On Fri, 13 Aug 2004, Nathan Ollerenshaw wrote:
> But, to fix the problem for everyone, you could fork off a 'fake'
> delivery process which just reports back to the parent regarding quota
> usage? A bit heavy I know, but its better than letting the messages sit
> in the queue IMHO.
It would have to be fork + exec, in order to regain root privilege. Far
too heavy and kludgy, IMHO. If you don't want to write a daemon, you
could write a setuid program that did the job, and call it using ${run}.
>> The "right" way is probably for somebody to write a daemon that can
>> run as root and be accessed via a socket.
>
> Can I twist your arm somehow?
No. I have no time at present for peripheral activities.
> Failing a feature like that, what should I do with these messages
> queuing up due to full mailboxes? Can I make them bounce or is it ok to
> let them sit there for the timeout period and let them bounce later?
Add this line at the start of your retry rules:
* quota
That will bounce them without ever retrying.
> Some of the big mailservers you hit out there, hotmail etc, will tell
> you if a mailbox is full at RCPT time. I'd really like to be able to do
> that in a clean way.
People have wanted this for a long time.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
