John Dalbec wrote:
> Johann Spies wrote:
>
>> On Wed, Aug 04, 2004 at 10:14:17AM -0400, John Dalbec wrote:
>>
>>> Johann Spies wrote:
>>>
>>>
>>>> One of our campuses experience a power failure for about 2½ hours. I
>>>> would expect exim to keep the emails in the queue in accordance with
>>>> these retry rules:
>>>>
*duh* OK, try adding:
* timeout_connect F,2h,12m; G,16h,2h,1.5; F,2d,4h
as the first retry rule. (If you're getting local connection timeouts, trying
once a minute is probably excessive.)
HTH,
John
>>>> # If DNS lookups timeout, drop the mail after trying for a hour.
>
>
> The "timeout" keyword seems to apply to connection timeouts also:
>
> # timeout_connect_MX: connection timeout from a host obtained from an MX
> record
> # timeout_connect_A: connection timeout from a host not obtained from an
> MX record
> # timeout_connect: any connection timeout
> # timeout_MX: any timeout from a host obtained from an MX record
> # timeout_A: any timeout from a host not obtained from an MX record
> # timeout: any timeout
>
> There doesn't appear to be a keyword for DNS timeouts only. Since you're
> getting connection timeouts (as shown below), this rule is being applied.
> Maybe Phil can add that to the wishlist?
>
>>>>
>>>> * timeout F,1h,12m;
>>
>>
>>
>> No. The main campus, where the dns-servers are, was not affected.
>>
>>
>>
>
>> This clearly was a temporary error as shown in:
>>
>> $ sudo exinext ralf@???
>> Transport:
>> mail.belpark.sun.ac.za [146.232.96.209] error 110: Connection timed out
>> first failed: 04-Aug-2004 12:38:40
>> last tried: 04-Aug-2004 15:37:51
>> next try at: 04-Aug-2004 15:49:51
>> past final cutoff time
>>
>>
>>> If it's still happening, try deleting the retry database.
>>
>>
>>
>> I did a
>>
>> exim_tidydb -t 10m /var/spool/exim4 retry
>
>
> What about "rm /usr/local/exim/spool/db/retry*"?
> John
>
>> and
>> exim_tidydb -t 10m /var/spool/exim4 wait-remote_smtp
>>
>> but that did not make any difference.
>
>
>
> --
>
> ## List details at http://www.exim.org/mailman/listinfo/exim-users Exim
> details at http://www.exim.org/ ##
>
>
>
