Re: MCI Cache / ESMTP facilities?
- From: JonB <jfretby@xxxxxxxxxxxxxx>
- Date: Fri, 23 May 2008 15:12:15 -0700 (PDT)
On May 23, 12:12 am, p...@xxxxxxxxxxxx (Per Hedeland) wrote:
confMCI_CACHE_SIZE=1
confMCI_CACHE_TIMEOUT=0
In order to enable the HostStatusDirectory facility?
What makes you think that they have anything to do with
HostStatusDirectory? They don't, AFAIK, but refer to the in-memory
connection cache of a running sendmail process.
The HostStatusDirectory option is disabled automatically if you
disable the MCI connection cache. Sendmail warns you if you try to use
HostStatusDirectory and have the MCI cache size set to zero:
"
Warning: HostStatusDirectory disabled with ConnectionCacheSize = 0
"
Does the remote send the word "SIZE" without a limit? That's valid too,
see RFC 1870. Otherwise you would appear to have found a bug in
sendmail, I find it extremely unlikely that this particular bug exists
though.
I can't see that the other side sent 'SIZE' at all.
Sending SIZE=xxxx on MAIL being related to the connection cache
settings? Sounds extremely unlikely too, but if there is a bug, who
knows - e.g. after having speculated that there is a bug, one could
speculate that the bug consists of sendmail getting confused about which
of the servers in the connection cache that supports the SIZE extension.
But since you're reducing the cache size to 1 from the default of 2,
that ought to make the speculative bug *less* likely to be triggered.
I was more concerned that having to set the Cache Connection time out
to zero (which should negate having the Connection Cache size set to 1
(which keeps the HostStatusDirectory option 'happy') might be an
issue.
I'd guess the MCI cache has/would cache the available facilities, as
sendmail only does a NOOP (and can do little else) when re-using the
connection, so the initial greeting/facilities chatter would only be
seen once - hence my concern that setting the timout to zero (to get
us the HostStatusDirectory) might be triggering a bug, if it's not a
recommended way of getting persistant host status, with only single
'one at a time' connections.
Afterall, if the timeout is zero, it shouldn't be re-using a
connection at all - and should connect/re-negotiate all a fresh.
JB.
.
- References:
- MCI Cache / ESMTP facilities?
- From: JonB
- Re: MCI Cache / ESMTP facilities?
- From: Per Hedeland
- MCI Cache / ESMTP facilities?
- Prev by Date: Re: Solaris 8 - Configuring sendmail relay (NoAuth inbound -> SSL outbound)
- Next by Date: Re: Extra DNS lookups with local domain appended
- Previous by thread: Re: MCI Cache / ESMTP facilities?
- Next by thread: Failure Responses not sent
- Index(es):
Relevant Pages
|