lots of corruption of delived.db file

[Eric Doutreleau]

I'm using cyrus2.3.8 on a RHEL4 machine and from time to time i got 
massive corruption of deliverd.db file

for example today

Mar 12 15:27:13 pasargades lmtp[31407]: skiplist: checkpointed 
/var/lib/imap/deliver.db (1282 records, 123104 bytes) in 5 seconds
Mar 12 15:27:18 pasargades lmtp[30946]: skiplist: recovered 
/var/lib/imap/deliver.db (1282 records, 123104 bytes) in 5 seconds
Mar 12 15:27:24 pasargades lmtp[30948]: skiplist: checkpointed 
/var/lib/imap/deliver.db (1289 records, 123800 bytes) in 6 seconds
Mar 12 15:27:28 pasargades lmtp[31407]: skiplist: recovered 
/var/lib/imap/deliver.db (1289 records, 123800 bytes) in 4 seconds
Mar 12 15:27:35 pasargades lmtp[30981]: skiplist: checkpointed 
/var/lib/imap/deliver.db (1300 records, 124788 bytes) in 7 seconds
Mar 12 15:27:40 pasargades lmtp[31677]: skiplist: recovered 
/var/lib/imap/deliver.db (1300 records, 124788 bytes) in 5 seconds
Mar 12 15:28:20 pasargades lmtp[31679]: skiplist: checkpointed 
/var/lib/imap/deliver.db (2541 records, 244756 bytes) in 11 seconds
Mar 12 15:28:26 pasargades lmtp[31675]: skiplist: recovered 
/var/lib/imap/deliver.db (2541 records, 244756 bytes) in 6 seconds
Mar 12 15:28:38 pasargades lmtp[30947]: skiplist: checkpointed 
/var/lib/imap/deliver.db (2578 records, 248632 bytes) in 11 seconds
Mar 12 15:28:45 pasargades lmtp[30951]: skiplist: recovered 
/var/lib/imap/deliver.db (2578 records, 248632 bytes) in 7 seconds
Mar 12 15:29:01 pasargades lmtp[31676]: skiplist: checkpointed 
/var/lib/imap/deliver.db (2622 records, 253440 bytes) in 15 seconds
Mar 12 15:29:08 pasargades lmtp[31678]: skiplist: recovered 
/var/lib/imap/deliver.db (2622 records, 253440 bytes) in 7 seconds
Mar 12 15:29:22 pasargades lmtp[31462]: skiplist: checkpointed 
/var/lib/imap/deliver.db (2699 records, 260880 bytes) in 12 seconds
Mar 12 15:29:31 pasargades lmtp[31678]: skiplist: recovered 
/var/lib/imap/deliver.db (2699 records, 260880 bytes) in 9 seconds
Mar 12 15:29:43 pasargades lmtp[30952]: skiplist: checkpointed 
/var/lib/imap/deliver.db (2706 records, 261556 bytes) in 12 seconds
Mar 12 15:29:49 pasargades lmtp[31679]: skiplist: recovered 
/var/lib/imap/deliver.db (2706 records, 261556 bytes) in 6 seconds
Mar 12 15:30:03 pasargades lmtp[30950]: skiplist: checkpointed 
/var/lib/imap/deliver.db (2720 records, 263040 bytes) in 14 seconds
Mar 12 15:30:11 pasargades lmtp[30981]: skiplist: recovered 
/var/lib/imap/deliver.db (2720 records, 263040 bytes) in 8 seconds

cyrus recover by itself but the problem is that when the deliver.db 
grows it take about one hour to recover and the delivery of mail is stopped.
Has someone already seen that and have an idea on how to solve the problem.

thank in advance for any help