| Subject: | Ran out of accept retries! (UDP - DNS Proxy) |
Hi there!
We use this great module in a DNS Proxy, relaying queries from one interface to another.
Going through the syslog on our production machines, I've noticed, that for every UDP DNS Query, we get a "Ran out of accept retries!" entry.
Problem with the particular entry is, that it's being raised as a warning in syslog, which then triggers the Server Admins to send us tickets, asking to fix it.
I had a quick look at the "sub accept" in Server.pm, which seems to be triggering this message.
From my point of view it looks like "$prop->{'client'} = $sock;" in the UDP recieve part of the above mentioned sub, is not set to anything that can be evaluated to true further down the road at "return 1 if $prop->{'client'};".
This makes it loop until $retries is reached and then exits, raising the "Ran out of accept retries!" message.
Please let me know whether I'm on the right track here or not...
Unfortunately I've got very limited possibilities to test in our environment.
Thanks & Regards!
-----------------------------
Two example syslog entries:
/var/log/messages:
Oct 3 11:35:14 ***** named-proxy[17526]: [ID 702911 daemon.notice] UDP ***.***.***.50:21601 ****.********.net A = (56 bytes)
Oct 3 11:35:14 ***** named-proxy[17527]: [ID 702911 daemon.notice] UDP ***.***.***.20:19864 ****.****.net A = (56 bytes)
Oct 3 11:35:14 ***** named-proxy[17524]: [ID 702911 daemon.warning] Ran out of accept retries!
Oct 3 11:35:14 ***** named-proxy[17525]: [ID 702911 daemon.warning] Ran out of accept retries!
-----------------------------
Versions:
$ perl -v
This is perl, v5.8.4 built for sun4-solaris-64int
(with 33 registered patches, see perl -V for more detail)
$ uname -a
SunOS ***** 5.10 Generic_144488-06 sun4v sparc SUNW,Sun-Fire-T200
Server.pm,v 1.153 2012/06/20 22:44:58 rhandom