Have you tried -autocdb? It offers much better performance than -autodbm, and might be a simpler and more reliable alternative to what you're trying to do below.
What are the hardware specs on this server of yours BTW? "Gregory Malsack" <[EMAIL PROTECTED]> writes: > Hello All, > > I recently had a server crash because the TMDA processes were taking too > much time and too much ram. I was using text file whitelists and > blacklists. I had a little over 12,000 email addresses between the > blacklist and whitelist files. I tried the -autodbm flag in the filter > file and this didn't help enough. I then did some additional research on > tmda.net and found configuration directives to attach the tmda program > to an sql backend. This was awesome! The average time to scan a message > was 34 seconds, the processor load was 100%, and the average amount of > ram was 18mb. After I implemented the sql backend, the average time to > scan a file is down to 3 seconds, processor load is at 0.8%, and ram is > 5mb. This resolves my problems, however now I have a new problem. I have > the append directives still adding to a text whitelist and blacklist > which is checked after the sql lists, and every night the lists are > copied to the sql databases and cleared from the text lists. This is > working, however some people and only some people are having to confirm > everyday. Some people are allowed through based on an sql whitelist > match and some (who are in the sql whitelist) are not. > > Can anyone tell me how debug this? I have the tmda_error.log and > tmda_in.log files enabled, however the actual sql query is not output to > the log files just the sql syntax I have in the filter_in file. > > Thanks in advance, > Greg _____________________________________________ tmda-users mailing list (tmda-users@tmda.net) http://tmda.net/lists/listinfo/tmda-users
