deferred dp> (temporary failure) If this is still a problem, you might check your imapd.log or where ever you have Cyrus logging to. You could also edit Postfix's startup script in /etc/postfix so that the master process is started up with one or more of "-v". You might also try using the LMTP delivery instead of the now somewhat depreciated "cyrus" mailer. -- Amos
Re: moving from 2.0 to 2.1.3
> On Wed, 3 Apr 2002 22:19:02 +0300, > Nick Ustinov <[EMAIL PROTECTED]> (nu) writes: nu> We are planning to switch production server from 2.0.16 to 2.1.3/skiplist. nu> The system has about 60 mailboxes and is quite heavily loaded (seeing nu> alot of db lockers). Are there any hidden catches people i should avoid? nu> Besides converting mailboxes file using cvt_cyrusdb is there anything else I nu> should know about? How about sieve scripts? Are these compatbile? Also, is nu> it recommended to switch to skiplist only for mailbox file or for the rest nu> as well? We're using skiplist for mailboxes.db and used the defaults for the others (BerkeleyDB 4.0.14 for deliver and tls stuff). Out of paranoia, this is what I did: 1. shutdown cyrus 2. drop in new binaries for 2.1.3, how ever you do that locally 3. su - cyrus # the user owning the /var/imap contents 4. cd /var/imap 5a. rm -f deliver.db tls_sessions.db# SEE NOTE BELOW!!! 5b. rm -rf db && mkdir db 6. cvt_cyrusdb `pwd`/mailboxes.db db3 `pwd`/mailboxes.db.new skiplist 7. mv mailboxes.db.new mailboxes.db 8. start cyrus 2.1.3 Note about #5a and #5b: I did this out of paranoia that I didn't want any old BerkeleyDB log files to have any relationship what so ever with my new skiplist mailboxes.db. I don't know if it would have happened, but I feared BerkeleyDB might try to replay any mailboxes.db log files. I don't think this would have been a problem because BerkeleyDB should have spit out mailboxes.db as being untasty. But, like I said, I was just being paranoid. Proceed at your own risk! ;-) Further note: the skiplist back-end also keeps a file in the db directory. So just because you should at some point drop BerkeleyDB all together, I wouldn't go and blow away that db directory without care. Just something to keep in mind. I ran the steps above on a test box first before doing them on our production server. Our mailboxes file had something like 65K folders in it. I forget the exact time, but it was incredibly fast to convert this on our Sun E250. Something like a couple of minutes. It has been doing swell since. :-) Oh, and using BerkeleyDB for deliver and tls stuff seems to be fine. The number of lockers will climb during peak periods, but then taper back down when traffic subsides. It just seems like mailboxes.db was an application not terribly well suited for it. -- Amos
Re: sieve and shared folders
> On Tue, 02 Apr 2002 09:47:23 -0700, > Justin Wood <[EMAIL PROTECTED]> (jw) writes: jw> Does anyone know of any problems using sieve scripts with Shared jw> Folders? I have set altnamespace to yes and have normal user.name jw> users with sieve working just fine. I also have a couple mailboxes jw> in the shared space, but the scripts don't seem to be getting read. jw> I have run the test program on them just to make sure the syntax is jw> okay. Sieve only works with real users (folder prefix of "user."). There has been some talk about how to approach this, but at least for me, nothing terribly satisfactory. One idea that seems to be the most straight forward to implement would be to use a pseudo-user for non-user folders. However, my concern with this is that I would not want to have *all* the non-user folders run through Sieve. If you've got a lot of Shared Folders it seems to me this would become a nasty bottleneck, but maybe not? Also, how do you then allow different individuals to edit a Sieve script for different Shared Folders if there is only one script under this one pseudo-user? -- Amos
Re: Sieve and header manipulation?
> On Thu, 28 Mar 2002 10:11:58 +0100, > GOMBAS Gabor <[EMAIL PROTECTED]> (gg) writes: gg> Hello! gg> Are there any plans to extend sieve for manipulating the message during gg> delivery? It is very annoying that some mailing list software (like the gg> one at CMU (hint, hint :)) do not add "Reply-To:" headers to messages gg> sent to the list. With procmail, I can pipe the message through formail gg> to add a "Reply-To:" header if there isn't one already; it would be nice gg> to have similar functionality with sieve. Sieve is based on a standard with particular design goals in mind. You might want to visit http://www.cyrusoft.com/sieve/ -- Amos
Re: Cyrus-imap folders
> On Thu, 28 Mar 2002 11:02:33 +0100, > Thomas Jarosch <[EMAIL PROTECTED]> (tj) writes: tj> You have to tell Outlook to use "INBOX" as the root folder tj> of Cyrus and to store the outgoing mail on the server. Or use altnamespace, depending on version used. -- Amos
Re: tools
> On Mon, 25 Mar 2002 10:36:54 -0500, > Kiarna Boyd <[EMAIL PROTECTED]> (kb) writes: kb> Good Morning, kb> Any one using the cyrus-graphtools.1.0 in 2.1.1? Kinda. One thing I discovered is that it is reporting the current number of active sessions and the total number of *forks*. In one of the numerous hacks of this I've got running (learning rrdtool), I got it to report the number of active sessions and the number of connections. I've got to cleanup my mess, though. -- Amos
Re: Cyrus SASL Auth/IMAPd with Postfix errors....
> On Fri, 22 Mar 2002 13:09:07 -0800, > Andrew Klino <[EMAIL PROTECTED]> (ak) writes: ak> doesn't solve the following issue -- I have no idea WTF LMTP ak> is/does, nor how to configure it. No documentation I have looked at ak> yet has said anything about it. Could it possibly have anything to ak> do at all with the fact that my aliases do not work? I did a ak> "postconf |grep lmtp", and this is what it gave me... Did you look at the LMTP_README file in the README_FILES directory in the Postfix source distribution? That LMTP_README is old, but should still be valid. -- Amos
Re: Cyrus SASL Auth/IMAPd with Postfix errors....
> On Fri, 22 Mar 2002 21:24:30 +0100, > Birger Toedtmann <[EMAIL PROTECTED]> (bt) writes: bt> Do you use LMTP? LMTP can only handly deliveries of one recipient per bt> mail, so you should configure postfix appropriately (there is a flag for That's not exactly correct. From my /etc/postfix/main.cf: # Cyrus will hard link messages to multiple recipients # on the same Cyrus partition. lmtp_destination_recipient_limit = 3000 HOWEVER, local delivery in Postfix is only to one recipient at a time. This is a limitation of the local delivery agent, not of LMTP. To get around this limitation I use virtual table to route mail (via LMTP) from MX host to Cyrus host. In the case of the problem report, it is not clear how mail is being routed or how lmtp is being called. -- Amos
Re: multiple cyruses via SAN
> On Thu, 21 Mar 2002 08:32:37 -0800, > David Mendenhall <[EMAIL PROTECTED]> (dm) writes: dm> I know I've seen snippets about this run across this list, but does anybody dm> have a step-by-step for switching from Berkeley DB to skiplist in cyrus dm> 2.0.16 ? For 2.0.16? I don't think you can. You need to upgrade to 2.1.3. -- Amos
Re: linker wierdness
> On Thu, 21 Mar 2002 16:26:33 +0100, > Birger Toedtmann <[EMAIL PROTECTED]> (bt) writes: bt> The problem mostly lies with the header files. I had the same effect bt> once. I had installed BDB3 and 4 on one machine, and the db header file bt> (include/db.h) pointed towards BDB3 whereas the linker itself grabbed bt> /lib/libdb-4.so. at the end of the compile. Just be sure that db.h bt> points to the BDB version you're going to link in (I did it with a soft- bt> link to include/db4/db.h). It is for this reason that when I install BDB on our Solaris systems I no longer attempt to link everything into /usr/local/include and /usr/local/lib. I think there is a reason why Sleepycat by default wants to put these headers/libraries into an application-specific directory. Yes that makes compiling a bit more involved, but it also avoid all these double link problems that folks go on about. It also means I have no problems with BDB-3.x and BDB-4.x both being installed on the same system. No collisions. To some extent I think it was a mistake of Redhat to try to defeat the version-specific directories that Sleepycat does by default. Been there, done that -- Amos
Re: multiple cyruses via SAN
> On Thu, 21 Mar 2002 07:40:30 +0100, > Simon Matter <[EMAIL PROTECTED]> (sm) writes: sm> OT, but did you upgrade all dbs from DB3 to skiplist? Is it possible at sm> all? I'm not sure about this but as I understand the configure script it sm> should be possible now to use skiplist for every db. So my question is sm> whether this is true, if yes then it means we can compile cyrus without sm> depending on any DB[3,4] whathever, right? sm> I'd like this idea because building RPM (or other bin packages) is much sm> easier if you don't depend on too many libs. At this time I believe it would not be good to replace deliver.db or tls.db with skiplist. While skiplist works well for mailboxes.db, I don't think that would be the case for deliver.db or tls.db. There are many more writes going on to these than mailboxes.db. -- Amos
Re: multiple cyruses via SAN
> On Wed, 20 Mar 2002 10:40:20 +0100, > Birger Toedtmann <[EMAIL PROTECTED]> (bt) writes: bt> If you want a production system that scales and is high available/has failover bt> capablilities, set up several boxes with _separate_ userspace and multiplex bt> all connections to and from it according to the userspace definition. bt> You can then provide high availability for each box with shared disk access bt> and an additional _offline_ cyrus (only jumping in mounting the shared disk bt> in a failover situation). I've been following these threads pretty closely because it is an issue I'm interested in as well. Based on what I've seen, the glimpses I've had into the code, and my gut, at least right now I'd have to concur: it is definitely a possibility to have host failover in a cluster, but true load-balancing raises several nasty issues. Even if true multi-node cluster load-balancing were possible, I fear the performance might suck pretty bad. No matter how much clustering improves, there will always be a performance penalty for all that overhead, at least from what I've seen. Some have suggested MySQL as db backend. Is it now possible to have MySQL shared between two nodes, a parallel database server? If not, then you have that as a single point of failure. Will the extra overhead of a full-sized SQL engine be a bottleneck? I dunno. I did happen to notice that Sleepycat now offers "Berkeley DB High Availability" which apparently supports a clustered environment. Could possibly that in conjunction with a SAN lead the way to a fully clustered Cyrus environment? I dunno. Though, I will say this: we recently switched to skiplist for mailboxes.db and at least so far, it beats the pants off of Berkeley DB. So as much as a clustered Berkeley DB intrigues me, I fear performance will be the deciding issue. As for Berkeley DB performance, I suspect the only way---and once again this is more my gut speaking than anything else---to get better performance is to fully thread Cyrus so that it is just one process and cache the hell out of Berkeley DB in memory. At least this seems to be what the Netscape folks did with Directory Server a while ago. (I believe they also highly customized Berkeley DB, but maybe some of those tweaks eventually found their way to Sleepycat. I dunno.) Even if all of this was done, would Berkeley DB performance be acceptable in the end compared to what we're currently seeing with skiplist? I dunno. I kinda doubt it. Well, that's my rambling for this morning. -- Amos
Re: Silkymail: shm_attach()-error
> On Wed, 20 Mar 2002 11:55:08 +0100, > Christoph Krempe <[EMAIL PROTECTED]> (ck) writes: ck> Hi, ck> I'm just tring to install the latest Silkymail on Linux. Using I think you need to ask Cyrusoft about this -- Amos
Re: [Re: [Re: removing mail from box on command line]]
> On Thu, 14 Mar 2002 20:00:42 -0500, > E M Recio (emr) writes: >> ipurge -f -d user..Trash emr> I tried to run ipurge but it gives me an error, it doesn't recognise the -f emr> option. Time to upgrade! :-) -- Amos
Re: [Fwd: Locking in Cyrus / db3 - Some more evidence]
> On Tue, 12 Mar 2002 08:27:42 +0100, > Birger Toedtmann <[EMAIL PROTECTED]> (bt) writes: bt> [...] >> >> Previously, our Cyrus servers were compiled to use db3 for the mailboxes >> file and for the delivery databases. Users mailboxes would frequently >> lock. They could not move, delete, or copy messages from within their >> accounts until their imap PIDs were killed, or the Cyrus master was >> restarted. bt> Does anyone know whether this behaviour applies to BDB4 as well? I *think* this was a problem with an older release. What version was this, anyway? It's all a blur right now. -- Amos
Re: cyrus-imapd-2.1.3 w/ BDB4: small typo in log output ofimapd,pop3d?
> On Tue, 12 Mar 2002 08:10:26 +0100, > Birger Toedtmann <[EMAIL PROTECTED]> (bt) writes: bt> I just set up cyrus-imapd-2.1.3 on a RH7.2 box that has BDB4 installed (re- [...] bt> Mar 12 07:57:50 testbox pop3d[19141]: DBERROR db3: 9 lockers That's okay. I don't know why it logs that with "db3", but that's nothing to be alarmed over. -- Amos
Re: e-mail expiry (a strange request)
> On Mon, 11 Mar 2002 17:57:03 -0800, > Ian Macdonald <[EMAIL PROTECTED]> (im) writes: im> Hi, im> I've been asked by my employer to investigate possible tools for im> expiring e-mail after it reaches a certain age. The motivation behind Depending on the version you have, try ipurge that comes with Cyrus: NAME ipurge - delete mail from IMAP mailbox or partition based on age or size SYNOPSIS ipurge [ -f ] [ -C config-file ] [ -x ] [ -d days | -b bytes | -k Kbytes | -m Mbytes ] [ mailbox-pattern... ] -- Amos
Re: restoring email
> On Mon, 11 Mar 2002 15:24:27 -0600, > Connie S Fensky <[EMAIL PROTECTED]> (csf) writes: csf> Yes, you need to do a reconstruct. Also, you cannot create directories under csf> the inbox with the usual unix procedures--must do it with an imap client or csf> cyradm. csf> c* If you're using the latest and greatest, and the directory that's created using mkdir has a "cyrus.header" (probably can use touch with correct perms), then the '-f' option to reconstruct is allegedly supposed to do all the rest. At least that's what the reconstruct man page for 2.1.3 indicates. I don't know if we've actually used this already or not. I think so. -- Amos
Re: lmtp or deliver from postfix
> On 08 Mar 2002 10:22:40 +0100, > Ivo Panacek <[EMAIL PROTECTED]> (ip) writes: ip> My question is: would it help to use lmtp instead ? ip> And in case it would: how to do that (from cyrus side)? ip> Just uncomment #lmtp and comment out lmtpunix in cyrus.conf ? Search around in the Postfix distribution you have for LMTP_README. While that might be a little dated, it should still be pretty darn close. If you can't find a LMTP_README file, then you should really think about upgrading Postfix. -- Amos
Re: cyrus and SSL/stunnel
> On Thu, 7 Mar 2002 21:40:50 -0800, > Jeff Bert <[EMAIL PROTECTED]> (jb) writes: jb> I'm trying to get cyrus secured via SSL using stunnel and haven't been jb> successful yet... this is what I've tried: jb> editted cyrus.conf: jb> SERVICES { jb> ... jb> ... jb> pop3cmd="/usr/sbin/stunnel -p /etc/stunnel/stunnel.pem -l pop3d" jb> listen="pop3" prefork=0 jb> ... jb> ... jb> } jb> is anything like this possible? i need it secured via SSL for Windoze users. You're working too hard. You can provide SSL (TLS) alternatives like this: SERVICES { ... imaps cmd="imapd -s" listen="imaps" prefork=0 ... pop3s cmd="pop3d -s" listen="pop3s" prefork=0 ... } Then tell Cyrus where to find the certs using the imapd.conf settings tls_key_file, tls_cert_file, tls_ca_path, and tls_ca_file. See imapd.conf(5) for more info. Oh, and don't forget to list the ports in /etc/services: imaps 993/tcp # imap via ssl pop3s 995/tcp # pop via ssl -- Amos
Re: Which module in cyrus provides return receipt ?
> On Thu, 07 Mar 2002 14:14:46 -0500, > Cyrus Daboo <[EMAIL PROTECTED]> (cd) writes: cd> actually sees the message. I think the DSN extension is pretty widely cd> deployed in SMTP servers so this ought to be a fairly reliable method, cd> though sometimes firewalls can get in the way. I know Postfix is one that currently does not support DSN. Though, since we use virtual table to route via LMTP to Cyrus host, I wonder if Postfix would even send the DSN if it could. Maybe it would if it detected that the next hop, in this case the LMTP host, didn't support DSN. -- Amos
Re: Which module in cyrus provides return receipt ?
> On Thu, 7 Mar 2002 10:08:39 -0500, > Theodore J Knab <[EMAIL PROTECTED]> (tjk) writes: tjk> Problem: Management wants a return receipt feature for emails. For what, read receipt or delivery receipt? For delivery receipt, I wonder if that's something lmtpd could do? Isn't read receipt something the mail client would do? Though, if delivery failed the message would be bounced, so the sender *should* know from that whether the message has been delivered. -- Amos
Re: Configuration Cyrus-Imap 2.0.16
Did you compile with tcp_wrappers support? If so, do you have appropriate entries in /etc/hosts.allow? Just a guess > On Wed, 6 Mar 2002 18:39:59 +0100, > Ana Belén Díez <[EMAIL PROTECTED]> (abd) writes: abd> Hi, abd> I have a problem. I start /usr/cyrus/bin/master and I make a netstat with abd> the next answer: abd> # netstat -a | grep imap abd> *.imap *.*0 0 24576 0 LISTEN abd> localhost.32886 localhost.imap 32768 0 32768 0 abd> ESTABLISHED abd> localhost.imap localhost.32886 32768 0 32768 0 abd> ESTABLISHED abd> 30001c1f448 stream-ord 30002b0dc90 /var/imap/socket/lmtp abd> But when I'am testing only appear: abd> # telnet 127.0.0.1 143 abd> Trying 127.0.0.1... abd> Connected to 127.0.0.1. abd> Escape character is '^]'. abd> or abd> # imtest -m login localhost abd> C: C01 CAPABILITY abd> My /etc/imapd.conf contains: abd> configdirectory: /var/imap abd> partition-default: /var/spool/imap abd> admins: cyrus abd> sasl_pwcheck_method: pwcheck abd> My configure for SASL: abd> ./configure --with-pwcheck=/var/pwcheck --enable-login \ abd> --without-saslauthd --enable-plain --disable-krb4 --disable-gssapi abd> and for Cyrus-Imap: abd> ./configure --with-dbdir=/opt/sfw --with-sasl=/usr/local/lib --includedir=/ abd> usr/local/include \ abd> --with-cyrus-user=cyrus \ abd> --with-cyrus-group=mail \ abd> --with-auth=unix abd> ¿Any idea? abd> Ana. -- Amos
Re: INBOX of shared account
> On Wed, 6 Mar 2002 16:28:53 +0100, > Christoph Krempe <[EMAIL PROTECTED]> (ck) writes: ck> I use Communicator 4.5. I see "user.teser", but not the mails in it ... I think that's too old. I think you'll need at least 4.76 to access another user's inbox. Maybe try 4.79. -- Amos
Re: INBOX of shared account
> On Wed, 6 Mar 2002 09:26:52 +0100, > Christoph Krempe <[EMAIL PROTECTED]> (ck) writes: ck> Hi, ck> I set up a cyurs-account "tester" which I want to share with the group ck> users. Therefore I did a "sam" on user.tester with the following result: ck> group:users lrswipcda ck> tester lrswipcda ck> cyrus lrswipcda Oh, I would recommend dropping the 'c' flag for group:users at least. Otherwise before long somebody will accidentally delete this folder and then mail will bounce. Been there, done that. ;-) -- Amos
Re: INBOX of shared account
> On Wed, 6 Mar 2002 14:33:54 +0100, > Christoph Krempe <[EMAIL PROTECTED]> (ck) writes: ck> That's right, but I can't see any mails in the INBOX of "tester", only=20 ck> mails in it's subsolders, allthough there a mailfiles in=20 ck> /var/spool/imap/user/tester ck> (I use Netscape Communicator a Client). What version of Communicator? I seem to remember experiencing that with older versions of Communicator. Note, "INBOX" is sorta like an alias just for your inbox. Though, as Cillian indicated, you should be able to see "user.tester". -- Amos
Playing with SNMP
Okay, now I'm playing with the SNMP reporting. Doing this with ucd-snmp-4.2.1.pre1 (yeah, I noticed pre2 is out, but if I restart the snmpd, I have to restart cyrus.) I'm now trying to figure out some of the graphing. Under the daily charts for "imap" there are two graphs, "Current" and "Total". The heading over the "Current" graph says "imap in use". Okay, that makes sense. I can see there are that number of imapd running on the box at that instant. The "Total" graph has the heading "imap connections". I presume that means "Total imap connections", as in an accumulative running total? However, the number seems a bit off for that. Or maybe I'm just not clear as to what that means. This is with a CVS snapshot from 20020216. I am using the DRAC patch if that matters. As to the predicament of snmp logging getting hosed if snmpd is restarted, any thoughts on re-initializing the connection upon master receiving a HUP signal? -- Amos
Re: pros over courier
> On Tue, 5 Mar 2002 08:13:37 -0800 (PST), > Dave C (dc) writes: dc> Hello, dc> I was wondering if people could give me some pros/cons dc> with courier and cyrus. I'm currently testing both. dc> The package must use ldap auth, do the maildir format dc> and also be setup for tls. any thoughts or suggestions dc> would be welcome Cyrus doesn't do maildir, so that in itself my conclude your query. While is it similar to maildir, it is not the same. -- Amos
Re: RECENT/SEEN flags
> On Wed, 27 Feb 2002 09:39:42 -0700, > Alec H Peterson <[EMAIL PROTECTED]> (ahp) writes: ahp> OK, so I've been poking my way through the code a little more, and ahp> found the database refresh routine (lib/cyrusdb_flat.c, ahp> map_refresh()). Is there some reason why the file size is the ahp> conditional for refreshing the database instead of the actual file ahp> modification time? It seems to me that would address this problem ahp> rather nicely. I wonder if this issue would resolve itself if skiplist was used as the backend for seen.db instead of flat. -- Amos
Re: daemonize
> On Thu, 28 Feb 2002 02:19:08 +0900, > Hajimu UMEMOTO <[EMAIL PROTECTED]> (hu) writes: hu> Hi, hu> The current master doesn't become daemon. The master should run in hu> background and detach controling terminal. Following patch also fixes hu> compilation problem under FreeBSD (maybe *BSD). I would also suggest that master do a chdir to some appropriate place. Maybe configdirectory would be good? Might also help in tracking down any dropped core files? -- Amos
SNMP (on Solaris)?
This seems to come up from time to time, so it's my turn. Is anybody using the SNMP support in Cyrus on Solaris (8)? Care to share what you did to get things going? One thing that puzzles me is the configure.in test for SNMP: #include "confdefs.h" /* Override any gcc2 internal prototype to avoid an error. */ /* We use char because int might match the return type of a gcc2 builtin and then its argument prototype would still apply. */ char sprint_objid(); int main() { sprint_objid() ; return 0; } Now why is this particular function used? It looks like it expects 3 arguments (yeah, I'm not too familiar with SNMP--yet). Wouldn't it be better to have something like this for the test? void init_mib(void); void init_mib_internals(void); int main() { init_mib(); init_mib_internals(); return 0; } -- Amos
Re: sieveshell authentication failed on Solaris
> On Tue, 26 Feb 2002 16:59:28 +0100, > Simon Josefsson <[EMAIL PROTECTED]> (sj) writes: sj> I was wondering why all copies from Simon Matter on this list ended up sj> in my junk folder instead of my info-cyrus folder, and it seems to be sj> because S.M.'s messages contain invalid RFC 2822 header lines sj> (">Received:"), thereby presumably indicating the end of the RFC 2822 sj> header and not allow the Sieve parser to reach the Sender: header I sj> match on. I *think* this was fixed in CVS, right? Ken will have to confirm. -- Amos
Re: sieve vacation not working
> On Fri, 22 Feb 2002 17:48:35 -0500, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> The "unspecified-domain" is almost certainly being added by lmtpd/sieve. If that's the case, then I bet the ".google.com" is getting tacked on by Postfix because it don't care for unqualified addresses. What a weird afternoon. Oh, is "postmaster" set in imapd.conf? Could it maybe be "cyrus" in this case? Could that be a factor? -- Amos
Re: OT: Sieve woes#2, but closer maybe? --- file closed!
> On Fri, 22 Feb 2002 22:15:59 +0100, > marc olejak <[EMAIL PROTECTED]> (mo) writes: mo> thx again... FYI? (F*** YOU IDIOT?)... as a german, i am not so deep mo> within your shortcuts... sorry... For Your Information. Perhaps this might help: http://www.ucc.ie/info/net/acronyms/acro.html -- Amos
Re: sieve vacation not working
> On Fri, 22 Feb 2002 13:18:33 -0800, > Ian Macdonald <[EMAIL PROTECTED]> (im) writes: im> Well, it's not being sent to the MAIL FROM address either, I'm im> afraid. The MAIL FROM is a fully qualified on-site address that I im> receive mail at, but instead the [EMAIL PROTECTED] im> address is used. I'm folloiwng this thread (trying to) and I'm afraid I'm totally baffled. I'm not seeing this with our Postfix/Cyrus setup. -- Amos
Re: howto use the skiplist backend
> On Fri, 22 Feb 2002 13:32:32 +0100, > Simon Matter <[EMAIL PROTECTED]> (sm) writes: sm> How do I compile and use the new skiplist backend? I want to try it but sm> I don't know how. Try: LOCALDEFS="-DCONFIG_DB_MBOX=\"(&cyrusdb_skiplist)\""; export LOCALDEFS (adapt accordingly to the shell you're using.) Then run the configure script for 2.1.2. Or grab Cyrus from CVS. The SMakefile shell script will do this for you. If you grab it from CVS, looks like you can try using seen db as skiplist as well. I *think* the conversion process is still: 1. Stop cyrus. 2. Become user "cyrus". 3. Dump out current mailboxes.db: ctl_mboxlist -d > /var/tmp/mboxes.txt mv /var/imap/mailboxes.db /var/tmp 4. Update your cyrus binaries. 5. Become user "cyrus". 6. Upload db: ctl_mboxlist -u < /var/tmp/mboxes.txt I think I've noticed something that looks like it might be a conversion thing or another. I dunno. Yes, I've played with skiplist too. It does indeed look way cool. Though, I'm still just a tad nervous about using it in production right at this moment. I think it is still evolving. -- Amos
Re: cyrus and db4 gives me DBERROR db3: 7 lockers
> On Fri, 22 Feb 2002 10:09:37 +0100, > Simon Matter <[EMAIL PROTECTED]> (sm) writes: sm> I have upgraded to db4 today, db3 libs are still there for other sm> applications. I have then recompiled my cyrus-imapd RPM version 2.1.2. sm> Feb 22 10:06:46 dhcp-141-104 pop3d[2135]: DBERROR db3: 8 lockers sm> Feb 22 10:06:46 dhcp-141-104 pop3d[2136]: DBERROR db3: 9 lockers sm> Feb 22 10:06:46 dhcp-141-104 pop3d[2139]: DBERROR db3: 10 lockers sm> Feb 22 10:07:13 dhcp-141-104 imapd[2149]: DBERROR db3: 14 lockers To some extent these are normal for BDB. I'm not sure why Sleepycat chose to log these at error level. I've also never been able to figure out exactly what trips this message to be logged. Very roughly there should be somewhere around as many lockers as Cyrus processes. This will vary somewhat depending on whether you're using SSL/TLS or not. Here's my understanding of things. The lmtpd process will typically have two read locks: one for deliver.db and another for mailboxes.db. Other Cyrus processes can have: one for mailboxes.db and another for tls_sessions.db, depending on whether SSL/TLS is being used. Then any of these processes can open write locks (via cursor) to update these db files. You can check the current status of lockers by using the db_stat command. First become the "cyrus" user, then cd to the "db" directory where the "__db.*" files are at, typically "/var/imap/db". Then you can run db_stat with the -c option: $ /usr/local/BerkeleyDB.4.0/bin/db_stat -c 18898 Last allocated locker ID. 9 Number of lock modes. 5 Maximum number of locks possible. 5 Maximum number of lockers possible. 5 Maximum number of objects possible. 216MCurrent locks. 4294M Maximum number of locks so far. 2147Current number of lockers. 2152Maximum number lockers so far. 0 Current number lock objects. 16 Maximum number of lock objects so far. 274MNumber of lock requests. 274MNumber of lock releases. 0 Number of lock requests that would have waited. 1188Number of lock conflicts. 2 Number of deadlocks. 0 Number of transaction timeouts. 0 Number of lock timeouts. 19MB 624KB Lock region size (20561920 bytes). 267MThe number of region locks granted after waiting. 2271M The number of region locks granted without waiting. You might see a rise in lockers if you experience a lot of Cyrus processes that crash unexpectedly. Search your logs for the string 'exited, signaled'. Now, 2.1.2 goes a long way to improve the run-away lockers I was seeing previously. However, I *think* there still might be a tiny leak somewhere. I'm not sure if it is something stupid I'm doing, a leak elsewhere in the code, or something crashing my Cyrus processes. One thing I have observed is that someone using Mozilla periodically seems to generate an 'exited, signaled to death by 10' when using SSL with POP. What's odd is that it doesn't seem to happen all the time, just occasionally. Though, I'm not entirely sure that alone is the cause of the rising lockers since the number of lockers seems to be growing faster than his 'death by 10' incidents. Still looking into it. -- Amos
Re: duplicate.c:138: too many arguments to function
> On Wed, 20 Feb 2002 17:28:10 -0800, > lucas <[EMAIL PROTECTED]> (l) writes: l> This is starting to be a long story. Anyway, I downloaded l> cyrus-imapd and learned that I needed Berkeley DB 3.x or above, l> so I got l> Berkeley DB 4.0.14 and installed it. I guess the libraries got l> installed to the wrong directory, because my configure line l> needed to know their l> location: l> env CPPFLAGS="-I/usr/local/BerkeleyDB.4.0/include" \ l> LDFLAGS="-L/usr/local/BerkeleyDB.4.0/lib" ./configure Try ./configure --help. I think there are some options for specifying BDB stuff l> Stop in /usr/local/src/cyrus-imapd-2.0.16/lib. This version does not support BDB 4.0.14. For that you need 2.1.2. -- Amos
Re: [Fwd: Vacation.. yes again..]
> On Tue, 19 Feb 2002 11:25:45 -0600, > Tyrone Vaughn <[EMAIL PROTECTED]> (tv) writes: tv> -- example sieve vacation filter -- tv> require ["fileinto"]; tv> require ["reject"]; tv> require ["vacation"]; tv> # %VACATION_START% tv> vacation tv> :days 1 tv> :subject "Out to lunch" tv> # %VAC_TEXT_START% tv> "Always"; tv> # %VAC_TEXT_END% tv> # %VACATION_END% tv> --- I can't answer why it might have quit working, but is the recipient address in the header consistent? I mean, if you're sending the message to <[EMAIL PROTECTED]>, perhaps you need: :addresses "[EMAIL PROTECTED]" -- Amos
Re: Skiplist in 2.0? (was Re: Cyrus IMAP 2.1.2 released)
> On Wed, 20 Feb 2002 11:29:24 +1100, > Jeremy Howard <[EMAIL PROTECTED]> (jh) writes: jh> We've been getting bitten lots by the BDB deadlock problem and would like to jh> switch to skiplist ASAP. However we're not yet quite ready to upgrade to jh> 2.1. I can report that we're not so stressed over DB now that we're using 2.1.2. If you're seeing runaway lockers, you might try the upgrade. -- Amos
Re: sieveshell not working with saslauthd
> On Mon, 11 Feb 2002 10:13:38 -0500, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> The problem is that timsieved is advertising DIGEST and CRAM, which km> sieveshell will try to use before any plaintext mechanism. An --auth km> option needs to be added to sieveshell (like cyradm) to get around km> this. I plan on looking into this. Uh oh, option collision. Maybe --mech would be less confusing, at least to my poor mind. ;-) SYNOPSIS $ cyradm [--user user] [--[no]rc] [--systemrc file] [--userrc file] \ > [--port n] [--auth mechanism] [--server] server SYNOPSIS sieveshell [--user=user] [--authname=authname] [-- realm=realm] [--exec=script] server -- Amos
Re: How can a program securely get new/unread msg status on lots of Cyrus mailboxes?
> On Fri, 15 Feb 2002 09:41:27 -0500, > Ken Murchison <[EMAIL PROTECTED]> (km) writes: km> As Cillian stated, you can proxy as the user, so doing this in perl or km> some other scripting language is fairly easy. Alternatively, if you km> want to do this outside of the IMAP protocol take a look at fud.c, which km> is a finger-type service that tells you the number of unread messages in km> the INBOX and the last time the users read his/her INBOX. This should km> be a good starting point for writing your own utility. Is there any reason why this fud client can't be put into contrib? -- Amos
Re: imap and folders
> On Fri, 08 Feb 2002 16:03:42 -0500, > Joe Ellis <[EMAIL PROTECTED]> (je) writes: je> isn't what they really have is user/userA/INBOX No. "INBOX" is an alias, if you will, for "user/userA". je> I have several folders that are in the "root" directory tree for my je> user. not for everyone else. for example, I have folder names as je> follows: je> user/joee/ je> user/joee/Cyrus je> user/joee/Cyrus/info je> user/joee/Cyrus/sasl je> (Now I'm logged into cyradm as the admin when I list these mailboxes) As Ken has pointed out before, even if altnamespace is set to yes, the admin, usually the "cyrus" user, will see exactly how the folders exist on the server. je> When I create a new folder, I create it in "root". Its root to the je> user. When I log into cyradm as the user and list the mailboxes, I je> get: je> Cyrus je> Cyrus/info je> Cyrus/sasl je> INBOX (\Noinferiors) je> So can't the user create sub folders in INBOX or at the "root" level je> to the user. je> *I have altnamespace: yes in /etc/imapd.conf. Thats why _I_ can't je> create folders in INBOX* Correct. This is something Ken has also pointed out before. -- Amos
Re: imap and folders
> On 08 Feb 2002 09:24:17 -0800, > julesa <[EMAIL PROTECTED]> (j) writes: j> However, as Ken pointed out, using the altnamespace option in Cyrus j> 2.1.1 will cause the server to report subdirectories of the INBOX as j> existing at the same level as the INBOX. So, the structure above would j> still be the same on disk, but mail clients would see: j> /INBOX j> /archive j> The only limitation with this approach is that now the INBOX never j> appears to have any subfolders. So far our experience has been that most clients tend to find this altnamespace easier to deal with. It also seems to require less of a learning curve to the user. -- Amos
Re: Vacation.. yes again..
> On Tue, 12 Feb 2002 12:43:23 -0600, > Tyrone Vaughn <[EMAIL PROTECTED]> (tv) writes: tv> -- example sieve vacation filter -- tv> require ["fileinto"]; tv> require ["reject"]; tv> require ["vacation"]; Can you do that? I thought there could only be one require line? Hmm... it's not in the RFC, so maybe not. Wonder why I thought that -- Amos
Re: sharing containers by group with cyrus
> On Tue, 12 Feb 2002 10:37:20 -0500, > Theodore J Knab <[EMAIL PROTECTED]> (tjk) writes: tjk> I was wondering if there was a way to share mailboxes by group with tjk> Cyrus. If you're using UNIX auth, I believe you can specify the ACL item as "group:XXX" where "XXX" is a UNIX group. -- Amos