https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6127
--- Comment #37 from Karsten Bräckelmann <[email protected]> 2009-06-11 16:26:27 PST --- (In reply to comment #36) > Still, it would be best if SA responds gracefully to DB-File blowing up and > logs meaningful error messages while continuing to process the message. Without really thinking about this -- I'm not sure if there is any possibility to do that. The perl process (or the lib) itself made an invalid memory reference. Not sure how to catch this if at all, and to continue skipping DB access. Hmm, now that I said this, there's the ls vs du discrepancy. An invalid memory reference, and a file that's reported to be 4 TB or 600-odd MB... Maybe the underlying issue is much more fundamental to the system. Need to think about this. Just don't have the time right now. And didn't even read all recent comments carefully. ;) > Karsten: We know AWL and Bayes will get big, perhaps it would be prudent to > wrap their DB_File calls in signal trapping and recover gracefully from major > failures? Bayes gets roughly about as big as it is set to in the config. It only keeps growing beyond borders, if the expiry run times out. In that case, the admin's choice of BDB backend was't the best... Growing Bayes DBs are an entirely unrelated topic. Unlike AWL, Bayes generally is bound by DB size, and does have an automatic expiry mechanism. -- Configure bugmail: https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug.
