http://issues.apache.org/SpamAssassin/show_bug.cgi?id=5203
[EMAIL PROTECTED] changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WORKSFORME
------- Additional Comments From [EMAIL PROTECTED] 2006-11-22 21:29 -------
(In reply to comment #2)
> Sorry I just realised that I did a typo: the bayes_toks is not 160M,
> bayes_seen
> is. bayes_toks is 10M.
aha. I misread what you sent. Sorry about that. :|
> Could you give me the commands for debugging?
For read vs write, use whatever tools you have on your system to do that kind of
thing (iostat, sar, etc.) As for what SA is doing during those times, that's a
bit harder. If the I/O only happens periodically, it's probably the Bayes
expiry. Learning vs scanning is possibly viewable in the log, though you could
disable learning for a little while and see what happens.
> Also, I thought the database file was called "bayes_seen", are you saying that
> it is called "DB_File", where is it located ? Sorry, the system was set up by
> my
> predecessor.
seen is the file w/ the message id's of the messages that've been learned. toks
is the actual database. DB_File is the perl module which lets SA use Berkeley
DB to do the DB files.
> What are the command for recompiling it ? Should I back it up first ? Can I
> easily roll back if something happens ?
Upgrading libdb (not part of SA) is a decent sized undertaking, and not really
appropriate for this forum. I said that as part of the toks 160M size thing,
which was my misreading, so ... While this still may be useful, I'd skip it for
now.
anyway, this sounds like an issue for the users@ list as opposed to specifically
a bug in SA, so I'm going to close the ticket for now. if through the
discussion it comes up that there's an actionable SA issue to address this, we
can reopen the issue.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
