FC3 memory (evolution)

Scot L. Harris webid at cfl.rr.com
Thu Jan 27 15:55:39 UTC 2005


On Thu, 2005-01-27 at 10:17, Michael A. Peters wrote:
> On 01/27/2005 06:23:58 AM, Sharon Kimble wrote:
> 
> > I think that the culprit is 'spamassassin', with one 'spamd' being
> > spawned
> > for each email account that is being checked. Simplest solution is to
> > either not use spamassassin or not check so many email accounts. Next
> > simplest solution, is to increase them amount of ram if you can. That
> > way
> > it won't appear so sluggish and will speed everything else up as  
> > well.
> 
> Spamassassin is slow.
> If there is any way you can set up a different box to pop for you and  
> process the mail through spamassassin, and then get evolution to get  
> mail from that, your user experience will increase.
> 
> What I ended up doing with balsa, and you can do the same with  
> evolution, is I use fetchmail to get my mail, process it through  
> spamassassin, and deposit them into my Inbox. Balsa then checks for new  
> mail in my Inbox ( /var/spool/mail/username )
> 
> That way it happens wether I am actively checking my mail or not, and  
> then my filters filter based upon the headers SpamAssassin creates. It  
> has made using mail so much better to use. I do this with three mail  
> accounts (fetchmail grabs them all).
> 
> That's probabably the best way to do it if you (like me) can't run  
> spamassassin on a separate box.

That is good advice.  I also recommend, if possible, skipping
spamassassin processing on most mailing lists.  I have found very little
spam on these lists.  In my current setup at home I use the filter in
evolution to invoke spamassassin.  That filter is after my other filters
which move mailing list messages to different folders.  So only stuff
that has a high probability of being spam invokes spamassassin.  Prior
to doing that downloading email was very slow since spamassassin was
called on everything.  

-- 
Scot L. Harris
webid at cfl.rr.com

There's no such thing as a free lunch.
		-- Milton Friendman 




More information about the users mailing list