Re: retries of delivery failure

[Andy Poling <andy@globalauctions.com>]

On Mon, 9 Mar 1998, David Trueman wrote:
> We run zmailer 2.99.49p9 system with local delivery to a partition
> with disk quotas.  What I would like to happen when delivery is attempted
> to the mailbox of a user who has exceeded their quota is for a few fairly
> quick retries followed by only occasional retries (e.g. once an hour)
> for a couple of days and then give up.
> 
> Despite playing with parameters in scheduler.conf, I have not been able to
> achieve this -- the retries are very frequent, no doubt having a
> performance impact.  Any hints on how I can do this?

Back when I administered a large system where user disk quotas would cause
delivery failures, I made 2 changes to mailbox:

	1) I made it so that mailbox checked their quota and didn't even
		attempt delivery if the msg wouldn't fit.
	2) I made mailbox return a retryat response to the scheduler to
		schedule a retry much later than the scheduler would by
		default.

The second is what you probably want to do.  It won't cause quick retries at
first, but what good would that do anyway - if they're over quota, that's not
likely to change in a minute or so...

-Andy

Global Auctions
http://www.globalauctions.com