On 08 Oct 08, at 0536, Bron Gondwana wrote:
On Mon, Oct 06, 2008 at 11:33:18AM -0400, Ken Murchison wrote:
I just put together a second release candidate for Cyrus 2.3.13. I'd
appreciate any independent testing before I release this to the
masses.
Sorry about the delay in testing -
We were only looking at SQL to replace BDB (deliver.db,
tls_sessions.db), because we still think that skiplist is superior
from mailboxes.db and seen.db. If you do some heavy testing with
the BDB code we'd be interested in your results.
OK, that's interesting. I'll get 2.3.13 onto
On 08 Oct 08, at 1119, Bron Gondwana wrote:
On Wed, 8 Oct 2008 09:24:49 +0100, Ian G Batten [EMAIL PROTECTED]
said:
What's the testing status of the SQL backend for cyrusdb? I'll
switch batten.eu.org over to it, but that only has a dozen or so
users; ftel.co.uk's 1000+ users might
the mail, the indices, and the mailboxes.db. Each can be on
separate storage with optimal price vs performance characteristics
for the proposed load.
We have 48000 mailboxes on behalf of 2000 users, with typically 1700
imapd processes during the day and ~1000 overnight. There's about
On 11 Aug 08, at 1648, Martin Schweizer wrote:
Hello
I have two mail server (FreeBSD 7.0, sendmail/cyrus v2.3.12p2), incl.
replication with sync_client/-master. Until now they works perfect.
Now I changed on the sync_server from sasldb (Berkeley db1.85) to
saslauthd (with saslauthd -a
On 21 Jul 08, at 2135, Steve Webb wrote:
I've got pop users
that can't access IMAP (using phones for checking email when on travel
with leave messages on server then suck down the emails when they
arrive
back at a desktop).
Have you confirmed that they can't use IMAP, or is that just what
On 19 Jul 08, at 1313, Bron Gondwana wrote:
Unfortunately, the cyrus database interface sort of sucks from
a consistency perspective, it's dangerous to call any function
that needs database access if you have a database transaction
open
I understand some of the technical, philosophical and
On 12 Mar 08, at 1654, Marco Broglia wrote:
I have problems to enable Cyrus pop because I don't know their
passwords and I don't want to change or ask to users (for several
reasons) for them.
saslauthd and auto_transition will cope with this. People with
entries in sasldb are
On 05 Mar 08, at 1549, Simon Matter wrote:
On Tue, 4 Mar 2008, Ian G Batten wrote:
software RAID5 is a performance
disaster area at the best of times unless it can take advantage of
intimate knowledge of the intent log in the filesystem (RAID-Z does
this),
actually, unless you have top
On 28 Feb 08, at 2256, Kenneth Marshall wrote:
It may be that the software RAID 5 is your problem. Without the
use of NVRAM for a cache, all of the writes need all 3 disks.
That will cause quite a bottle-neck.
In general, RAID5 writes require two reads and two writes,
independent of the
On 29 Jan 08, at 2004, Paul Dekkers wrote:
But I don't think it will be very easy to
combine the two: the Vasco tokens provide you with one-time passwords,
and for IMAP access, you'll have more then just one connection. My
Thunderbird client already makes a new connection for each folder I
The Cyrus server I run for my employer is sat on our internal
network, and remote users access either the IMAP port or the
associated Squirrelmail instance via our VPN. They come in via a
Cisco IPSec VPN server, secured with SecureID.
My private Cyrus server, which sits in borrowed space
I once had to delete the cyrus.* files prior to a reconstruct: they
had become toxic to the point that Cyrus's attempts to make use of
the information in them was causing problems.
ian
Cyrus Home Page: http://cyrusimap.web.cmu.edu/
Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki
On 30 Nov 07, at 0234, Vincent Fox wrote:
We had sym-linked imap/proc directory to a size-limited
TMPFS a while back.
Now I'm thinking to do the same for imap/socket and imap/log
Any reasons against? Other candidates?
What's the point? The socket directory only contains a handful of
On 20 Nov 07, at 1756, David Lang wrote:
however a fsync on a journaled filesystem just means the data needs
to be
written to the journal, it doesn't mean that the journal needs to
be flushed to
disk.
on ext3 if you have data=journaled then your data is in the journal
as well and
all
On 19 Nov 07, at 2139, Scott Adkins wrote:
This does bring up an important point... we had to disable duplicate
suppression on our server because it was just causing too much
contention.
When every single mail message being delivered has to check against
the
database, it is a problem.
In a NON-replicated setup, do the changes to the GUID have an
impact? Can I just put 2.3.10 on with a quick restart of the
mailsystem, or is there More To It?
I have 1.7TB of mail, about 40K mailboxes, about 10 million pieces of
mail. So I don't want to do an upgrade which will kick off
On 20 Nov 07, at 1146, Bron Gondwana wrote:
The index files are pretty small, and they rebuild fast :) They
get streamed
into new copies every single expunge anyway.
What's involved in the rebuild? I have users with tens of thousands
of messages in a single mailbox, so delaying that
On 17 Nov 07, at 0909, Rob Mueller wrote:
This shouldn't really be a problem. Yes the whole file is locked
for the
duration of the write, however there should be only 1 fsync per
transaction, which is what would introduce any latency. The
actual writes
to the db file itself should be
On 15 Nov 07, at 1504, Michael Bacon wrote:
Interesting thought. We haven't gone to ZFS yet, although I like
the idea
a lot. My hunch is it's an enormous win for the mailbox
partitions, but
perhaps it's not a good thing for the meta partition. I'll have to
let
someone else who knows
I was interested to see someone suggesting putting proc into tmpfs.
That's slightly painful if /var/imap is in ZFS: the order in which
mounts take place means you can't just put /var/imap/proc tmpfs into /
etc/vfstab if /var/imap is coming in through ZFS. A glance at the
source code says
On 15 Nov 07, at 1045, Simon Matter wrote:
I was interested to see someone suggesting putting proc into tmpfs.
That's slightly painful if /var/imap is in ZFS: the order in which
mounts take place means you can't just put /var/imap/proc tmpfs
into /
etc/vfstab if /var/imap is coming in
On 13 Nov 07, at 1505, David Chait wrote:
One key piece of functionality that seems to be missing from every OSS
solution mentioned thus far is mobile device push support
(Activesync),
this is not to be underestimated as it is for us, a key reason why we
are ultimately being forced to
On 13 Nov 07, at 0724, Rudy Gevaert wrote:
If we could ever get a decent calendar system that works together with
Cyrus or other software many people would be happy.
We run Cyrus for mail, Oracle Collaboration Suite ($$, but not
Exchange-$$) for calendaring. The Outlook plugin that allows
On 13 Nov 07, at 0812, Scott M. Likens wrote:
No it's not
as seamless as Exchange, but it works just fine and it's an open
standard.
However, people don't want calendaring, they want Outlook. Offer a
solution which doesn't allow Outlook to work ``like it should'' and
you risk learning
We've been having a chat about how useful it would be to have timed
actions in sieve: so that a vacation message could be set up for a
duration which would automatically revert, so that a forwarding could
be set up for the duration of a short-term project, etc, etc. The
naive way is to
On 13 Nov 07, at 1335, Adam Tauno Williams wrote:
3. Can't handle high load very well, in fact it handles load
horribly.
I have a friend who works at a small shop who reports exactly the same
issue with Zimbra, s..ll...ooo.....
3) ClamAV. Do note how much email I said we dealt with
A common scenario when we are moving users between partitions is to
want to move their archived mail first, because there's less risk of
disrupting them, and then their top-level mailbox last.
So we want to do:
rename --partition new-partition user.name.box user.name.box
and then later
).
Begin forwarded message:
From: Ian G Batten [EMAIL PROTECTED]
Date: Thu 25 Oct 07 12:30:57 BDT
To: Ken Murchison [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED], Cyrus Mailing List info-
[EMAIL PROTECTED]
Subject: Re: Cyrus IMAPd 2.3.10 Released
I've just compiled 2.3.10 on batten.eu.org (my
user.markr has, through a sequence of events, ended up with data on
both partition8 (/var/imap/partition-8) and default (/var/imap/
partition-1). All his mailboxes have been consolidated into
default, but there are still copies in partition-8.
So I think ctl_mboxlist -v should report that
On 31 Oct 07, at 1527, Ken Murchison wrote:
Ian G Batten wrote:
user.markr has, through a sequence of events, ended up with data
on both partition8 (/var/imap/partition-8) and default (/var/
imap/ partition-1). All his mailboxes have been consolidated
into default
To what extent has Cyrus been tested with 64 bit builds? One of my
users has managed to create a mailbox with a 2.0G cyrus.cache, which
of course can't be mmap'd. The mere act of finding out how many
files there are in there is itself a problem: I've got
perl -e 'use IO::Dir; my $d=new
I've just compiled 2.3.10 on batten.eu.org (my private x86 servers)
and although it looks OK on the Solaris 10 system, it's in deep
trouble on the elderly Linux machine. Both are upgrades from 2.3.7,
the Solaris box is a replication target, the Linux box is a
replication master that
On 25 Oct 07, at 1230, Ian G Batten wrote:
I've just compiled 2.3.10 on batten.eu.org (my private x86 servers)
and although it looks OK on the Solaris 10 system, it's in deep
trouble on the elderly Linux machine. Both are upgrades from
2.3.7, the Solaris box is a replication target
idled_shutdown_check: 0
Are you applying third-party patches? 'idled_shutdown_check' isn't
a valid option in the stock distro.
No: the config dates back to the dawn of time, but the installation
today is a straight download and compile.
ian
Cyrus Home Page:
:50:25 -0500 (CDT) Returned mail:
User unknown ((NIL NIL postoffice batten.eu.org)) ((NIL NIL
postoffice batten.eu.org)) ((NIL NIL postoffice
batten.eu.org)) ((Mail Delivery Subsystem NIL MAILER-DAEMON
batten.eu.org)) NIL NIL NIL [EMAIL PROTECTED])
and then a lot more of the same.
Ian G
On the Linux box, all fresh compilations aside from the sasl 2.1.15
binaries:
imapd 2.3.7 + sasl 2.1.15: works
imapd 2.3.7 + sasl 2.1.22: works
imapd 2.3.9 + sasl 2.1.15: not tried
imapd 2.3.9 + sasl 2.1.22: works
imapd 2.3.10 + sasl 2.1.15: fails (cannot examine mailboxes, then
coredumps
On 25 Oct 07, at 1248, Ken Murchison wrote:
What does imapd.conf look like?
Does the output of 'ctl_mboxlist -d' look reasonable?
Does 'mbexamine user.igb' look reasonable?
OK, there's a steady stream of imapd processes being forked and then
dying on SIGSEGV. I've caught one in the act.
On 25 Oct 07, at 1501, Ken Murchison wrote:
Ian G Batten wrote:
On the Linux box, all fresh compilations aside from the sasl
2.1.15 binaries:
imapd 2.3.7 + sasl 2.1.15: works
imapd 2.3.7 + sasl 2.1.22: works
imapd 2.3.9 + sasl 2.1.15: not tried
imapd 2.3.9 + sasl 2.1.22: works
imapd
On 23 Oct 07, at 1943, David Carter wrote:
On Tue, 23 Oct 2007, Ken Murchison wrote:
Your problem is most likely related to using NFS. NFS has never been
recommended for Cyrus because is doesn't play nice with mmap() and
flock(), both of which are critical to the operation of Cyrus.
Since we moved to 2.2.6 on Solaris 10, with storage coming via NFS,
we've been seeing ``IOERROR: reading message: unexpected end of
file'' intermittently. It doesn't seem to cause any harm. I went
through and placed the function name in each occurrence of the
message (as a note,
, and it all worked fine.
ian
Regards.
On 10/11/07, Ian G Batten [EMAIL PROTECTED] wrote:
I'm performing a migration from 2.2.8 on an E450 running Solaris 10
Beta Build 58 to 2.3.9 on a T2000 running Solaris 10 latest. The
former isn't zoned (it wasn't available in that build), isn't smf'd
On 12 Oct 07, at 1901, Ian G Batten wrote:
I've just brought up our 2.3.9 installation. My logs are filled with
`idled: unrecognized message: 3', happening around the time of
login. Should I worry?
Tracked that one down. In 2.2 you need --with-idle=idled. With 2.3
you need --enable
I've just brought up our 2.3.9 installation. My logs are filled with
`idled: unrecognized message: 3', happening around the time of
login. Should I worry?
ian
Cyrus Home Page: http://cyrusimap.web.cmu.edu/
Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki
List Archives/Info:
I'm performing a migration from 2.2.8 on an E450 running Solaris 10
Beta Build 58 to 2.3.9 on a T2000 running Solaris 10 latest. The
former isn't zoned (it wasn't available in that build), isn't smf'd
(ditto) and isn't ZFS'd (ditto).
The new machine has Cyrus in a zone, rooted on ZFS, with
Looking at the migration script, it's essentially find . -name
cyrus.* -exec mv {} /meta/part/{} \;. That's quite slow, because it
has to stat() all the messages to find the files which are metadata.
I've written an alternative for the migration we're planning for
later this week, which
We have clients which issue LIST * when they start up. We think
we need them to do this, because we are making quite heavy use of
shared mailboxes so a mailbox may arrive in a hierarchy other than
INBOX.* to which the user should subscribe.
We have ~35K mailboxes (as reported by
)
/[EMAIL PROTECTED]: - starttxn_or_refetch(0x1f22f8, 0x0,
0x0,
0x0)
/1: stat(/var/imap/mailboxes.db, 0xFFBFC838) = 0
:wes
On 09 Oct 2007, at 09:19, Ian G Batten wrote:
WeBut at root I don't understand why LIST * should take any
longer than, for example
On 09 Oct 07, at 1522, Blake Hudson wrote:
Sorry this doesn't help solve your problem but it proves it should
be a lot faster than that.
Joseph Brennan
Columbia University Information Technology
Could database type differences (or contention) be an issue here? What
database format are
On 09 Oct 07, at 1522, Blake Hudson wrote:
Could database type differences (or contention) be an issue here? What
database format are each of you using?
Yes. With skiplist (took me several stabs at it to get the
conversion to work) it takes 0.19s. Versus ~250s with BDB. A slight
On 09 Oct 07, at 1748, Ian G Batten wrote:
On 09 Oct 07, at 1522, Blake Hudson wrote:
Could database type differences (or contention) be an issue here?
What
database format are each of you using?
Yes. With skiplist (took me several stabs at it to get the
conversion to work) it takes
I've compiled 2.2.12 on a Linksys NSLU2. It appears to work --- I can
rsync a mailbox on from a Sun and access it correctly. I've also got
Sendmail 8.13.3 built and awaiting a config file, so I can start
delivering mail to the slug soonest.
I have real work for this, honest, and it's not just
On Wed, 19 Jan 2005, Gregory Harris wrote:
Hi. Last Sunday I migrated all of my users mailboxes from a traditional
[...]
I have had a similar problem with one user that uses Ximian Evolution
1.4.6. Same error message on the server side, on the client side, says
I have only a tiny
I've put in an exception, but it may be worth someone looking at why the
Cyrus list trips over the Sendmail greet_pause code.
1090338598.297 Jul 20 16:49:58 i6KFntRU028281: rejecting commands from
LISTS2.andrew.cmu.edu [128.2.10.216] due to pre-greeting traffic
This implies that LISTS2 is
On Sun, 11 Jul 2004, Elizabeth Schwartz wrote:
I'm planning to build a new, larger, server with Solaris 9, and cyrus,
I'm running my new mailserver on Solaris 10, but I'm using ZFS for
several of the spool partitions. The performance rocks (even on some
mangy old A5000 arrays) and you get
On Thu, 10 Jun 2004, Florin Andrei wrote:
What's the most simple method to connect to a Cyrus IMAP server, login
as a user, go to a certain folder (the path is known), delete all
messages in it, then logout - and do all that from a shell script? (a
cron job)
There's bound to be an option to
As part of my move from 1.6.22 on Solaris 7 to 2.2.4 on Solaris 10, I
need to bulk move about six hundred mailboxes totalling about 120GBytes.
For the pilot users I've used imapsync, which works correctly (including
moving flags) but is quite slow and resource intensive.
For service use, I'd
On Thu, 10 Jun 2004, Ian G Batten wrote:
As part of my move from 1.6.22 on Solaris 7 to 2.2.4 on Solaris 10, I
need to bulk move about six hundred mailboxes totalling about 120GBytes.
For the pilot users I've used imapsync, which works correctly (including
moving flags) but is quite slow
On Thu, 03 Jun 2004, Rob Siemborski wrote:
Why on earth were you killing processes at random?
As a test of the stability of the databases. A database should be able
to take the explosive failure of any reader or writer with, at worst, a
reconstruction. Consider the scenario when the machine
On Thu, 03 Jun 2004, Rob Siemborski wrote:
I'll also ask the obvious -- did you subsequently stop the server and run
recovery on the database?
Well, master couldn't prefork anything, citing an inability to read
mailboxes.db. As reconstruct -m is currently unavailable I used
reconstruct -p to
On Fri, 04 Jun 2004, Rob Siemborski wrote:
I ment ctl_cyrusdb -r.
I thought that was run automatically when master started?
ian
---
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
For what it's worth, I'm building a new server to replace my
long-standing Solaris 7 + Cyrus 1.6.22 server --- one hour of downtime
since 1999. I've used the gcc from the SFW collection, Cyrus 2.2.4 and
Solaris 10 build 55 (which is probably the current Solaris Express bits:
I'm a Platinum Beta
I wiped out access to my Cyrus store for my father and my wife which I
went to 2.2.3. It turned out that the problem was that when they
connected with TLS (I didn't test that, sadly) imapd immediately exited
with ``imaps: required OpenSSL options not present''.
I traced through the source to
On Wed, 04 Feb 2004, Rob Siemborski wrote:
On Wed, 4 Feb 2004, Ken Murchison wrote:
Try servicename_tls_cert_file and servicename_tls_key_file where
servicename is the name of the service from cyrus.conf (note that this
is NOT the name of the binary).
I've fixed install-configure.html
On my small personal server, supporting a few family members, I run
Cyrus latest where possible. I was running 2.1.16 with every database
set to skiplist. Upgrading to 2.2.3 with the same ./configure command
./configure --with-openssl --with-sasl --without-cmulocal
--with-cyrus-user= cyrus
On Tue, 03 Feb 2004, Rob Siemborski wrote:
On Tue, 3 Feb 2004, Ian G Batten wrote:
On my small personal server, supporting a few family members, I run
Cyrus latest where possible. I was running 2.1.16 with every database
set to skiplist. Upgrading to 2.2.3 with the same ./configure
On Tue, 03 Feb 2004, Rob Siemborski wrote:
On Tue, 3 Feb 2004, Ian G Batten wrote:
It's in changes.html. It didn't leap out at me when I switched, so
perhaps it might be worth making louder noises about.
Actually, I was referring to install-upgrade.html.
I'm not sure where we'd put
On Tue, 03 Feb 2004, Rob Siemborski wrote:
On Tue, 3 Feb 2004, Ian G Batten wrote:
I thought that meant ``no change''. Looking at it more closely, the
default format for deliver.db is still Berkeley, hence my problem. I
think it would be worth being more specific than ``the default
We've not upgraded for ages, and we must do so. But this should set an
interesting lower floor on the amount of hardware you need.
On Tue, 02 Dec 2003, [EMAIL PROTECTED] wrote:
The version(s) of Cyrus you are running
1.6.24
The specs of the servers you have (cpu,mem,disk,os, etc)
Sun Ultra
It has always been my understanding that Cyrus isn't supported over NFS.
Clearly, getting locking working for the scenario of two machines
running Cyrus sharing /var/imap is Very Hard. However, what if I only
want to use _one_ instance of Cyrus, with failover (essentially, one Sun
sat in front of
On Fri, 04 Apr 2003, Phil Chambers wrote:
I am concerned that because Cyrus is a black box system which keeps track of its
own internal organisation we may have problems if we restore a disc from our
backups. It will take hours to do a backup and the files within the Cyrus structure
will
On Mon, 31 Mar 2003, Dave O wrote:
2 level hashing would work, but I don't know if Cyrus supports that. It
would most likely be trivial to implement.
eg spool/s/sm/user/smith
spool/s/m/user/smith?
ian
On Mon, 24 Feb 2003, Peter Lawler wrote:
For those who may have missed it,
http://www.openssl.org/news/secadv_20030219.txt
As a datapoint, although I could compile Cyrus against OpenSSL 0.9.7a
and it appeared to work for imaps, both apache 1.3.27 and sendmail
8.12.7 didn't work correctly
On Mon, 30 Dec 2002, Scott Smith wrote:
group and put cyrus and MTA user in it. Or, you can run LMTP over TCP (keep
it on loopback) with SASL.
I must confess that as a general rule I've given up on using AF_UNIX
sockets now that we're all aware that running all daemons as root is A
Bad Idea.
On Wed, 14 Mar 2001, Rob Tanner wrote:
privileges. Since all the mailboxes are owned by the Cyrus user, what
would be more secure of a system that just does mail delivery woulkd
be a hack to sendmail so that once it attaches to port 25 it drops root
and runs as the Cyrus user. Show me a
75 matches
Mail list logo