Re: [exim] exim bug(?)

Philip Hazel wrote:
PH> On Mon, 2 Jan 2006, Vladimir Sharun wrote:

>> Ahha, looks like Philip want for this:
>> Dec 30 01:13:52 dun exim[65422]: 1Es6rr-000H1A-Fo == *****@ukr.net <***@*******>
>> R=dnslookup T=remote_smtp defer (0): SMTP error from remote mail server after end of data:
>> host mxs.ukr.net [212.42.65.66]: 421 too many messages in this connection

PH> That is a "defer" line, meaning Exim will try again later.

>> greppin' exim logfile on qID 1Es6rr-000H1A-Fo gaves full users list. Processing lines
>> was started from =>, then lot of ->'s, then lot of =='s, then '=>' and again lot of ->'s
>> and in circle.

PH> Please shows us *all* the *exact* log lines that match 1Es6rr-000H1A-Fo,

I can't send them all, it's a few tens of megabytes of output (hundreds of
thousands RCPT).

PH> preferably not obfuscated (you can obfuscate the local parts if you
PH> must, but try to keep them distinguishable so we can see which is which)
PH> and then we can tell you exactly what happened to the message from the
PH> sending end. The first such line should be a <= line and the last one
PH> should be "Completed".

PH> If you read the relevant chapter on log lines in the manual, it will
PH> tell you that => and -> lines are deliveries for those recipients,
PH> whereas == is a defer that means Exim will try later.

Of course I read relevant chapter 4 years ago, and I understand, that
=='s mean that these messages must be keepd is spool until all the
recipients are delivered or message will be canceled in other ways, but
message isn't canceled and deferred recipients not placed in queue for
further resending.

Why I'm asking: is there any possibilities to check, is exim will
resend any of tempfailed after data messages in expanded via "include"
mailing lists ? I think there is some error in exim algorhytm, but
I'm not clear, that I can be wrong here.

I'll put some testing later (I think tomorrow) on subject and will
send further information related to subject.

--
UKR.NET Postmaster