[Dovecot] 1.0.rc17 released
http://dovecot.org/releases/dovecot-1.0.rc17.tar.gz http://dovecot.org/releases/dovecot-1.0.rc17.tar.gz.sig
- MySQL authentication caused username to show up as "OK" in rc16.Timo Sirainen wrote:
I installed rc17 on my test server (SUSE Linux Enterprise 10) with only IPV6 disabled (otherwise I took the defaults). The end of configure show:
Install prefix ...................... : /usr/local File offsets ........................ : 64bit I/O loop method ..................... : poll File change notification method ..... : dnotify Building with SSL support ........... : yes (OpenSSL) Building with IPv6 support .......... : no Building with pop3 server ........... : yes Building with mail delivery agent .. : yes Building with GSSAPI support ........ : no Building with user database modules . : static prefetch passwd passwd-file checkpassword vpopmail (modules) Building with password lookup modules : passwd passwd-file shadow checkpassword vpopmail (modules) Building with SQL drivers ............:
When trying to move large numbers of messages to Trash using Thunderbird 1.5.0.9, I would get "Lost connection with server" messages at anything over about 1500 messages. The strange thing is that when I checked the trash folder, the deleted messages were there multiple times, i.e. everytime I attempted to move all messages to Trash, plus everytime it failed until I got below the magic number.
Nothing shows up in the log (which I'm sending to /dev/stderr, in case that matters); I'll try and enable rawlog. More details when I have them...
John
-- John Peacock Director of Information Research and Technology Rowman & Littlefield Publishing Group 4501 Forbes Boulevard Suite H Lanham, MD 20706 301-459-3366 x.5010 fax 301-429-5748
On 9.1.2007, at 20.35, John Peacock wrote:
When trying to move large numbers of messages to Trash using
Thunderbird 1.5.0.9, I would get "Lost connection with server"
messages at anything over about 1500 messages. The strange thing
is that when I checked the trash folder, the deleted messages were
there multiple times, i.e. everytime I attempted to move all
messages to Trash, plus everytime it failed until I got below the
magic number.
Loses connection immediately, or sometimes later? straceing the imap
process might show something useful. Also do you use
maildir_copy_with_hardlinks=yes?
Timo Sirainen wrote:
Loses connection immediately, or sometimes later? straceing the imap process might show something useful. Also do you use maildir_copy_with_hardlinks=yes?
30 seconds maybe (certainly not minutes). It may or may not be enough time to get strace involved; I'll try. I'm a little confused about the total lack of logging; could it just be TBird getting impatient?
FWIW, when I perform a successful delete there are two distinct phases:
- Cylon - the little status bar goes back and forth and back and forth
- Seizure - the mouse pointer flashes wildly
For the unsuccessful runs, I only get the "Cylon" phase... ;-)
I don't have maildir_copy_with_hardlinks=yes enabled (I was trying to be as vanilla as possible). I'll try it again with that set (I have to figure out how to de-dupe the messages in my trash folder to put them back into my Inbox for testing)...
John
-- John Peacock Director of Information Research and Technology Rowman & Littlefield Publishing Group 4501 Forbes Boulevard Suite H Lanham, MD 20706 301-459-3366 x.5010 fax 301-429-5748
On 9.1.2007, at 21.59, John Peacock wrote:
30 seconds maybe (certainly not minutes). It may or may not be
enough time to get strace involved; I'll try. I'm a little
confused about the total lack of logging; could it just be TBird
getting impatient? .. I don't have maildir_copy_with_hardlinks=yes enabled (I was trying
to be as vanilla as possible). I'll try it again with that set (I
have to figure out how to de-dupe the messages in my trash folder
to put them back into my Inbox for testing)...
30 seconds sounds a bit little for Thunderbird to wait. Without
hardlinks I think it's easily possible that copying 1500 messages is
going to take time..
On 1/9/07, Timo Sirainen tss@iki.fi wrote:
On 9.1.2007, at 21.59, John Peacock wrote:
30 seconds maybe (certainly not minutes). It may or may not be enough time to get strace involved; I'll try. I'm a little confused about the total lack of logging; could it just be TBird getting impatient? .. I don't have maildir_copy_with_hardlinks=yes enabled (I was trying to be as vanilla as possible). I'll try it again with that set (I have to figure out how to de-dupe the messages in my trash folder to put them back into my Inbox for testing)...
30 seconds sounds a bit little for Thunderbird to wait. Without hardlinks I think it's easily possible that copying 1500 messages is going to take time..
We have Maildir on a netapp filer and experience this with Thunderbird as well, 1.0rc15. It takes about a minute to trigger.
Now that I see this, is there really any compelling reason not to use maildir_copy_with_hardlinks? I think it is time to enable it!
On 10.1.2007, at 20.46, bofh list wrote:
Now that I see this, is there really any compelling reason not to
use maildir_copy_with_hardlinks? I think it is time to enable it!
There are only two reasons that I've found so far:
OSX + HFS+ sucks with it
Copying to shared mailboxes (with maildir-shared file) doesn't
update the group/permissions
I'll probably change 2) some day so that it fallbacks to copying the
normal way. I think in Dovecot v2.0 I'll change to
maildir_copy_with_hardlinks=yes by default, with everything except OSX.
On 1/10/07, Timo Sirainen tss@iki.fi wrote:
On 10.1.2007, at 20.46, bofh list wrote:
Now that I see this, is there really any compelling reason not to use maildir_copy_with_hardlinks? I think it is time to enable it!
There are only two reasons that I've found so far:
OSX + HFS+ sucks with it
Copying to shared mailboxes (with maildir-shared file) doesn't update the group/permissions
I'll probably change 2) some day so that it fallbacks to copying the normal way. I think in Dovecot v2.0 I'll change to maildir_copy_with_hardlinks=yes by default, with everything except OSX.
Thanks Timo.
Is it safe to make this change with shutdown_clients = no ?
- John Peacock, 2007-01-09 19:35
[...] When trying to move large numbers of messages to Trash using Thunderbird 1.5.0.9, I would get "Lost connection with server" messages at anything over about 1500 messages. The strange thing is that when I checked the trash folder, the deleted messages were there multiple times, i.e. everytime I attempted to move all messages to Trash, plus everytime it failed until I got below the magic number. [...]
I think this is a Thunderbird problem rather than a Dovecot one. I had the same thing happening recently when I moved last year's logfiles to my archive folder (36000+ messages).
Thunderbird would pop up the "Lost connection" error, and the (re)moved messages showed up in Trash multiple times. I solved the problem by using Mutt instead of Thunderbird to move the messages - it took a while, but it worked on the first try without any hiccups whatsoever.
I haven't tried it, but tuning Thunderbirds timeout values might help. It has a (global?) mailnews.tcptimeout which defaults to 60 seconds, and a per-server mail.server.serverX.timeout which does not have a default value assigned, but seems to get set to 29 seconds when you create your account.
You can check/modify the timeout values in Thunderbird's "about:config" page.
HTH, Thomas
=-------------------------------------------------------------------------=
- Thomas "ZlatkO" Zajic zlatko@gmx.at Linux-2.6.19 & Thunderbird-1.5 -
"It is not easy to cut through a human head with a hacksaw." (M. C.) -
I had that problem with an installation of 1.0rc10 in Solaris9 + maildir
- plugin Quota. I installed the same version and same configuration in RedHat Enterprise Linux ES 4.0 Update 4 and it worked well.
I left Solaris…
John Peacock escribió:
Timo Sirainen wrote:
I installed rc17 on my test server (SUSE Linux Enterprise 10) with only IPV6 disabled (otherwise I took the defaults). The end of configure show:
Install prefix ...................... : /usr/local File offsets ........................ : 64bit I/O loop method ..................... : poll File change notification method ..... : dnotify Building with SSL support ........... : yes (OpenSSL) Building with IPv6 support .......... : no Building with pop3 server ........... : yes Building with mail delivery agent .. : yes Building with GSSAPI support ........ : no Building with user database modules . : static prefetch passwd passwd-file checkpassword vpopmail (modules) Building with password lookup modules : passwd passwd-file shadow checkpassword vpopmail (modules) Building with SQL drivers ............:
When trying to move large numbers of messages to Trash using Thunderbird 1.5.0.9, I would get "Lost connection with server" messages at anything over about 1500 messages. The strange thing is that when I checked the trash folder, the deleted messages were there multiple times, i.e. everytime I attempted to move all messages to Trash, plus everytime it failed until I got below the magic number.
Nothing shows up in the log (which I'm sending to /dev/stderr, in case that matters); I'll try and enable rawlog. More details when I have them...
John
participants (5)
-
Antonio Casado Rodríguez
-
bofh list
-
John Peacock
-
Thomas Zajic
-
Timo Sirainen