I realized that I can pull the backups for the last couple of weeks
and see what is different between them.  Interesting results.

On 01/15:

-rw------- 1     1272 Jan 15 20:04 bayes_journal
-rw------- 1 43421696 Jan 15 20:03 bayes_seen
-rw------- 1  5042176 Jan 15 20:03 bayes_toks

0.000          0          3          0  non-token data: bayes db version
0.000          0      25735          0  non-token data: nspam
0.000          0      47082          0  non-token data: nham
0.000          0     118870          0  non-token data: ntokens
0.000          0 1358129333          0  non-token data: oldest atime
0.000          0 1358305649          0  non-token data: newest atime
0.000          0 1360353964          0  non-token data: last journal sync atime
0.000          0 1358305654          0  non-token data: last expiry atime
0.000          0     172800          0  non-token data: last expire atime delta
0.000          0       1456          0  non-token data: last expire reduction 
count

On 01/22:

-rw------- 1    46704 Jan 22 19:59 bayes_journal
-rw------- 1 43663360 Jan 22 19:58 bayes_seen
-rw------- 1  5373952 Jan 22 19:58 bayes_toks

0.000          0          3          0  non-token data: bayes db version
0.000          0      26480          0  non-token data: nspam
0.000          0      50182          0  non-token data: nham
0.000          0     123870          0  non-token data: ntokens
0.000          0 1358734372          0  non-token data: oldest atime
0.000          0 1358910216          0  non-token data: newest atime
0.000          0 1360354143          0  non-token data: last journal sync atime
0.000          0 1358906886          0  non-token data: last expiry atime
0.000          0     172800          0  non-token data: last expire atime delta
0.000          0       1300          0  non-token data: last expire reduction 
count

On 01/29:

-rw------- 1     3624 Jan 29 20:07 bayes_journal
-rw------- 1 77619200 Jan 29 20:03 bayes_seen
-rw------- 1  5447680 Jan 29 20:07 bayes_toks

0.000          0          3          0  non-token data: bayes db version
0.000          0        276          0  non-token data: nspam
0.000          0      53159          0  non-token data: nham
0.000          0     145782          0  non-token data: ntokens
0.000          0 1359166043          0  non-token data: oldest atime
0.000          0 1359515005          0  non-token data: newest atime
0.000          0 1360354262          0  non-token data: last journal sync atime
0.000          0 1359515271          0  non-token data: last expiry atime
0.000          0     345600          0  non-token data: last expire atime delta
0.000          0       1477          0  non-token data: last expire reduction 
count

Clearly something blew up in the last week.

  01/15: nspam:25735 nham:47082 ntokens:118870
  01/22: nspam:26480 nham:50182 ntokens:123870
  01/29: nspam:276 nham:53159 ntokens:145782

I will restore to the 01/22 version and wait and see if it happens
again.  If it does then I suspect some rollover problem.  If it does
not then I suspect a transient file lock glitch of some sort.

Bob

Reply via email to