Constant DBERROR issues

[Jason Fesler]

I seem to have this problem pretty consistently:

Dec  9 10:37:48 heaven lmtpd[80568]: DBERROR db3: region error detected; run recovery.
Dec  9 10:37:48 heaven lmtpd[80568]: DBERROR: dbenv->open '/var/imap/db' failed: DB_RUNRECOVERY: Fatal error, run database recovery
Dec  9 10:37:48 heaven lmtpd[80568]: DBERROR: init /var/imap/db: cyrusdb error
Dec  9 10:37:48 heaven lmtpd[80568]: lmtpd: unable to init duplicate delivery database

This is after using a configure line of:
./configure  --with-dbdir=/usr/local/BerkeleyDB.3.3 --with-cyrus-prefix=/usr/cyrus2 --with-cyrus-user=cyrus --with-cyrus-group=cyrus --disable-sieve --with-auth=unix --with-openssl=/usr --enable-listext --enable-annotatemore --with-perl=/usr/local/bin/perl5 --with-libwrap --without-ucdsnmp --with-mboxlist-db=skiplist

Things are at least "up" - but definately spewing errors quite a bit.
"./ctl_cyrusdb -r" did not make the messages go away at all.

------------------------------------------------
On the bright side:  before rebuilding as above, it flat out started
locking up.  Something would die, and all the ltmpd's, pop3d's, and imapd's would hang
until master were restarted.  This appears to be a common problem :( .

Dec  6 10:20:42 heaven lmtpd[40652]: telling master 3
Dec  6 10:20:42 heaven lmtpd[40652]: lmtp connection preauth'd as postman
Dec  6 10:20:42 heaven lmtpd[40652]: duplicate_check: <20021206094751.H48964 at bart.bartsite.com> user.dlangley        0
Dec  6 10:20:42 heaven lmtpd[40652]: mystore: starting txn 2151872938
Dec  6 10:20:42 heaven lmtpd[40652]: DBERROR db3: PANIC: Invalid argument
Dec  6 10:20:42 heaven lmtpd[40652]: DBERROR: critical database situation
Dec  6 10:21:08 heaven lmtpd[40427]: telling master 2
Dec  6 10:21:08 heaven lmtpd[40427]: accepted connection
Dec  6 10:21:08 heaven lmtpd[40427]: telling master 3
Dec  6 10:21:08 heaven lmtpd[40427]: lmtp connection preauth'd as postman
Dec  6 10:21:08 heaven lmtpd[40427]: DBERROR: error fetching user.jfesler: DB_RUNRECOVERY: Fatal error, run database recovery
Dec  6 10:21:08 heaven lmtpd[40427]: DBERROR: error fetching user.jfesler: cyrusdb error
Dec  6 10:21:08 heaven lmtpd[40427]: telling master 1
Dec  6 10:21:09 heaven lmtpd[40428]: telling master 2
Dec  6 10:21:09 heaven lmtpd[40428]: accepted connection
Dec  6 10:21:09 heaven lmtpd[40428]: telling master 3
Dec  6 10:21:09 heaven lmtpd[40428]: lmtp connection preauth'd as postman
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR db3: region error detected; run recovery.
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR: dbenv->open '/var/imap/db' failed: DB_RUNRECOVERY: Fatal error, run database recovery
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR: init /var/imap/db: cyrusdb error
Dec  6 10:21:09 heaven lmtpd[40764]: lmtpd: unable to init duplicate delivery database
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR db3: environment not yet opened
Dec  6 10:21:09 heaven lmtpd[40428]: DBERROR: error fetching user.jfesler: DB_RUNRECOVERY: Fatal error, run database recovery
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR: opening /var/imap/mailboxes.db: Invalid argument
Dec  6 10:21:09 heaven lmtpd[40428]: DBERROR: error fetching user.jfesler: cyrusdb error
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR: opening /var/imap/mailboxes.db: cyrusdb error
Dec  6 10:21:09 heaven lmtpd[40764]: FATAL: can't read mailboxes file
Dec  6 10:21:09 heaven lmtpd[40764]: DBERROR db3: Database handles open during environment close

Looking at the list archives, I've only seen one patch for master, and
nothing really good overall about the various DBERROR and DB_RUNRECOVERY
messages.

Has anyone else come across these problems, and found ways to make things
behave?

Not that it _should_ matter, but..

OS: FreeBSD heaven.gigo.com 4.7-RELEASE-p2 FreeBSD 4.7-RELEASE-p2
cyrus: cyrus-imapd-2.1.11 and cyrus-sasl-2.1.9


-- 
 Jason Fesler, <jfesler at gigo.com> http://gigo.com/resume.html
 "You'll finish first - or not at all" - Death Race 2000