Bad index files.

[Earl R Shannon]

Hello,

I've got a question about the behavior we see when we shut down the imap 
server processes. When the master gets killed and the imap processes 
start to go away the state of the index files in some of the user 
accounts seems to become corrupted for some users. While a small 
percentage doesn't seem to be a bother the fact that it happens at all 
is bothersome. And it consumes time at server startup to find and 
correct the affected accounts.

The index files suffixed with .NEW are left behind apparently in some
unrecoverable state. If not fixed with a reconstruct the user is unable
to access their account.

Is this a known bug with the version we are using ( 2.1.11 )? We are
running on Solaris 2.8 and 2.7. The 2.8 machines seem to be the ones
affected.

ID NIL command output:

name: Cyrus IMAPD
version: v2.1.11 2002/12/04 14:53:12
vendor: Project Cyrus
support-url: http://asg.web.cmu.edu/cyrus
os: SunOS
os-version: 5.8
environment: Cyrus SASL 2.1.10; Sleepycat Software: Berkely DB 3.2.9: 
(January 24, 2001); OpenSSL 0.9.7c 30 Sep 2003; CMU Sieve 2.2;
mmap = shared;
lock = fcntl;
nonblock = fcntl;
auth = krb (NCSU Krbcyrus);
idle  = poll;
mboxlist.db = db3;
subs.db = flat;
seen.db = flat;
duplicate.db = db3-nosync;
tls.db = db3.nosync

I've not seen any info in the wiki FAQ or the mail list archive.
Perhaps I've missed it. Point me to it if its there. Thanks.

Regards,
Earl Shannon

---
Home Page: http://asg.web.cmu.edu/cyrus
Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html