Re: Cyrus IMAP 2.0.13 released
God I love these guys! Thanks Larry! Being part of the Cyrus crowd is great. I was just thinking about the fact that I hadn't seen any announcements about 2.0.13 just yesterday and was wondering if the coding was still coming along. -- Michael -- - Original Message - From: Lawrence Greenfield [EMAIL PROTECTED] To: info-cyrus [EMAIL PROTECTED] Sent: Thursday, April 26, 2001 3:51 PM Subject: Cyrus IMAP 2.0.13 released I've placed Cyrus IMAP 2.0.13 on ftp.andrew. This version adds a couple of new features and some slight code reorganization. We'd love to have people try it out. Of especial interest may be the partial virtual domain support via the multiple config files (I'd love a contributed doc/virtual-domains.html) and support for a single imapd process to service multiple clients in succession, cutting down the number of forks. This approach will be extended to lmtpd and pop3d in the future. Finally, the master process now accepts the maxchild service modifier to limit how many instances of a particular service may be active at one time. Cyrus IMAP 2.0.13 is available at ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/cyrus-imapd-2.0.13.tar.gz Have fun, Larry Greenfield
Re: Patch to change dots to slashes.
This is a modification of my earlier patch to enable changing the mailbox delimiter to a different character (eg. a slash), which was based almost entirely on David Fuchs' original patch to change the delimiter to a slash. This patch fixes a problem with both earlier patches which broke setting quotas on sub-folders. The actual changes from the previous patch are those in mailbox.c:mailbox_hash_quota(), but I'm including the full patch here for completeness, along with my original note. Cheers, Chris, OnTheNet This is a patch to enable changing the mailbox delimiter to an different character, such as a slash. This patch is mostly a modified version of David Fuchs' original patch to change the dot delimiter to a slash. The major differences are: o You can change the delimiter to any character (within reason!) by changing two #defines in imap/mailbox.h: #define DELIMITER_CHAR '/' #define DELIMITER_STR / o It adds some sanity checking for mailbox names. In particular it explicity disallows mailbox names of cyrus.cache, cyrus.index etc., and explicitly disallows patterns like './' (which includes '../'). See badmboxpatterns in imap/mboxname.c. o There are one or two additional places over David's patch where I think the delimiter needs to be changed. Note: this patch has not yet seen extensive use in a high traffic environment, so... Buyer Beware! One thing it might be nice to see, but I haven't had the time to implement, is setting the delimiter character using the configure script. Cheers, Chris, OnTheNet diff -u -r cyrus-imapd-2.0.12/imap/acapmbox.c cyrus-imapd-2.0.12/imap/acapmbox.c --- cyrus-imapd-2.0.12/imap/acapmbox.c Sat Feb 17 05:54:43 2001 +++ cyrus-imapd-2.0.12/imap/acapmbox.c Thu Apr 12 17:26:16 2001 @@ -93,12 +93,12 @@ const char *server, const char *name) { if (!server) server = config_servername; -if (!strncmp(name, user., 5)) { +if (!strncmp(name, user DELIMITER_STR, 5)) { char *p; /* user+detail */ strcpy(postaddr, name + 5); - p = strchr(postaddr, '.'); + p = strchr(postaddr, DELIMITER_CHAR); if (p) *p = '+'; strcat(postaddr, @); strcat(postaddr, server); @@ -243,9 +243,11 @@ /* needs to convert from mUTF7 to UTF-8 */ snprintf(ret, MAX_MAILBOX_PATH, %s/%s, global_dataset, mailbox_name); +#if DELIMITER_CHAR != '/' for (j = strlen(global_dataset); ret[j] != '\0'; j++) { - if (ret[j] == '.') ret[j] = '/'; + if (ret[j] == DELIMITER_CHAR) ret[j] = '/'; } +#endif return 0; } @@ -266,9 +268,11 @@ } strcpy(ret, entryname + gdlen + 1); +#if DELIMITER_CHAR != '/' for (j = 0; ret[j] != '\0'; j++) { - if (ret[j] == '/') ret[j] = '.'; + if (ret[j] == '/') ret[j] = DELIMITER_CHAR; } +#endif return 0; } diff -u -r cyrus-imapd-2.0.12/imap/arbitron.c cyrus-imapd-2.0.12/imap/arbitron.c --- cyrus-imapd-2.0.12/imap/arbitron.c Wed May 24 06:52:12 2000 +++ cyrus-imapd-2.0.12/imap/arbitron.c Thu Apr 12 13:40:53 2001 @@ -162,10 +162,10 @@ int useridlen = tab - line; /* Don't report users reading their own private mailboxes */ -if (!strncasecmp(arbitronargs-name, user., 5) - !memchr(line, '.', useridlen) +if (!strncasecmp(arbitronargs-name, user DELIMITER_STR, 5) + !memchr(line, DELIMITER_CHAR, useridlen) !strncasecmp(arbitronargs-name+5, line, useridlen) - (arbitronargs-name[5+useridlen] == '.' || + (arbitronargs-name[5+useridlen] == DELIMITER_CHAR || arbitronargs-name[5+useridlen] == '\0')) { return 0; } diff -u -r cyrus-imapd-2.0.12/imap/duplicate.c cyrus-imapd-2.0.12/imap/duplicate.c --- cyrus-imapd-2.0.12/imap/duplicate.c Sat Jan 27 13:05:52 2001 +++ cyrus-imapd-2.0.12/imap/duplicate.c Thu Apr 12 15:08:49 2001 @@ -75,6 +75,7 @@ #include imapconf.h #include exitcodes.h #include xmalloc.h +#include mailbox.h #include duplicate.h @@ -160,11 +161,11 @@ char *idx; char c; -idx = strchr(mbox,'.'); /* skip past user. */ -if (idx == NULL) { /* no '.' so just use mbox */ +idx = strchr(mbox, DELIMITER_CHAR); /* skip past user. */ +if (idx == NULL) { /* no DELIMITER so just use mbox */ idx = mbox; } else { - idx++; /* skip past '.' */ + idx++; /* skip past DELIMITER */ } c = (char) tolower((int) *idx); if (!islower((int) c)) { diff -u -r cyrus-imapd-2.0.12/imap/idle_idled.c cyrus-imapd-2.0.12/imap/idle_idled.c --- cyrus-imapd-2.0.12/imap/idle_idled.cSat Jan 20 03:48:58 2001 +++ cyrus-imapd-2.0.12/imap/idle_idled.cThu Apr 12 13:42:07 2001 @@ -190,7 +190,7 @@ /* fill the structure */ idledata.msg = msg; idledata.pid = getpid(); -strcpy(idledata.mboxname, mailbox ? mailbox-name : .); +strcpy(idledata.mboxname, mailbox ?
deliver bus error core dump
Hello, I'm running Cyrus IMAP4 v1.6.24 for several months under solaris and since yesterday deliver -E 2 doesn't work any more /usr/local/imap/imapd/bin/deliver -E 2 Bus Error - core dumped I've tried to analyse why and made a truss of /usr/local/imap/imapd/bin/deliver -E 2 I've got this: stat(/export/bases/imap/imapd/bin/deliver, 0xFFBEFB00) = 0 [stuff deleted] open(/var/imap/mailboxes, O_RDWR) = 3 fstat(3, 0xFFBEFB48)= 0 mmap(0x, 14789, PROT_READ, MAP_SHARED, 3, 0) = 0xFF25 umask(077) = 077 brk(0x001538E8) = 0 brk(0x001578E8) = 0 getuid()= 6131 [6131] time() = 988358704 fstat(-1, 0xFFBEDFF8) Err#9 EBADF open(/dev/conslog, O_WRONLY) = 4 fcntl(4, F_SETFD, 0x0001) = 0 fstat(4, 0xFFBEDFF8)= 0 fstat(4, 0xFFBEEA58)= 0 time() = 988358704 open(/usr/share/lib/zoneinfo/MET, O_RDONLY) = 5 read(5, \0\0\0\0\0\0\0\0\0\0\0\0.., 8192) = 758 close(5)= 0 getpid()= 20945 [20944] putmsg(4, 0xFFBEE110, 0xFFBEE104, 0)= 0 open(/etc/.syslog_door, O_RDONLY) = 5 door_info(5, 0xFFBEE048)= 0 getpid()= 20945 [20944] door_call(5, 0xFFBEE030)= 0 close(5)= 0 open(/var/imap/deliverdb/deliver-a.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08) = 0 open(/var/imap/deliverdb/deliver-q.lock, O_RDWR|O_CREAT, 0666) = 6 fcntl(6, F_SETLKW, 0xFFBECAF0) = 0 brk(0x001578E8) = 0 brk(0x001598E8) = 0 open64(/var/imap/deliverdb/deliver-q.db.pag, O_RDWR|O_CREAT, 0666) = 7 open64(/var/imap/deliverdb/deliver-q.db.dir, O_RDWR|O_CREAT, 0666) = 8 fstat64(8, 0xFFBEDB00) = 0 llseek(7, 0, SEEK_SET) = 0 read(7, \01203D203CE03A0039C03 n.., 1024) = 1024 close(6)= 0 Incurred fault #5, FLTACCESS %pc = 0x00025634 siginfo: SIGBUS BUS_ADRALN addr=0x00156E49 Received signal #10, SIGBUS [default] siginfo: SIGBUS BUS_ADRALN addr=0x00156E49 *** process killed *** It seems to stop when processing /var/imap/deliverdb/deliver-q.db.* It's maybe a coincidence but I've got only one user whose mailbox begins with q and I've created it yesterday Why does it processes deliver-q just after processing deliver-a ? I'd have expected first a, then b, the c .. If I remove /var/imap/deliverdb/deliver-q.db.* rm /var/imap/deliverdb/deliver-q.db.* deliver seems to work again but if I truss it, I notice that it opens /var/imap/deliverdb/deliver-q.lock very often open(/var/imap/deliverdb/deliver-a.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08) = 0 open(/var/imap/deliverdb/deliver-q.lock, O_RDWR|O_CREAT, 0666) = 6 fcntl(6, F_SETLKW, 0xFFBECAF0) = 0 brk(0x001578E8) = 0 brk(0x001598E8) = 0 open64(/var/imap/deliverdb/deliver-q.db.pag, O_RDWR|O_CREAT, 0666) = 7 open64(/var/imap/deliverdb/deliver-q.db.dir, O_RDWR|O_CREAT, 0666) = 8 fstat64(8, 0xFFBEDB00) = 0 llseek(7, 0, SEEK_SET) = 0 read(7, 0x00156B80, 1024) = 0 close(6)= 0 close(8)= 0 close(7)= 0 close(5)= 0 open(/var/imap/deliverdb/deliver-b.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08) = 0 open(/var/imap/deliverdb/deliver-q.lock, O_RDWR|O_CREAT, 0666) = 6 fcntl(6, F_SETLKW, 0xFFBECAF0) = 0 close(6)= 0 close(8)Err#9 EBADF close(7)Err#9 EBADF close(5)= 0 open(/var/imap/deliverdb/deliver-c.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08) = 0 open(/var/imap/deliverdb/deliver-q.lock, O_RDWR|O_CREAT, 0666) = 6 fcntl(6, F_SETLKW, 0xFFBECAF0) = 0 close(6)= 0 close(8)Err#9 EBADF close(7)Err#9 EBADF close(5)= 0 open(/var/imap/deliverdb/deliver-d.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08)
Re: deliver bus error core dump
I think you've hit the bug I mailed in on 19th April (Subject: deliver -E (Solaris, 1.6.24 without DB). In part it read... --- On Solaris platforms which don't have DB available, deliver -E fails to prune anything, ever. The problem starts in deliver.c, in the _prune_actual_db routine. In it, there's a call to... checkdelivered(, 0, , 0); (only called when DB not available) Within checkdelivered, a call is then made to _get_db_name which will then end up with it (_get_db_name) always returning .../deliverdb/deliver-q to checkdelivered (which in itself is disastrous if you actually have real entries in the q database). --- Note the bit which in itself is disastrous if you actually have real entries in the q database. I'm afraid that I can't offer you any solution. Cheers... On Fri, 27 Apr 2001 11:22:10 +0200 [EMAIL PROTECTED] wrote: Hello, I'm running Cyrus IMAP4 v1.6.24 for several months under solaris and since yesterday deliver -E 2 doesn't work any more /usr/local/imap/imapd/bin/deliver -E 2 Bus Error - core dumped I've tried to analyse why and made a truss of /usr/local/imap/imapd/bin/deliver -E 2 I've got this: stat(/export/bases/imap/imapd/bin/deliver, 0xFFBEFB00) = 0 [stuff deleted] open(/var/imap/mailboxes, O_RDWR) = 3 fstat(3, 0xFFBEFB48)= 0 mmap(0x, 14789, PROT_READ, MAP_SHARED, 3, 0) = 0xFF25 umask(077) = 077 brk(0x001538E8) = 0 brk(0x001578E8) = 0 getuid()= 6131 [6131] time() = 988358704 fstat(-1, 0xFFBEDFF8) Err#9 EBADF open(/dev/conslog, O_WRONLY) = 4 fcntl(4, F_SETFD, 0x0001) = 0 fstat(4, 0xFFBEDFF8)= 0 fstat(4, 0xFFBEEA58)= 0 time() = 988358704 open(/usr/share/lib/zoneinfo/MET, O_RDONLY) = 5 read(5, \0\0\0\0\0\0\0\0\0\0\0\0.., 8192) = 758 close(5)= 0 getpid()= 20945 [20944] putmsg(4, 0xFFBEE110, 0xFFBEE104, 0)= 0 open(/etc/.syslog_door, O_RDONLY) = 5 door_info(5, 0xFFBEE048)= 0 getpid()= 20945 [20944] door_call(5, 0xFFBEE030)= 0 close(5)= 0 open(/var/imap/deliverdb/deliver-a.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08) = 0 open(/var/imap/deliverdb/deliver-q.lock, O_RDWR|O_CREAT, 0666) = 6 fcntl(6, F_SETLKW, 0xFFBECAF0) = 0 brk(0x001578E8) = 0 brk(0x001598E8) = 0 open64(/var/imap/deliverdb/deliver-q.db.pag, O_RDWR|O_CREAT, 0666) = 7 open64(/var/imap/deliverdb/deliver-q.db.dir, O_RDWR|O_CREAT, 0666) = 8 fstat64(8, 0xFFBEDB00) = 0 llseek(7, 0, SEEK_SET) = 0 read(7, \01203D203CE03A0039C03 n.., 1024) = 1024 close(6)= 0 Incurred fault #5, FLTACCESS %pc = 0x00025634 siginfo: SIGBUS BUS_ADRALN addr=0x00156E49 Received signal #10, SIGBUS [default] siginfo: SIGBUS BUS_ADRALN addr=0x00156E49 *** process killed *** It seems to stop when processing /var/imap/deliverdb/deliver-q.db.* It's maybe a coincidence but I've got only one user whose mailbox begins with q and I've created it yesterday Why does it processes deliver-q just after processing deliver-a ? I'd have expected first a, then b, the c .. If I remove /var/imap/deliverdb/deliver-q.db.* rm /var/imap/deliverdb/deliver-q.db.* deliver seems to work again but if I truss it, I notice that it opens /var/imap/deliverdb/deliver-q.lock very often open(/var/imap/deliverdb/deliver-a.lock, O_RDWR|O_CREAT, 0666) = 5 fcntl(5, F_SETLKW, 0xFFBEDC08) = 0 open(/var/imap/deliverdb/deliver-q.lock, O_RDWR|O_CREAT, 0666) = 6 fcntl(6, F_SETLKW, 0xFFBECAF0) = 0 brk(0x001578E8) = 0 brk(0x001598E8) = 0 open64(/var/imap/deliverdb/deliver-q.db.pag, O_RDWR|O_CREAT, 0666) = 7 open64(/var/imap/deliverdb/deliver-q.db.dir, O_RDWR|O_CREAT, 0666) = 8 fstat64(8, 0xFFBEDB00) = 0 llseek(7, 0, SEEK_SET) = 0 read(7, 0x00156B80, 1024) = 0 close(6)= 0 close(8)= 0 close(7)= 0 close(5)= 0 open(/var/imap/deliverdb/deliver-b.lock, O_RDWR|O_CREAT, 0666) = 5
Re: Cyrus IMAP 2.0.13 released
Hi, this is greate that the development of the server is moving along. I hope that you guys at cmu also watch to mailing list, and have seen the 'forking problem' that ppl here have been describing. We will try to test the newest release and will report ASAP if the problem persists or not. Thank you Helmut Apfelholz --- Lawrence Greenfield [EMAIL PROTECTED] wrote: I've placed Cyrus IMAP 2.0.13 on ftp.andrew. This version adds a couple of new features and some slight code reorganization. We'd love to have people try it out. Of especial interest may be the partial virtual domain support via the multiple config files (I'd love a contributed doc/virtual-domains.html) and support for a single imapd process to service multiple clients in succession, cutting down the number of forks. This approach will be extended to lmtpd and pop3d in the future. Finally, the master process now accepts the maxchild service modifier to limit how many instances of a particular service may be active at one time. Cyrus IMAP 2.0.13 is available at ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/cyrus-imapd-2.0.13.tar.gz Have fun, Larry Greenfield __ Do You Yahoo!? Yahoo! Auctions - buy the things you want at great prices http://auctions.yahoo.com/
Re: deliver bus error core dump
[EMAIL PROTECTED] wrote: I think you've hit the bug I mailed in on 19th April (Subject: deliver -E (Solaris, 1.6.24 without DB). In part it read... Thanks for your answer. You're right, it seems to be the same bug Does anyone know how to correct this ? Is it corrected under 2.0.12 ? Regards, Nicolas PS: (More info) the cyrus binaries are statically built with Berkeley DB 3.1.17
Re: Cyrus IMAP 2.0.13 released
I was wondering if perhaps the maxchild parameter was inspired, at least in part, to keep a check on processes like lmtpd from getting so numerous that the possibility of contention increases. Though, that's just a wild guess On Fri, 27 Apr 2001 03:55:41 -0700 (PDT), Helmut Apfelholz [EMAIL PROTECTED] (ha) writes: ha Hi, ha this is greate that the development of the server is ha moving along. I hope that you guys at cmu also watch ha to ha mailing list, and have seen the 'forking problem' that ha ppl here have been describing. We will try to test the ha newest release and will report ASAP if the problem ha persists or not. ha Thank you ha Helmut Apfelholz
Re: Virtual Domain Implementation (example)
Marius, We use standard DNS, not DNS entries in LDAP or anything like that. I assume (I could be wrong - someone help me out...) that your LDAP directory structure would be completely independent of your DNS registrations. If you had 2 domains registered to your organization, you could have 2 Cyrus/Postfix combos installed on 1 physical server, where your LDAP tree that looks something like this (LDAP can be running on any machine...): o=LDAPadmin (or whatever) ou=MailDomains (or whatever) ou=Acme miller smith ou=Rutgers miller smith Now, lets assume you have email.acme.com registered for the Acme domain and email.rutgers.edu registered for the Rutgers domain. In your DNS data, all mail for acme.com will be handled by email.acme.com and all mail for rutgers.edu will be handled by email.rutgers.edu As long as Postfix or Sendmail is configured to accept mail for their respective domains, if you send to [EMAIL PROTECTED], the message should get dropped to [EMAIL PROTECTED] At this point, the only purpose your LDAP directory serves is for account authentication (storage container just for usernames and passwords...). You can make the structure look anyway you want. You don't not even need to reference the domains at all - as long as you know the users for domain A are at this basedn, and users for domain B are at this basedn, just configure each imapd.conf's sasl configuration to point to that location. In my example, all I am using LDAP for is username password. The mail server or the ldap server does not care whether or not you are using domains or subdomains. -John Tegomoh, Marius N. wrote: John, Thanks for the further clarification. It certainly helps. Unfortunately, my original query wasn't clear enough. Your directory service is for Rutgers.edu and everything else is a subdomain of that. What if you wanted your directory service to hold information for Acme.com as well as Rutgers.edu? I'm hoping for an example multiple unrelated domains in the same directory server (I suppose that's we need to define multiple BaseDNs). Marius -Original Message- From: John C. Amodeo [mailto:[EMAIL PROTECTED]] Sent: Thursday, April 26, 2001 5:20 PM To: Tegomoh, Marius N. Cc: [EMAIL PROTECTED] Subject: Re: Virtual Domain Implementation (example) Marius, Actually, we are not using OpenLdap - we are using Novell NDS, which exports all its information in LDAP compliant form. We translate each context in the NDS tree into a separate Virtual Server or department - for example: o=Rutgers.edu ou=Units ou=Economics miller smith ou=Biology miller smith So, one Cyrus server's Sasl config would point to ou=Economics,ou=Units,o=Rutgers.edu and the other would point to ou=Biology,ou=Units,o=Rutgers.edu These LDAP entries go in the imapd.conf files for each Cyrus server (1 Cyrus server can point to one LDAP server OR different BaseDN's in an LDAP tree...) As far as logging 2 users from different departments in to their mail server is pretty easy... Cyrus server 1, whose sasl config points to ou=Economics,ou=Units,o=Rutgers.edu is running on an IP address that is registered as email.economics.rutgers.edu and Cyrus server 2, whose sasl config points to ou=Biology,ou=Units,o=Rutgers.edu is running on an IP address that is registered as email.biology.rutgers.edu For each user (smith, for instance) on any given server, a login would look like: username smith imap server email.economics.rutgers.edu smtp server email.economics.rutgers.edu OR username smith imap server email.biology.rutgers.edu smtp server email.biology.rutgers.edu Each user will use just their username and the server for their department. What's nice about our setup is that since most of our users are on Windows (some things you can't change... ), when they log into Novell to get to the network the same account gets them their email. I assume you could apply this directory structure where each on of our departments would equal a domain. At this point, it would be up to your DNS servers and MX records to drop mail to the proper server. What we do is for a department like Economics, whose Novell server is economics.rutgers.edu, and whose Cyrus e-mail server is email.economics.rutgers.edu, the MX record for economics.rutgers.edu has its mail handling set to email.economics.rutgers.edu So, when you send to miller@economics it actually goes to [EMAIL PROTECTED] Hope this helps. -John -- __ John C. Amodeo, Associate Director Information Technology and Computer Operations Faculty of Arts Sciences, Rutgers University 732.932.9455-voice 732.932.0013-fax
RE: Virtual Domain Implementation (example)
John, Thanks. It makes sense now. It was an issue of my limited understanding of LDAP and how its trees are designed. If I need further help on this I'll bother you privately since this is no longer a Cyrus issue. Thanks for your help. Marius -Original Message- From: John C. Amodeo [mailto:[EMAIL PROTECTED]] Sent: Friday, April 27, 2001 8:59 AM To: Tegomoh, Marius N. Cc: [EMAIL PROTECTED] Subject: Re: Virtual Domain Implementation (example) Marius, We use standard DNS, not DNS entries in LDAP or anything like that. I assume (I could be wrong - someone help me out...) that your LDAP directory structure would be completely independent of your DNS registrations. If you had 2 domains registered to your organization, you could have 2 Cyrus/Postfix combos installed on 1 physical server, where your LDAP tree that looks something like this (LDAP can be running on any machine...): o=LDAPadmin (or whatever) ou=MailDomains (or whatever) ou=Acme miller smith ou=Rutgers miller smith Now, lets assume you have email.acme.com registered for the Acme domain and email.rutgers.edu registered for the Rutgers domain. In your DNS data, all mail for acme.com will be handled by email.acme.com and all mail for rutgers.edu will be handled by email.rutgers.edu As long as Postfix or Sendmail is configured to accept mail for their respective domains, if you send to [EMAIL PROTECTED], the message should get dropped to [EMAIL PROTECTED] At this point, the only purpose your LDAP directory serves is for account authentication (storage container just for usernames and passwords...). You can make the structure look anyway you want. You don't not even need to reference the domains at all - as long as you know the users for domain A are at this basedn, and users for domain B are at this basedn, just configure each imapd.conf's sasl configuration to point to that location. In my example, all I am using LDAP for is username password. The mail server or the ldap server does not care whether or not you are using domains or subdomains. -John
Re: cyrus configuration
On Tue, Apr 24, 2001 at 08:48:01AM +0200, Werner Reisberger wrote: as in Debian/Linux. You need to use imap 143/tcp i have same problem: telnet localhost 143 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. Connection closed by foreign host. in syslog: Apr 24 13:17:45 klfmanik service-imapd[10095]: executed Apr 24 13:17:45 klfmanik service-imapd[10095]: unable to set close on exec: Bad file descriptor Cyrus 2.x isn't supposed to be run under inetd (which is what this error means). You need to use the Cyrus master process instead. debian/woody - kernel 2.4.3 i have problem with /dev/random too -- 5o Peter.Mann at tuke.sk KLFMANiK ICQ 12491471 PM2185 RIPE -- Larry M. Rosenbaum [EMAIL PROTECTED] Bldg 4500-N, Room E-218 865 574-8155 phone PO Box 2008, MS 6271865 241-4000 fax Oak Ridge, TN 37831-6271 Oak Ridge National Laboratory, Network Computing Services group
Re: How to add virtual domain support
Michael Fair wrote: Yes, the Cyrus server supports realms though it is largely unused. Currently SASL just fills the realm info in with the hostname of the machine. You can see this when you do sasldblistusers. In regards to all clients supportnig SASL. Perhaps it's only a matter of time, but will they implement it according to the standards (specifically Microsoft). I have a suggestion on this subject. What about the possibility of binding a realm to a local address for cyrus (IP based vhost)? Yes, authentication and named vhosts via username and realm is ideal, but given that that information is usually not explicitly send by the client, if the imap server could assign the realm based on some implicit information such as the IP address, then there is an answer that should work while we all wait for more widespread support of SASL realms. If there was a patch to do this, would it be accepted into CVS? -- Todd Nemanich [EMAIL PROTECTED] Protecting the opulent and staging moral standard, They expect redemption of character and self growth Bad Religion - Inner Logic
Re: Cyrus IMAP 2.0.13 released
I've placed Cyrus IMAP 2.0.13 on ftp.andrew. This version adds a couple of new features and some slight code reorganization. We'd love to have people try it out. I noticed that pop3d.c now has some APOP code. Does that mean there's a new version of Cyrus SASL containing the sasl_checkapop() function? Where can I get it? Have fun, Larry Greenfield -- Larry M. Rosenbaum [EMAIL PROTECTED] Bldg 4500-N, Room E-218 865 574-8155 phone PO Box 2008, MS 6271865 241-4000 fax Oak Ridge, TN 37831-6271 Oak Ridge National Laboratory, Network Computing Services group
Re: Cyrus IMAP 2.0.13 released
this forking problem is related only in the case of Linux environment ? I have not noticed it on FreeBSD. thanks Rick On Fri, 27 Apr 2001, Helmut Apfelholz wrote: Hi, this is greate that the development of the server is moving along. I hope that you guys at cmu also watch to mailing list, and have seen the 'forking problem' that ppl here have been describing. We will try to test the newest release and will report ASAP if the problem persists or not. Thank you Helmut Apfelholz --- Lawrence Greenfield [EMAIL PROTECTED] wrote: I've placed Cyrus IMAP 2.0.13 on ftp.andrew. This version adds a couple of new features and some slight code reorganization. We'd love to have people try it out. Of especial interest may be the partial virtual domain support via the multiple config files (I'd love a contributed doc/virtual-domains.html) and support for a single imapd process to service multiple clients in succession, cutting down the number of forks. This approach will be extended to lmtpd and pop3d in the future. Finally, the master process now accepts the maxchild service modifier to limit how many instances of a particular service may be active at one time. Cyrus IMAP 2.0.13 is available at ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/cyrus-imapd-2.0.13.tar.gz Have fun, Larry Greenfield __ Do You Yahoo!? Yahoo! Auctions - buy the things you want at great prices http://auctions.yahoo.com/
Re: Cyrus IMAP 2.0.13 released
Larry M. Rosenbaum wrote: I've placed Cyrus IMAP 2.0.13 on ftp.andrew. This version adds a couple of new features and some slight code reorganization. We'd love to have people try it out. I noticed that pop3d.c now has some APOP code. Does that mean there's a new version of Cyrus SASL containing the sasl_checkapop() function? Where can I get it? I sent the sasl_checkapop() patch to Larry for either inclusion in the library itself or as a contrib patch. I'm assuming it will be included with 1.5.28+. In the meantime, I have attached the patch against CVS for your enjoyment. Keep in mind that the current implementation is just a stop-gap until Cyrus SASL v2. Ken -- Kenneth Murchison Oceana Matrix Ltd. Software Engineer 21 Princeton Place 716-662-8973 x26 Orchard Park, NY 14127 --PGP Public Key--http://www.oceana.com/~ken/ksm.pgp sasl_apop_patch.gz
Re: Cyrus IMAP 2.0.13 released
Helmut Apfelholz [EMAIL PROTECTED] writes: this is greate that the development of the server is moving along. I hope that you guys at cmu also watch to mailing list, and have seen the 'forking problem' that ppl here have been describing. Yes, we read the mailing lists but have been busy with some issues internally. In chatting briefly with Larry, we agreed that we both need to ponder the issue a little more and so we haven't responded (beyond throwing in a quick hack in this release). We saw some problems, though not as severe and perhaps the problem is most severe under Linux. Speaking of Linux, did you do the chattr +S to make the disk writes synchronous? If you are willing to lose data on a crash, you may want to see if performance improves by changing that. If you aren't willing to lose data on a crash, maybe you can see if resierfs performs better? Do you have a set of test scripts that shows this behavior? Some other notes... a) you don't have to use duplicate delivery suppression if it appears that the locking overhead on that is getting in the way. b) If you are primarily using POP some of the recent changes probably don't help since we only implemented the process reuse in imapd (and probably will do this in lmtpd eventually too) c) We got around some of our problems by running sendmail in queue only mode. In this manner, email comes in but gets dumped into a queue. A queue processor fires up, connects to a lmtpd and sends down a bunch of messages. This helps reduce the number of simultaneous lmtpds Walter
Re: Cyrus IMAP 2.0.13 released
I've had a problem that might be in under FBSD 3.5stable. I haven't implemented Cyrus yet, but under load testing I see a large number of sendmail processes that are in a Client Greeting state. I presume that sendmail is trying to talk to lmtpd and not getting a response. I've gotten around the problem by using the O DeliveryMode=q option in the cf that that has been mentioned here several times. rj45 wrote: this forking problem is related only in the case of Linux environment ? I have not noticed it on FreeBSD. thanks Rick On Fri, 27 Apr 2001, Helmut Apfelholz wrote: Hi, this is greate that the development of the server is moving along. I hope that you guys at cmu also watch to mailing list, and have seen the 'forking problem' that ppl here have been describing. We will try to test the newest release and will report ASAP if the problem persists or not. Thank you Helmut Apfelholz --- Lawrence Greenfield [EMAIL PROTECTED] wrote: I've placed Cyrus IMAP 2.0.13 on ftp.andrew. This version adds a couple of new features and some slight code reorganization. We'd love to have people try it out. Of especial interest may be the partial virtual domain support via the multiple config files (I'd love a contributed doc/virtual-domains.html) and support for a single imapd process to service multiple clients in succession, cutting down the number of forks. This approach will be extended to lmtpd and pop3d in the future. Finally, the master process now accepts the maxchild service modifier to limit how many instances of a particular service may be active at one time. Cyrus IMAP 2.0.13 is available at ftp://ftp.andrew.cmu.edu/pub/cyrus-mail/cyrus-imapd-2.0.13.tar.gz Have fun, Larry Greenfield __ Do You Yahoo!? Yahoo! Auctions - buy the things you want at great prices http://auctions.yahoo.com/