Re: [EXIM] check_spool_space not working?

Página superior
Eliminar este mensaje
Responder a este mensaje
Autor: Karsten Thygesen
Fecha:  
A: Philip Hazel
Cc: exim-users
Asunto: Re: [EXIM] check_spool_space not working?
>>>>> "Philip" == Philip Hazel <ph10@???> writes:

Philip> On 30 Nov 1998, Karsten Thygesen wrote:
>> It seems (from reading accept.c) that exim is using the bfree ffree
>> vfstat entries. Is that the right way to do it? The favail and
>> bavail returns the files/space available *to non root users*


Philip> Exim is indeed normally not running as root when receiving a
Philip> message (though some sites do run it as root). Would any body
Philip> else on the list care to comment on whether favail/bavail
Philip> should be used? Are they standard fields that are available in
Philip> all the OS Exim supports?

Solaris is OK.

>> And second, then I would like to submit a request for an option to
>> check for free diskspace in the /var/mail spool^H^H^H^H^H partition
>> :-)


Philip> Where mail is delivered is not a single special partition in
Philip> Exim. You could have lots of different partitions in use. The
Philip> only possible specification would have to be something messy
Philip> like

Philip> check_partition_space = 100M/partition1 : 200M/partition2 :
Philip> ...

Philip> and I'm not sure how helpful that really is, even with just
Philip> one partition. You are wanting to shut down the processing of
Philip> *any* mail when the partition into which local deliveries are
Philip> made gets full. Messages that don't have local deliveries, or
Philip> messages to local users that get forwarded somewhere else are
Philip> not going to use space in that partition, so why block them?

Hmm.. Or perhaps it should be an transport-option? Like in:

local_delivery:
driver = appendfile;
file = /var/mail/${local_part},
check_free_space = /var/mail:100M


and then hold back (in the spool) all mails to that transport? Then
mails to other transports could continue... A kind of fine-granulated
conditional delivery...

Karsten

--
*** Exim information can be found at http://www.exim.org/ ***