Re: [Dovecot] Dovecot Ldap Problem
> > Dovecot never disconnects from LDAP server. You're using > passdb/userdb ldap and not pam_ldap, right? I am using passdb/userdb ldap not pam_ldap. -- Rgds, Aravind M D
[Dovecot] Dovecot Ldap Problem
Hi All, My server is running with Dovecot 1.2.11 and openldap 2.4.17 it was working fine but suddenly my dovecot auth started to failing sometimes. After restarting dovecot everything will work fine. The slapd log file is showing "deffering operation: too many executing". When i googled i have seen that if the client is not properly disconnecting the connection slapd will give this error. Can you please let me know whether it is a dovecot problem and if so why it is occuring. -- Rgds, Aravind M D aravind.divaka...@yukthi.com Helpdesk Team Yukthi Systems Pvt Ltd Ph:- 08042351119 08042351120
Re: [Dovecot] Duplicate Elimination
> On 24.2.2010, at 16.15, Steffen Kaiser wrote: > >>> group1: us...@example.com,us...@example.com >>> group2: us...@example.com,us...@example.com >>> >>> If i am sending a mail to group1 and group2. It is going two times a >>> single mail to the user1 mailbox. But in dovecot it is not eliminating >>> since it has the same message-id. Previously i was able to accomplish >>> this >>> task with cyrus duplicate elimination. >> >> "two times a single mail to the user1 mailbox" >> actually I do expect that the MTA detects the duplication of the two >> aliases and does NOT send the mail two times to user1. > > Yeah, I think that's what I answered before too. Dovecot currently has no > duplicate elimination code. > > (I hate the whole concept of dropping incoming messages as duplicates > based on Message-ID. I would definitely want it disabled for my own > mails, for example that would mean I couldn't filter messages to mailing > lists based on List-ID: header, because the first mail usually comes > directly from the other user, and then the second mail with the List-ID: > would get dropped.) > > Thanks for your valuable information. Actually its not a mailing list, its just an aliases created in my mail server. I had made changes in my mta to handle duplicate mails. Rgds, Aravind M D
Re: [Dovecot] Duplicate Elimination
On Wed, 24 Feb 2010 13:11:04 +0100, Robert Schetterer wrote: > Am 24.02.2010 12:58, schrieb Aravind Divakaran: >> >> >> Hi All, >> >> In my mail server i have two groups and one user is belonging >> to both the groups. Previously i was using cyrus at that time when i am >> sending a mail to the two groups there will only one mail in the user >> mailbox . But now the user is getting one more copy of same mail. Is >> there >> is a way i can avoid this. >> >> Rgds, >> >> Aravind M D > > Hi cant think that dovecot itself is the source of your problem > check sieve and/or forward rules as well as your smtp server logs > sometimes i.e amavis etc doubles mail Actually as i told there were two groups i have created in my mailserver which contain a common emailid. For example group1: us...@example.com,us...@example.com group2: us...@example.com,us...@example.com If i am sending a mail to group1 and group2. It is going two times a single mail to the user1 mailbox. But in dovecot it is not eliminating since it has the same message-id. Previously i was able to accomplish this task with cyrus duplicate elimination.
[Dovecot] Duplicate Elimination
Hi All, In my mail server i have two groups and one user is belonging to both the groups. Previously i was using cyrus at that time when i am sending a mail to the two groups there will only one mail in the user mailbox . But now the user is getting one more copy of same mail. Is there is a way i can avoid this. Rgds, Aravind M D
Re: [Dovecot] Dovecot Deliver CPU usage
> On Tue, 2009-12-15 at 15:08 -0500, Timo Sirainen wrote:
>> > Dec 14 12:46:04 server01 postfix/pipe[1870]: 57489E913:
>> > to=, orig_to=, relay=dovecot,
>> > delay=1703, delays=1369/328/0/5.9, dsn=5.7.0, status=bounced
>> (permission
>> > denied. Command output: Not enough disk space )
>>
>> Can you get a strace from this kind of a failed delivery? The strace
>> should contain ENOSPC string.
>
> Looking at the code .. I think that can happen with both EDQUOT and
> ENOSPC errors from a write, but nothing else..
Strace with ENOSPC
09:57:39.782569 write(10, "T5vfloaIeSOsJdT6IXB146wSWLNKqK\n4v"..., 28672)
= -1 ENOSPC (No space left on device)
09:57:40.695074 fstat(10, {st_mode=S_IFREG|0644, st_size=2068480, ...}) = 0
09:57:40.707778 munmap(0x7fa35c97b000, 2101248) = 0
09:57:40.708119 close(10) = 0
09:57:40.708162
unlink("/home/vmail/example.com/user1/tmp/1261456056.M913860P17478.node01")
= 0
09:57:40.711774 sendto(6, "<22>Dec 22 09:57:40 dovecot: deli"..., 201,
MSG_NOSIGNAL, NULL, 0) = 201
09:57:40.712087 munmap(0x7fa35dc47000, 200704) = 0
09:57:40.712148 write(2, "Not enough disk space\n"..., 22) = 22
09:57:40.712235 sendto(6, "<22>Dec 22 09:57:40 dovecot: deli"..., 189,
MSG_NOSIGNAL, NULL, 0) = 189
09:57:40.712353 exit_group(77)
Re: [Dovecot] Dovecot Deliver CPU usage
On Thu, 3 Dec 2009 09:05:21 -0600, Timo Sirainen wrote: > On Dec 3, 2009, at 5:50 AM, Aravind Divakaran wrote: > >>> strace -tt would have actually been also more helpful, so it would have >>> included timestamps. Are you using NFS or something similar for your >>> filesystem in /home? The strace does show that Dovecot seems to be >>> reading >>> from a file that it just wrote to, so there's something non-optimal but >>> it >>> necessarily shouldn't be that slow. >> >> You can find my strace with timestamp on the below link. >> >> http://pastebin.com/m2942fbdd > > That's only 400 milliseconds. I thought the delivers would be running for > at least a few seconds? Can you show such a strace -tt? > >> My /home partition using san storage with ocfs filesystem. > > Anyway I suppose this could be the reason, if the problem is the reads.. > > You could also try if it helps when you increase this in > src/deliver/deliver.c: > > #define MAIL_MAX_MEMORY_BUFFER (1024*128) > > to > > #define MAIL_MAX_MEMORY_BUFFER (1024*1024*100) > > i.e. 100 MB (or whatever is large enough for most mails). I had made the changes for MAIL_MAX_MEMORY_BUFFER and installed the new dovecot. I was monitoring my mail server for past few days. Today i experienced the same problem My log file shows Dec 14 12:38:19 server01 dovecot: deliver(us...@example.com): Created dotlock file's timestamp is different than current time (1260774499 vs 1260774377): /home/vmail/example.com/user1/dovecot-uidlist Dec 14 12:38:19 server01 dovecot: deliver(us...@example.com): Created dotlock file's timestamp is different than current time (1260774499 vs 1260774377): /home/vmail/example.com/user2/dovecot-uidlist The above erros will show when i change the maxproc from 4 to default in postfix. At this time dovecot deliver process is using lot of cpu. Dec 14 12:46:04 server01 postfix/pipe[1870]: 57489E913: to=, orig_to=, relay=dovecot, delay=1703, delays=1369/328/0/5.9, dsn=5.7.0, status=bounced (permission denied. Command output: Not enough disk space ) As my hardisk has enough space, why dovecot is giving this error? -Aravind
Re: [Dovecot] Dovecot Deliver CPU usage
> On Dec 1, 2009, at 4:48 AM, Aravind Divakaran wrote: > >>> You mean a single deliver process can run a long time with 100% CPU? >>> Getting a strace of the process could be helpful. >> >> Yes, dovecot deliver process is eating a lot of CPU. please find my >> strace >> output on the below link. >> >> http://pastebin.com/m6965f4ab > > strace -tt would have actually been also more helpful, so it would have > included timestamps. Are you using NFS or something similar for your > filesystem in /home? The strace does show that Dovecot seems to be reading > from a file that it just wrote to, so there's something non-optimal but it > necessarily shouldn't be that slow. You can find my strace with timestamp on the below link. http://pastebin.com/m2942fbdd My /home partition using san storage with ocfs filesystem. I have configured mailarchivel, where in which a copy of all mails will sent to one mailbox. In the archive mailbox using sieve script i am sorting the mail according to username, year, month and delivering it. When ever i am removing my max procs option limit from 4 to default in postfix. System cpu usage become high. Then i have to put the limit again to make my cpu usage go down. Rgds, Aravind
Re: [Dovecot] Dovecot Deliver CPU usage
> On Nov 28, 2009, at 3:21 AM, Aravind Divakaran wrote: > >> We have upgraded the dovecot from 1.0 to 1.2.8. After >> upgradation we have noticed that dovecot deliver process is utilizing a >> lot >> of CPU. As a result the mailsystem will stop working. When we limit the >> no >> of procs in postfix for dovecot deliver the CPU usage will come down. >> But >> we are not facing this problem with the previous version. Below is the >> log >> files for dovecot when the error occurs > > You mean a single deliver process can run a long time with 100% CPU? > Getting a strace of the process could be helpful. Yes, dovecot deliver process is eating a lot of CPU. please find my strace output on the below link. http://pastebin.com/m6965f4ab > >> write_full(/home/vmail/dhanbank.co.in/vinukumarvm/temp.server.be3e5a849a909847) >> failed: No space left on device >> dovecot: deliver(u...@example.com): >> stat(/tmp/Dovecot Delivery Mail) failed: No space left on device >> >> But my >> system has 54gb of free space for root filesystem where tmp residing. >> Can >> anyone help me to solve this issue. > > The /tmp/ part is wrong, I should get that fixed probably. Is there space > in /home/vmail? Or do you use filesystem quotas? (I see you're also using > Maildir++ quotas.) > My home partition has 353gb of free space. I am using maildir quota not using filesystem quota.
Re: [Dovecot] Dovecot Deliver CPU usage
On Sat, 28 Nov 2009 09:17:12 -0500, Charles Marcus wrote: > On 11/28/2009, Aravind Divakaran (aravind.divaka...@yukthi.com) wrote: >> write_full(/home/vmail/dhanbank.co.in/vinukumarvm/temp.server.be3e5a849a909847) >> failed: No space left on device >> dovecot: deliver(u...@example.com): >> stat(/tmp/Dovecot Delivery Mail) failed: No space left on device >> >> But my >> system has 54gb of free space for root filesystem where tmp residing. Can >> anyone help me to solve this issue. > > Its not looking at / for the temp, it is using the users home directory, > according to the above... > > Are you using quotas? > > dovecot -n output? Yes i am using quota, below is my configuration # 1.2.8: /etc/dovecot/dovecot.conf # OS: Linux 2.6.30 x86_64 Debian 5.0 ocfs2 log_timestamp: %Y-%m-%d %H:%M:%S protocols: imap imaps pop3 pop3s managesieve disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable(default): /usr/lib/dovecot/imap-login login_executable(imap): /usr/lib/dovecot/imap-login login_executable(pop3): /usr/lib/dovecot/pop3-login login_executable(managesieve): /usr/lib/dovecot/managesieve-login mail_privileged_group: mail mail_uid: 1003 mail_gid: 1003 mail_location: maildir:/home/vmail/%d/%n mbox_write_locks: fcntl dotlock mail_executable(default): /usr/lib/dovecot/imap mail_executable(pop3): /usr/lib/dovecot/pop3 mail_executable(managesieve): /usr/lib/dovecot/managesieve mail_plugins(default): quota imap_quota acl listescape mail_plugins(imap): quota imap_quota acl listescape mail_plugins(pop3): quota mail_plugins(managesieve): mail_plugin_dir(default): /usr/lib/dovecot/modules/imap mail_plugin_dir(imap): /usr/lib/dovecot/modules/imap mail_plugin_dir(pop3): /usr/lib/dovecot/modules/pop3 mail_plugin_dir(managesieve): /usr/lib/dovecot/modules/managesieve namespace: type: private separator: / inbox: yes list: yes subscriptions: yes lda: postmaster_address: postmas...@example.com mail_plugins: quota sieve auth default: mechanisms: login plain passdb: driver: ldap args: /etc/dovecot/dovecot-ldap.conf userdb: driver: ldap args: /etc/dovecot/dovecot-ldap.conf socket: type: listen client: path: /var/spool/postfix/private/auth mode: 432 user: postfix group: postfix master: path: /var/run/dovecot/auth-master mode: 384 user: vmail plugin: quota: maildir:storage=25 sieve: /home/vmail/sieve/%d/%n/.dovecot.sieve sieve_storage: /home/vmail/sieve/%d/%n/sieve sieve_extensions: +imapflags acl: vfile:/etc/dovecot/acls Users quota is getting from ldap. -Aravind
[Dovecot] Dovecot Deliver CPU usage
Hi All, We have upgraded the dovecot from 1.0 to 1.2.8. After upgradation we have noticed that dovecot deliver process is utilizing a lot of CPU. As a result the mailsystem will stop working. When we limit the no of procs in postfix for dovecot deliver the CPU usage will come down. But we are not facing this problem with the previous version. Below is the log files for dovecot when the error occurs write_full(/home/vmail/dhanbank.co.in/vinukumarvm/temp.server.be3e5a849a909847) failed: No space left on device dovecot: deliver(u...@example.com): stat(/tmp/Dovecot Delivery Mail) failed: No space left on device But my system has 54gb of free space for root filesystem where tmp residing. Can anyone help me to solve this issue. Rgds, Aravind
