RE: Database backend?

Title: Database backend? Hi,   I had an experience to implement mysql backend to store mailbox paths. The mysql backend was intended for special purpose. But it would help you to start to make your DBMS backend implementation.   Let me know if you want it. ;-)   Tawan Won   From: [EMAIL

[question] ipurge took too long time to purge mails from about 30,000 users.

We are using cyrus imap 2.1.9 on solaris 2.8 to serve about 30,000 users. In addition, we are using ipurge to periodically delete old mails When we executed ipurge, it took so much time and consumed all of CPU power. Our cyrus-imap configuration is 1. OS : solaris 2.8 2. mailbo

cyrus imap + mysql as backend db ?

I am considering to add mysql support as a backend of mboxlist DB. I have used skiplist and Berkeley DB so far. All of them have a serious contention problem. But, I want to know if it is a valuable work . Thanks in advance. Won

[Q] partition-balancing in mailbox creation

Hi there, I am thinking of automatic partition-balancing when creating mailbox. As of date, partition information should be given as a parameter of cyradm, otherwise default partition is used. When growing the number of mailbox users and partitions, it is getting difficult to decide which partitio

[question] using IMAP/SORT extension...

Hi, We are thinking of using the "SORT" extension which already implemented in Cyrus IMAP for our web-based mail client. Has any one experienced with the feature ? When I read IETF-DRAFT-SORT extension, I was confused with how to use the feature. What I guess to use the feature is : C: a

[Q] anybody has a benchmark result of cyrus IMAP ?

I was tested cyrus IMAP version 2.1.2 on SUN 420R . The 420R system is configured as following: - 4 450MHz SPARC CPU - 2G RAM - Fiber RAID channel adaptor - Solaris 2.7 (32bit) - UFS for mailboxes - CyrusIMAP changed to use skiplist for mailboxes.db

who has experiance with large number of mailboxes ?

Hi, I am trying to make about 500,000 mailboxes in a single server, which has 4CPU and 1G RAM. I am worry about disk access time in (1) lookup up mailbox path from mailbox db. (2) directory navigation to select a specific message file. I think I have to change the cyrus directory ha

A Question/Idea on using Berkeley DB in Cyrus IMAP

Hi, I have a question on the way to use Berkeley db in Cyrus. As you may see, some body posted on a problem that imapd hung. I found it several times whenever massively testing Cyrus with postfix/smtpstone with '-L' (lmtp) option. I think it would be a locking problem caused by Berkeley db engine

RE: transfer usermails to another machine

Hi, I think reconstruct would work. After transfering files, you may have to remove all cyrus.{index,cache,header), which would be in different format with other version. I tested reconstruct and fould it worked well. I also followed the source flow and believe it works. But you might lost any

RE: Running 2 Cyrus on one machine

I think we did it with cyrus. I have clustered two machines each of which runs cyrus in active/active way. When one machine goes down, the other has to run two instances of cyrus. To do so, we made each instance of cyrus listen on a specific IP address in config file. For example, 211.192.212.1:2

a question on cyrus-imap version 2.1 (mailbox creation)

Hi, I am testing cyrus imap 2.1beta. In the basic configuration, it works well. But, i found a different thing from the other imap server. for example MS Exchange Server. For example, MS Exchange server allows user to make a folder at the same level of INBOX. So, client program (like outlook