Subj : an interesting thing
To   : Peter Knapper
From : Roy J. Tellason
Date : Mon Oct 14 2002 05:06 am

Peter Knapper wrote in a message to Roy J. Tellason:

 PK> Hi Roy,

 RJT> But the ones that I'd also set to no more than a specified number 
 RJT> (a couple of the real high-volume ones) kept to no more than that 
 RJT> number,  even though SQPACK wasn't being run on them.  

 PK> This is normal and EXACTLY what happens here.

I figured it was.

 PK> When you set limits to the NUMBER of messages (EG I use a maximum 
 PK> number of Msgs between 100 - 1000 plus a maximum time of 1 year,) 
 PK> on a Squish message base, the Squish API ensures those limits are 
 PK> retained number is kept. What you really need to be aware of is 
 PK> that the actual disk SPACE consumed by the "deleted" messages is 
 PK> not "released" until SQPACK (or is equivalent) is run.

I figured that,  too.  I've set most of the message bases here for a time 
limit,  many to 30 days back when fido was seeing more traffic than it is now,  
these days I'm bumping that to 60 days,  and still my message bases are 
consuming a lot less in disk space than they used to.

 PK> This also means that once messages HAVE expired, the space they 
 PK> occupied is RE-USED for new messages. The net result is a slow 
 PK> increase in message base size as the FREE fragments of message 
 PK> base become too hold new messages to allow new message to be 
 PK> inserted over old ones, so the message base needs to expanded to 
 PK> allow the new ones to be inserted. The "problem" is similar to 
 PK> fragmentation of an old FAT drive, the message base becomes more 
 PK> and more inefficient over time...;-)

That's why I pack them as a part of maintenance here.  Though disk space is a 
lot less of a concern than it used to be.

--- 
 * Origin: TANSTAAFL BBS 717-838-8539 (1:270/615)

.