Re: NFSv4, anyone?
Hi, Quoting Nikola Milutinovic [EMAIL PROTECTED]: Now, one question that is a bit off topic, why is NFSv4 locking so bad? Couldn't they (Powers that Be) fix it, so it behaves like a normal (albeit, networked) file system? As far as i know NFSv4 has the locking support, but nobody has tested it with cyrus till know. All shared filesystems have to deal with the problem that they need to cache informations to be fast and to get all changes to keep consistent. M.Menge Tel.: (49) 7071/29-70316 Universitaet Tuebingen Fax.: (49) 7071/29-5912 Zentrum fuer Datenverarbeitung mail: [EMAIL PROTECTED] Waechterstrasse 76 72074 Tuebingen smime.p7s Description: S/MIME krytographische Unterschrift Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
Hi, i have to correct some information Quoting Michael Menge [EMAIL PROTECTED]: - Shared filesystem Active/Passive This setup can add HA to a normal Cyrus server by storing the Mails and databases on a shared filesystem and monitoring the server with heartbeat. This setup should work with every shared filesystem and there should be no problems with bdb Since there are many cyrus proceses on one server accessing the files they still need the file-locking. M.Menge Tel.: (49) 7071/29-70316 Universitaet Tuebingen Fax.: (49) 7071/29-5912 Zentrum fuer Datenverarbeitung mail: [EMAIL PROTECTED] Waechterstrasse 76 72074 Tuebingen smime.p7s Description: S/MIME krytographische Unterschrift Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
Hi, Quoting Gary Mills [EMAIL PROTECTED]: On Tue, Nov 28, 2006 at 05:28:02PM +0100, Michael Menge wrote: The following shared Filesystems seem support the file-locking GFShttp://sources.redhat.com/cluster/gfs/ Lustre http://www.lustre.org/ NFSv4 http://www.nfsv4.org/ Polyserve CFS http://www.polyserve.com Veritas CFS http://www.symantec.com/enterprise/products/overview.jsp?pcid=1020pvid=209_1 Sun Cluster Filesystem also supports file locking and other POSIX semantics on a cluster. It's part of Sun Cluster, which is a free product. I know that this list is not complete so feel free to add more shared filesystems to the list. Maybe we should share oure expirience with these filesystems and discus advantages and disadvantages. M.Menge Tel.: (49) 7071/29-70316 Universitaet Tuebingen Fax.: (49) 7071/29-5912 Zentrum fuer Datenverarbeitung mail: [EMAIL PROTECTED] Waechterstrasse 76 72074 Tuebingen smime.p7s Description: S/MIME krytographische Unterschrift Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: NFSv4, anyone?
Hi, Quoting Nikola Milutinovic [EMAIL PROTECTED]: Now, one question that is a bit off topic, why is NFSv4 locking so bad? Couldn't they (Powers that Be) fix it, so it behaves like a normal (albeit, networked) file system? As far as i know NFSv4 has the locking support, but nobody has tested It has been tested but didn't work. See the posts from some days ago. It's either a configuration problem with the tests, an NFSv4 implementation problem with the tested Linux kernel, or something else. I didn't investigate why it didn't work, and it would also be nice if someone on Solaris could try it there. Simon it with cyrus till know. All shared filesystems have to deal with the problem that they need to cache informations to be fast and to get all changes to keep consistent. M.Menge Tel.: (49) 7071/29-70316 Universitaet Tuebingen Fax.: (49) 7071/29-5912 Zentrum fuer Datenverarbeitung mail: [EMAIL PROTECTED] Waechterstrasse 76 72074 Tuebingen Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Sendmail, virtualdomain, alias
Hello, We try to build a virtualdomain cyrus server with sendmail. Sendmail must manage aliases. Here's the goal architecture : _ the fqdn of our cyrus server is imap.domain.com _ sendmail on this server must accept messages for both domain1.domain.com domain2.domain.com dealing with aliases for both these domains _ cyrus has two virtual domains domain1.domain.com domain2.domain.com We didn't find any tips for this. Any suggestions ? Regards. Jerome N. Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Sendmail, virtualdomain, alias
Hi, We try to build a virtualdomain cyrus server with sendmail. Sendmail must manage aliases. Here's the goal architecture : _ the fqdn of our cyrus server is imap.domain.com _ sendmail on this server must accept messages for both domain1.domain.com domain2.domain.com dealing with aliases for both these domains _ cyrus has two virtual domains domain1.domain.com domain2.domain.com We didn't find any tips for this. Any suggestions ? I think we have your goal architecture running. But what are your questions exactly? We use Sendmail + Cyrus + Webcyradm with mysql. Our userdata is in the mysql database and a script puts them in the sendmail files like aliases and virtusers. There are also an local aliases file, which could managed manually for example for integrating new mailman mailinglists. The Server accepts mails for various domains. So what information did you need? bye Bjoern Burger Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Sendmail, virtualdomain, alias
Jerome Nenert schrieb: Hello, We try to build a virtualdomain cyrus server with sendmail. Sendmail must manage aliases. Here's the goal architecture : _ the fqdn of our cyrus server is imap.domain.com _ sendmail on this server must accept messages for both domain1.domain.com domain2.domain.com dealing with aliases for both these domains _ cyrus has two virtual domains domain1.domain.com domain2.domain.com We didn't find any tips for this. Any suggestions ? Regards. Jerome N. http://www.sendmail.org/tips/virtual-hosting.php http://cyrusimap.web.cmu.edu/imapd/install-virtdomains.html http://anfi.homeunix.net/sendmail/#cyrus Alexander Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Sendmail, virtualdomain, alias
On Wed, Nov 29, 2006 at 09:54:21AM +0100, Jerome Nenert wrote: Hello, We try to build a virtualdomain cyrus server with sendmail. Sendmail must manage aliases. Here's the goal architecture : _ the fqdn of our cyrus server is imap.domain.com _ sendmail on this server must accept messages for both domain1.domain.com domain2.domain.com dealing with aliases for both these domains _ cyrus has two virtual domains domain1.domain.com domain2.domain.com We didn't find any tips for this. Any suggestions ? Two variants for sendmail: 1. use procmail. 2. sendmail don't use domain part of address, when deliver over lmtp. Edit sendmail.cf for adding full headers to lmtp transport: sendmail.cf Mcyrusv2, ... S=EnvFromSMTP/HdrFromSMTP, R=EnvToSMTP/HdrToL, E=\r\n, ... sendmail.cf Also, you can make this changes over MAILER_DEFINITIONS in .mc file, if you friendly with m4. This example don't checked by me: sendmail.mc MAILER_DEFINITIONS Mcyrusv2, P=[IPC], F=lsDFMnqA@/:|SmwXzW, E=\r\n, S=EnvFromSMTP, R=EnvToSMTP/HdrToL, T=DNS/RFC822/X-Unix, W=120s, A=FILE /var/imap/socket/lmtp sendmail.mc HOWTO in russian is: http://www.binkd.spb.ru/howto/cyrus-sendmail.howto.txt WBR -- Dmitriy Kirhlarov OILspace, 26 Leninskaya sloboda, bld. 2, 2nd floor, 115280 Moscow, Russia P:+7 495 105 7247 ext.208 F:+7 495 105 7246 E:[EMAIL PROTECTED] OILspace - The resource enriched - www.oilspace.com Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
Michael Menge wrote: I know that this list is not complete so feel free to add more shared filesystems to the list. Maybe we should share oure expirience with these filesystems and discus advantages and disadvantages. What about Lustre? Anybody has tried this in combination with Cyrus-IMAP? Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: cyrus and global filtering
On Wed, 29 Nov 2006 01:57:44 +0700 (ICT) Johannes Egger [EMAIL PROTECTED] wrote: Dear list, I run cyrus 2.1.18 and postfix 2.1.5 on a linux debian box, 2.6.8 kernel. I use cyrdeliver as a delivery agent for postfix, procmail is not involved. I am wondering whther it is possible to use global filtering scripts with cyrus, similar to /etc/procmailrc. The reason I ask - maybe there are other ways to do this - is that I would like to filter mails tagged as spam by spamassassin into a user's spam folder, with a minimum of maintenance (e.g. I'd like to automatically create the folder and subscribe the user to it). I have been told and read on the web that sieve scripts only work on a per-user basis. My users use the squirrelmail avelsieve plugin to modify filtering rules. I have considered solving the problem using plus addressing, e.g. [EMAIL PROTECTED] but so far I have not been successful. Any help? Thanks, Johannes I use plus addressing, but with LMTP instead of deliver, so perhaps that's why it works for me and not for you. (I also use Exim, not postfix.) Phil. --- Phil Chambers ([EMAIL PROTECTED]) University of Exeter Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
Quoting Yuri Gorshkov [EMAIL PROTECTED]: Michael Menge wrote: I know that this list is not complete so feel free to add more shared filesystems to the list. Maybe we should share oure expirience with these filesystems and discus advantages and disadvantages. What about Lustre? Anybody has tried this in combination with Cyrus-IMAP? We want to build a active/active cyrus cluster with lustre but don't have a testenviroment yet. M.Menge Tel.: (49) 7071/29-70316 Universitaet Tuebingen Fax.: (49) 7071/29-5912 Zentrum fuer Datenverarbeitung mail: [EMAIL PROTECTED] Waechterstrasse 76 72074 Tuebingen smime.p7s Description: S/MIME krytographische Unterschrift Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: NFSv4, anyone?
On Wed, Nov 29, 2006 at 09:36:55AM +0100, Simon Matter wrote: It has been tested but didn't work. See the posts from some days ago. It's either a configuration problem with the tests, an NFSv4 implementation problem with the tested Linux kernel, or something else. AFAIK there is a known NFSv4 bug in kernel 2.6.18 in file creation. You should try again with 2.6.19 (when it is released). Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences - Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Captive mailbox in Cyrus IMAP?
Thanks to all for the suggestions. They were very good so I ignored
them. :)
I've patched imapd.c (cmd_login and cmd_authenticate) so that the
presence and contents of {config_dir}/captive/{username} indicate the
actual user that should be logged in (providing it begins with
disabled). So for example if /var/lib/imap/captive/smith contains
disabled-archiving, then when smith logs in, it will really be
taken as disabled-archiving is logging in. disabled-archiving has
previously been primed with a message and has an ACL of lr to prevent
updates.
I've tried swapping things back and forth with Apple Mail and
Mulberry and things seem to work OK. I don't know how this will work
with a murder.
If anybody wants the patch (pretty small) I can send it somewhere
appropriate.
Thanks.
-nik
Cyrus Home Page: http://cyrusimap.web.cmu.edu/
Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Beyond rtcyrus2 (sendmail integration)
I would like to know who is interested in helping to develop next versions of advanced Sendmail and Cyrus IMAP integration methods. I would like to further improve methods available at links below: http://anfi.homeunix.net/sendmail/rtcyrus2.html http://anfi.homeunix.net/sendmail/#cyrus I have created google group for the purpose. Join it if you are interested: http://groups-beta.google.com/group/cyrus-sendmail -- [pl2en: Andrew] Andrzej Adam Filip : [EMAIL PROTECTED] : [EMAIL PROTECTED] Home site: http://anfi.homesite.net/ Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Sharing Murder?
Hi! While skimming throught the 2.3.7 changes, I found: --klip-- Changes to the Cyrus IMAP Server since 2.2.x * Added support for unified and replicated Murders. A Murder no longer has to have discrete frontend and backend servers; any one unified server can both proxy and serve local mailboxes (proxy functionality in proxyd and lmtpproxyd has been merged with imapd and lmtpd respectively), or all replicated servers can serve the same mailboxes from a shared filesystem. The new mupdate_config option in imapd.conf is used to determine whether a Murder is using a traditional, unified or replicated configuration. --klip-- The part '... all replicated servers can serve the same mailboxes from a shared filesystem' attracted my attention. Is this the LB that is reported not to work with current murder/replication? Or is this something else? --Janne Peltonen Email adm, Univ of Helsinki Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: NFSv4, anyone?
Hi, Simon Matter wrote: I just tried the in-memory NFSv4 server from citi (in newpynfs); it is very very slow, and I had to lower the number of concurrent users in imaptest (to prevent 'stalling errors') - but it seems to work better! I don't get any locking problems in the logs. So it could be just the server on Linux that is causing problems... and it might work on a Solaris server or NAS like NetApp... That's why I said someone with Solaris could try it. I'm not surprised if it works there... I took some time to try the Linux nfs4 client (RHEL4) with a Solaris 10 nfs4 server (default install) instead of the RHEL4 server. I copied over my spool, started cyrus... and this is actually more dramatic. [EMAIL PROTECTED] dovecot-1.0.rc7]# ./imaptest Error: STORE failed: s NO System I/O error Error: STORE failed: s NO System I/O error Error: STORE failed: s NO System I/O error Error: STORE (\Deleted) failed: sd NO System I/O error Auth Logi Sele Fetc Fet2 Stor Dele Expu Appe Logo Disc 0 24 24 22 22531110 Error: APPEND failed: a NO System I/O error Error: APPEND failed: a NO System I/O error Error: STORE failed: s NO System I/O error Error: EXPUNGE failed: e NO System I/O error After a couple of these errors I got a kernel panic on my Linux client, repetitive. I'm really surprised that it works this bad. (No, it worries me. I hope I made some serious mistakes.) I should try solaris - solaris now, of course... (And NetApp, but we don't have such a filer, yet. Hmm, I actually wanted to use that with nfs4 for other purposes (too)...) Paul Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Sharing Murder?
On 29 Nov 2006, at 10:46, Janne Peltonen wrote: The part '... all replicated servers can serve the same mailboxes from a shared filesystem' attracted my attention. Is this the LB that is reported not to work with current murder/replication? Or is this something else? In the replicated mupdate_config, each backend (nothing to do with replication) is meant to serve data from a common mail spool, while maintaining a replicated copy of mailboxes.db. The changes to implement this IMAP_ENUM_MUPDATE_CONFIG_REPLICATED are isolated to imap/mupdate-client.c, effectively to just remove references to the backend server. So, this configuration would remove the various locking problems surrounding mailboxes.db, since each backend would be maintaining a separate copy. Any concerns related to locking in the mail spool would still be relevant, tho that locking is typically less esoteric. You'd need to have an mupdate master as well. :wes Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Captive mailbox in Cyrus IMAP?
Nik Conwell wrote: If anybody wants the patch (pretty small) I can send it somewhere appropriate. Please, I'd be interested! :) Ben Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Beyond rtcyrus2 (sendmail integration)
--On Wednesday, November 29, 2006 16:11 +0100 Andrzej Adam Filip [EMAIL PROTECTED] wrote: I would like to know who is interested in helping to develop next versions of advanced Sendmail and Cyrus IMAP integration methods. I would like to further improve methods available at links below: http://anfi.homeunix.net/sendmail/rtcyrus2.html http://anfi.homeunix.net/sendmail/#cyrus I have created google group for the purpose. Join it if you are interested: http://groups-beta.google.com/group/cyrus-sendmail Sounds good to me. Joseph Brennan Lead Email Systems Engineer Columbia University Information Technology Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Captive mailbox in Cyrus IMAP?
On Tue, 2006-11-21 at 12:35 -0500, [EMAIL PROTECTED] wrote: I could see this being useful as well for password expirations. ...or billing issues. MJ -- Miroslaw Psyborg Jaworski GCS/IT d- s+:+ a C++$ UBI$ P+++$ L- E--- W++(+++)$ N++ o+ K- w-- O- M- V- PS+ PE++ Y+ PGP t 5? X+ R++ !tv b++(+++) DI++ D+ G e* h++ r+++ y? Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
IMAP/POP traffic accounting
I believe there was some mention of this in the past, but I'm wondering whether there are any known efforts to implement network traffic accounting in the later versions of Cyrus, or if this is still something that hasn't been dealt with. I am pondering possibly tackling this, but I'm not really familiar with the Cyrus code and would rather not reinvent the wheel if it's already in the works. On the surface, this appears to be fairly straight forward, what with the protocol abstraction already in place (where recording could be done at or around the same place as telemetry logging) ... or is there something I'm missing about this idea that makes it significantly more complicated than it first appears? Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
Great summary Michael. I have a few questions on what a shared filesystem hardware CAN be: - iSCSI storage ? (cheap GigaEthernet SAN) according to Wikipedia http://en.wikipedia.org/wiki/ISCSI , only 1 iSCSI-client can be connected to 1 iSCSI-server (disk) at a time...so this does not allow for a shared FS ? - Fibre Channel SAN with NFSv4, GSF support - NAS with with NFSv4, GSF support - if GFS is mainly for Linux, any NAS or SAN should work ?? regards, Michael Menge wrote: - Shared filesystem Active/Passive This setup can add HA to a normal Cyrus server by storing the Mails and databases on a shared filesystem and monitoring the server with heartbeat. This setup should work with every shared filesystem and there should be no problems with bdb - Shared filesystem Active/Active In this setup all users are on all server. The mailboxes and databases are on a shared filesystem so that changes on one server are visible on all other servers. Cyrus depends on the file-locking of the filesystem. NFSv4, GFS, Lustre, and some other shared filesystem affirm that they have file-locking across cluster nodes. The sockets, lock- and pid-files have to stay on a local filesystem or have to be made unique across cluster nodes. BDB seems to have problems in this setup, because the changes of the lock in the mmaped files are not instantly on all clients/nodes and the use of shared memory. You have to compile Cyrus without bdb-support to get rid of the errors. This setup needs no extra servers, and all use the same configuration. In case of server crash the user use one of the other servers. The server can be replaced with an clone of one of the other servers. The storage should be too on a HA cluster to make the whole mailsystem ha. The Active/Active Shared filesystem setup is discussed controvers. This setup is not widely used and therefor not tested that mutch. But Dave McMurtrie and Scott Adkins have reported of sucsessfull installations The following shared Filesystems seem support the file-locking GFShttp://sources.redhat.com/cluster/gfs/ Lustre http://www.lustre.org/ NFSv4 http://www.nfsv4.org/ Polyserve CFS http://www.polyserve.com Veritas CFS http://www.symantec.com/enterprise/products/overview.jsp?pcid=1020pvid=209_1 M.Menge Tel.: (49) 7071/29-70316 Universitaet Tuebingen Fax.: (49) 7071/29-5912 Zentrum fuer Datenverarbeitung mail: [EMAIL PROTECTED] Waechterstrasse 76 72074 Tuebingen Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html -- Marcelo Maraboli Rosselott Jefe Area de Redes y Comunicaciones (Network UNIX Systems Engineer) Ingeniero Civil Electronico, CISSP (Electronic Engineer, CISSP) Direccion Central de Servicios Computacionales (DCSC) Universidad Tecnica Federico Santa Mariaphone: +56 32 2654071 Chile. http://www.usm.cl http://elqui.dcsc.utfsm.cl Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
- iSCSI storage ? (cheap GigaEthernet SAN) according to Wikipedia http://en.wikipedia.org/wiki/ISCSI , only 1 iSCSI-client can be connected to 1 iSCSI-server (disk) at a time...so this does not allow for a shared FS ? Yes, much like only one host can connect to an FC LUN at once. :) iSCSI doesn't actually *prevent* you from doing this. John Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: NFSv4, anyone?
At Wed, 29 Nov 2006 08:52:08 +0100, Michael Menge wrote: As far as i know NFSv4 has the locking support, but nobody has tested it with cyrus till know. All shared filesystems have to deal with the problem that they need to cache informations to be fast and to get all changes to keep consistent. NFSv4 has a better design and better protocols for implementing filesystem level locking, but that doesn't mean any particular implementation will be better, or even that it will have it. Take for example the *BSD's. They haven't had client-side NFS locking implementations even for NFSv3, except for some very recently. (NetBSD had excellent and quite robust server-side NFSv3 locking a relatively long time ago, but still has no real client-side locking implementation. I.e. it can be used to implement a robust NFSv3 file server, but don't expect to do database work on any diskless NFS client implemented using NetBSD, yet.) -- Greg A. Woods H:+1 416 218-0098 W:+1 416 489-5852 x122 VE3TCP RoboHack [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED] Secrets of the Weird [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Cyrus HA LB cluster summary
At Wed, 29 Nov 2006 15:24:44 -0300, Marcelo Maraboli wrote: I have a few questions on what a shared filesystem hardware CAN be: - iSCSI storage ? (cheap GigaEthernet SAN) according to Wikipedia http://en.wikipedia.org/wiki/ISCSI , only 1 iSCSI-client can be connected to 1 iSCSI-server (disk) at a time...so this does not allow for a shared FS ? A shared FS is something completely different than storage that can be divvied up between multiple systems. A filesystem is just a construct of the operating system (a data structure that lends form and purpose to the data on some kind of external secondary storage), and thus for a shared filesystem the kernels of all the participants must share the same construct consistently and with integrity. I.e. if one system changes a block on the storage devices then it must be able to do so with integrity and perhaps also in such a way that the other systems will know about that change by the time they might need to reference that block. Networked block storage devices, such as Fibre Channel and iSCSI can lend themselves well to being used as components of a cluster system since they isolate points of failure related to the storage subsystem into a unit where they can be dealt with in a black-box way without the operating systems having to deal with it. (E.g. redundant network connections, mirrored controllers, mirrored storage, perhaps with the mirrors in locations remote from each other.) Of course you can also have networked storage connected to a single file server and then use a networked filesystem, just to complicate things. :-) Shared filesystems can also be implemented without shared storage though -- i.e. in much the way network filesystems are implemented. (Now why couldn't we have kept with the Multics unified view of storage?!?!?! Maybe someday we'll finally get back to it.) - if GFS is mainly for Linux, any NAS or SAN should work ?? I'm hoping that someone might find time/money/etc. to port GFS (or something similar) to *BSD (NetBSD in particular would be best from my perspective, especially since it runs on the most platforms) -- Greg A. Woods H:+1 416 218-0098 W:+1 416 489-5852 x122 VE3TCP RoboHack [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED] Secrets of the Weird [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: Captive mailbox in Cyrus IMAP?
At Wed, 29 Nov 2006 09:56:04 -0500,
Nik Conwell wrote:
Thanks to all for the suggestions. They were very good so I ignored
them. :)
I've patched imapd.c (cmd_login and cmd_authenticate) so that the
presence and contents of {config_dir}/captive/{username} indicate the
actual user that should be logged in (providing it begins with
disabled). So for example if /var/lib/imap/captive/smith contains
disabled-archiving, then when smith logs in, it will really be
taken as disabled-archiving is logging in. disabled-archiving has
previously been primed with a message and has an ACL of lr to prevent
updates.
That sounds like a nice and nasty little hack! It should work very fine
and it nicely separates the issues of showing the user a read-only
message (without worrying about decent MOTD support in the client, or
even that it's an IMAP client) from the issues of whether or not
incoming mail for the user is allowed to continue or is bounced, etc.
The only thing I don't like is use of a file in the filesystem as a flag
instead of using a proper configuration file whos contents can be
version controlled and more easily backed up, restored, documented,
centralized, shared, and understood. I.e. tables in files are much
better suited to this use. (unless perhaps you happen to have a brain
that is twisted into the kind of insanity that resulted in qmail and its
ilk :-))
(and don't worry about searching such a table -- a straight linear
search of a simple table in a normal text file will be fast enough and
quite obviously much simpler than anything more sophisticated, at least
until you've got many thousands of entries. Premature optimization is
the root of all evil in programming. -- C.A.R. Hoare, who said that a
very long time ago (in IT-years) and it's still 100% true, if not even
more so)
I've tried swapping things back and forth with Apple Mail and
Mulberry and things seem to work OK. I don't know how this will work
with a murder.
That could either be tricky, or relatively transparent so long as you
create the flag file on all backend machines.
Did you do anything about the seen state of the unread message? Maybe
I'm just ignorant of how a read-only inbox will behave though -- perhaps
the message will always appear new (and a POP client will always
download it anew too) and so nothing need be done special?
--
Greg A. Woods
H:+1 416 218-0098 W:+1 416 489-5852 x122 VE3TCP RoboHack [EMAIL PROTECTED]
Planix, Inc. [EMAIL PROTECTED] Secrets of the Weird [EMAIL PROTECTED]
Cyrus Home Page: http://cyrusimap.web.cmu.edu/
Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
some weirdness and broken error handling with sieve (in 2.2.12)
So I finally decided to try actually using SIEVE to allow me some mobility and separation between my INBOX sorting and my client(s). However my first experiment (the one taken directly from the install-sieve.html guide) ended in a minor disaster (because I did it live on my home mail server, not on my test system, sigh) when I didn't have the correct path configured for sendmail. (I thought I had corrected the hard-coded path for my installation, but I obviously hadn't.) Here's the log of what happened on the Cyrus side when the message to be rejected arrived: Nov 28 20:52:52 most lmtpunix[27427]: executed Nov 28 20:52:52 most lmtpunix[27427]: accepted connection Nov 28 20:52:52 most lmtpunix[27427]: lmtp connection preauth'd as postman Nov 28 20:52:52 most lmtpunix[27428]: FATAL: couldn't exec() sendmail Nov 28 20:52:52 most lmtpunix[27427]: sieve runtime error for woods id [EMAIL PROTECTED]: Reject: Sendmail process terminated normally, exit status 75 Nov 28 20:52:52 most lmtpunix[27427]: DBERROR : db4 Nov 28 20:52:52 most lmtpunix[27427]: DBERROR: error fetching [EMAIL PROTECTED]: Invalid argument Nov 28 20:52:52 most lmtpunix[27427]: duplicate_check: error looking up [EMAIL PROTECTED]/user.woods: cyrusdb error Nov 28 20:52:52 most lmtpunix[27427]: IOERROR: opening /var/spool/imap/stage./27427-1164765172-0: No such file or directory Nov 28 20:52:52 most lmtpunix[27427]: IOERROR: error unlinking file /var/spool/imap/stage./27427-1164765172-0: No such file or directory Nov 28 20:53:52 most master[21410]: process 27427 exited, signaled to death by 11 Nov 28 20:53:52 most master[21410]: service lmtpunix pid 27427 in READY state: terminated abnormally BTW, I don't have any problem with the duplicate DB normally -- it usually works just fine: Nov 29 15:58:53 most lmtpunix[22809]: dupelim: eliminated duplicate message to user.woods id [EMAIL PROTECTED] (delivery) The DBERROR lines above are the only ones in my log. From there on the LMTP service was dead. I had to stop and restart the master daemon. Worse yet all the deliver processes (I still use a pipe to deliver to store incoming mail) would just hang indefinitely (or at least for about an hour until I had the good sense to restart Cyrus from scratch). I tried to do a wee bit of debugging, of the latter issue and all I could see was that the deliver processes were sitting in a read(2) call on a unix socket that wasn't connected to anything. Maybe this latter issue is more a problem with the AF_LOCAL implementation on my server than anthing else, though I thought there was a timeout alarm to catch stuck LMTP connections I didn't think to try to get any clue of the state of the master daemon as I didn't realize that it had effectively stopped the LMTP service without warning. I tried a SIGHUP first, but that didn't help. Error handling of such an event is obviously not very robust yet. I also had some confusion as well as to whether de-activating the script allowed things to work again or not, though at this time I'm unsure as to whether I had restared the master daemon first or not. Maybe I'll try that again once I've got some other issues sorted out. BTW, this whole idea of generating new mail from SIEVE is bogus. The local mail service should NEVER EVER IN A MILLION YEARS ever generate new messages in response to incoming mail (think backscatter, unless if it had the kind of once-per-week protection built into the much older unix vacation(1)). The reject should happen within the LMTP transaction (and in the case of deliver it should result in a unique exit code). I'll be fixing that before I ever let my end users touch SIEVE, or at least use the reject or redirect actions, etc. I'm not even sure I want to give my users a real discard action -- I think I'll rewrite it to fileinto Junk or some such more undo-able action. -- Greg A. Woods H:+1 416 218-0098 W:+1 416 489-5852 x122 VE3TCP RoboHack [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED] Secrets of the Weird [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: NFSv4, anyone?
On Wednesday 29 November 2006 09:55, Paul Dekkers wrote: I took some time to try the Linux nfs4 client (RHEL4) with a Solaris 10 nfs4 server (default install) instead of the RHEL4 server. I copied over my spool, started cyrus... and this is actually more dramatic. What kernel version is RHEL4 using? If I am not mistaken, there are some fairly serious issued with RHEL4's kernel and NFSv4. Try making the server export it in sync mode and see if that does anything for you. wt -- Warren Turkal, Research Associate III/Systems Administrator Colorado State University, Dept. of Atmospheric Science Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
