DNS Timeouts - I think the threshold for overload needs raising
Hi all! I'm operating a TOR Exit on dedicated hardware. The load average is low (0.07) and the network load is fine (120Mb/sec out of a 1Gb/sec link). Connections aren't being dropped, and for all I can see things are fine. However, on the TOR Metrics relay search, my exit consistently shows as "Overloaded" due to DNS failures.
From my logs, I never go above 2% in DNS failures and from the debugging I've done so far it looks like most of the failures are in IPv6 lookups not returning an AAAA record because the destination site has no AAAA configured.
Could whoever runs the "Overloaded" flag consider upping the overload limit to 2% or so? Or, can we tune out ipv6 and ipv4 failures separately maybe? This was never an issue until I configured my relay for v6. =========== Jan 15 19:57:45.000 [notice] General overload -> DNS timeouts (65) fraction 1.0619% is above threshold of 1.0000% Jan 15 20:07:45.000 [notice] General overload -> DNS timeouts (67) fraction 1.0124% is above threshold of 1.0000% Jan 15 20:17:45.000 [notice] General overload -> DNS timeouts (71) fraction 1.1275% is above threshold of 1.0000% Jan 15 20:27:45.000 [notice] General overload -> DNS timeouts (94) fraction 1.4003% is above threshold of 1.0000% Jan 15 20:37:45.000 [notice] General overload -> DNS timeouts (72) fraction 1.2000% is above threshold of 1.0000% Jan 15 20:47:45.000 [notice] General overload -> DNS timeouts (75) fraction 1.2517% is above threshold of 1.0000% Jan 15 20:57:45.000 [notice] General overload -> DNS timeouts (70) fraction 1.1605% is above threshold of 1.0000% Jan 15 21:07:45.000 [notice] General overload -> DNS timeouts (92) fraction 1.6028% is above threshold of 1.0000% Jan 15 21:17:45.000 [notice] General overload -> DNS timeouts (73) fraction 1.3290% is above threshold of 1.0000% Jan 15 21:27:45.000 [notice] General overload -> DNS timeouts (73) fraction 1.2907% is above threshold of 1.0000% Jan 15 21:37:45.000 [notice] General overload -> DNS timeouts (65) fraction 1.0095% is above threshold of 1.0000% Jan 15 21:47:45.000 [notice] General overload -> DNS timeouts (83) fraction 1.2635% is above threshold of 1.0000%
Update your version of tor this is an old issue that has been fixed in the latest version. Thanks, John C.
On Jan 16, 2022, at 4:17 AM, AMuse <tor-amuse@foofus.com> wrote:
Hi all!
I'm operating a TOR Exit on dedicated hardware. The load average is low (0.07) and the network load is fine (120Mb/sec out of a 1Gb/sec link). Connections aren't being dropped, and for all I can see things are fine.
However, on the TOR Metrics relay search, my exit consistently shows as "Overloaded" due to DNS failures.
From my logs, I never go above 2% in DNS failures and from the debugging I've done so far it looks like most of the failures are in IPv6 lookups not returning an AAAA record because the destination site has no AAAA configured.
Could whoever runs the "Overloaded" flag consider upping the overload limit to 2% or so? Or, can we tune out ipv6 and ipv4 failures separately maybe?
This was never an issue until I configured my relay for v6.
===========
Jan 15 19:57:45.000 [notice] General overload -> DNS timeouts (65) fraction 1.0619% is above threshold of 1.0000% Jan 15 20:07:45.000 [notice] General overload -> DNS timeouts (67) fraction 1.0124% is above threshold of 1.0000% Jan 15 20:17:45.000 [notice] General overload -> DNS timeouts (71) fraction 1.1275% is above threshold of 1.0000% Jan 15 20:27:45.000 [notice] General overload -> DNS timeouts (94) fraction 1.4003% is above threshold of 1.0000% Jan 15 20:37:45.000 [notice] General overload -> DNS timeouts (72) fraction 1.2000% is above threshold of 1.0000% Jan 15 20:47:45.000 [notice] General overload -> DNS timeouts (75) fraction 1.2517% is above threshold of 1.0000% Jan 15 20:57:45.000 [notice] General overload -> DNS timeouts (70) fraction 1.1605% is above threshold of 1.0000% Jan 15 21:07:45.000 [notice] General overload -> DNS timeouts (92) fraction 1.6028% is above threshold of 1.0000% Jan 15 21:17:45.000 [notice] General overload -> DNS timeouts (73) fraction 1.3290% is above threshold of 1.0000% Jan 15 21:27:45.000 [notice] General overload -> DNS timeouts (73) fraction 1.2907% is above threshold of 1.0000% Jan 15 21:37:45.000 [notice] General overload -> DNS timeouts (65) fraction 1.0095% is above threshold of 1.0000% Jan 15 21:47:45.000 [notice] General overload -> DNS timeouts (83) fraction 1.2635% is above threshold of 1.0000%
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
Ah, thanks! I swear I looked through the forum archives and didn't see that. :P I'll see if the FreeBSD package maintainers are releasing an update soon. I'm trying to stick to maintained packages on the Tor box since I don't have a ton of time to tinker around with it. On Sun, Jan 16, 2022 at 6:14 AM John Csuti via tor-relays < tor-relays@lists.torproject.org> wrote:
Update your version of tor this is an old issue that has been fixed in the latest version.
Thanks, John C.
On Jan 16, 2022, at 4:17 AM, AMuse <tor-amuse@foofus.com> wrote:
Hi all!
I'm operating a TOR Exit on dedicated hardware. The load average is low (0.07) and the network load is fine (120Mb/sec out of a 1Gb/sec link). Connections aren't being dropped, and for all I can see things are fine.
However, on the TOR Metrics relay search, my exit consistently shows as "Overloaded" due to DNS failures.
From my logs, I never go above 2% in DNS failures and from the debugging I've done so far it looks like most of the failures are in IPv6 lookups not returning an AAAA record because the destination site has no AAAA configured.
Could whoever runs the "Overloaded" flag consider upping the overload limit to 2% or so? Or, can we tune out ipv6 and ipv4 failures separately maybe?
This was never an issue until I configured my relay for v6.
===========
Jan 15 19:57:45.000 [notice] General overload -> DNS timeouts (65) fraction 1.0619% is above threshold of 1.0000% Jan 15 20:07:45.000 [notice] General overload -> DNS timeouts (67) fraction 1.0124% is above threshold of 1.0000% Jan 15 20:17:45.000 [notice] General overload -> DNS timeouts (71) fraction 1.1275% is above threshold of 1.0000% Jan 15 20:27:45.000 [notice] General overload -> DNS timeouts (94) fraction 1.4003% is above threshold of 1.0000% Jan 15 20:37:45.000 [notice] General overload -> DNS timeouts (72) fraction 1.2000% is above threshold of 1.0000% Jan 15 20:47:45.000 [notice] General overload -> DNS timeouts (75) fraction 1.2517% is above threshold of 1.0000% Jan 15 20:57:45.000 [notice] General overload -> DNS timeouts (70) fraction 1.1605% is above threshold of 1.0000% Jan 15 21:07:45.000 [notice] General overload -> DNS timeouts (92) fraction 1.6028% is above threshold of 1.0000% Jan 15 21:17:45.000 [notice] General overload -> DNS timeouts (73) fraction 1.3290% is above threshold of 1.0000% Jan 15 21:27:45.000 [notice] General overload -> DNS timeouts (73) fraction 1.2907% is above threshold of 1.0000% Jan 15 21:37:45.000 [notice] General overload -> DNS timeouts (65) fraction 1.0095% is above threshold of 1.0000% Jan 15 21:47:45.000 [notice] General overload -> DNS timeouts (83) fraction 1.2635% is above threshold of 1.0000%
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
participants (2)
-
AMuse -
John Csuti