https://bugs.exim.org/show_bug.cgi?id=1797
Bug ID: 1797
Summary: Sender verify defer (DNS timeout?) after several days
running
Product: Exim
Version: 4.86
Hardware: x86-64
OS: Linux
Status: NEW
Severity: bug
Priority: medium
Component: ACLs
Assignee: jgh146exb@???
Reporter: alexandre@???
CC: exim-dev@???
Sender verification is enabled like so:
require verify = sender
domains = !+no_sender_verify_domains
After a few hours or days of running continuously, the sender verification will
start failing systematically for any received messages, even from domains which
are defined in the no_sender_verify_domains.
<mainlog redacted="true">
2016-02-15 08:54:03 H=(mail-ob0-f178.google.com) [209.85.214.178] sender verify
defer for <myfriend@???>: host lookup for mail.aei.ca did not complete
(DNS timeout?)
2016-02-15 08:54:03 H=(mail-ob0-f178.google.com) [209.85.214.178]
F=<myfriend@???> temporarily rejected RCPT <alexandre@???>:
Could not complete sender verify
2016-02-15 21:48:07 H=(my_apps_domain) [1.2.3.4] F=<app@my_apps_domain>
temporarily rejected RCPT <alexandre@???>: Could not complete sender
verify
2016-02-15 21:58:39 H=(my_apps_domain) [1.2.3.4] sender verify defer for
<app@my_apps_domain>: host lookup for mail.isp.ca did not complete (DNS
timeout?)
</mainlog>
I have one Exim instance at home configured with the smarthost router using
mail.isp.ca as the SMTP relay, and another instance on a Linode VPS configured
with dnslookup. The bug happens in both servers.
In this example, my_apps_domain is one of my apps sending me email alerts from
a non existent domain. So I added it to the no_sender_verify_domains list to
skip sender verification. When the bug occurs, even skipping sender
verification does not work.
my_friend@??? reports that he receives an SMTP mail delivery delayed
message when this happens, so the messages must be rejected temporarily, or the
SMTP connection times out after the MAIL FROM.
The problem will affect all incoming messages until the exim process is
restarted. Messages will then be accepted the next time sending SMTP servers
retry delivering their queued up mail.
I doubt there is an actual DNS timeout. The problem happens also on a Linode
VPS using different resolvers. We continuously use the Internet on multiple PCs
from the same subnet without DNS issues. When I detect the bug, I can resolve
any valid domain with ping or dig from the mail server.
--
You are receiving this mail because:
You are on the CC list for the bug.
