Le jeudi, 9 jan 2003, � 14:10 Europe/Paris, Sam Varshavchik a �crit :
Are you sure ? I have a cluster of 6 machines using courier-imap under solaris... With about > 600 000 users, so I may be don't know what I am talking about...Xavier Beaudouin writes:Last CVS snapshot of courier-imap introduces a cache system.... but in a file... This is also a strange idea, let me explain what.
Writting the cache a in file on a maildir can give some races on a complex cases. Imagine you have a pop3 account shared with 3 users and have 3 pop3 server (with data on NFS) behind a lot balancer. How can all pop3 servers knows that current pop3 cache file is being modified.
You don't know what you're talking about.
We (my company) had such problems with FreeBSD and Netapp with NFS (more than 1Tb and 1 Millon mailboxes) and mailboxes... Even if we still use dot-file locks there is sometimes mailboxes corrupted...
So that maybe I don't know what I am talking about... So please tell me what I had missed ?
Ah ?No. Just be sure your rsize and wsize on your NFS mount is 8192... Because buffers are kindly hardcoded in source code which is baaaaad...
And you still don't know what you're talking about.
Maybe this has been changed between the old courier-imap we used in production and last ones. But effectively changing rsize and wsize from 32768 to 8192 magicly give us speed only solaris.... But I'll maybe out of date in current code...
Or this is maybe (as usual ?) specific to solaris and NFS.... ?
/Xavier
-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
_______________________________________________
courier-users mailing list
[EMAIL PROTECTED]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
