I have one pretty large relay, MIGHTYWANG which is an IP4/6 guard, dedicated hardware running on a 1Gb line uncontended. It is usually one of the top 5 relays by consensus weight but on the morning of 14th October it lost Guard status on account of losing the stable flag.
I checked logs, connectivity and server health - nothing unusual, everything is generally pretty bullet proof in and around the relay and it had been running for well over a year without a reboot - just the very occasional Tor daemon restart following upgrades but no such activity prior to the 14th.
So next I checked the consensus and I see that around half of the directory authorities seem to be not assigning the stable flag. See attached screenshot showing current consensus.
The peering to each of those relays seems OK from what I can see (IP4 and IP6) so any idea what gives?
I’ve got a MIGHTYWANG sitting here twiddling it’s thumbs because have the directory authorities don’t want to use it. Bit of a waste.
I had similar things happen a few years ago with one of my old relays; again no obvious reason, just seemed to be the a random whim of the directory authorities.
I’ve noticed a couple of other long term relays are in a similar position - is this some time of attack, deliberate action or just Tor magic?
I have one pretty large relay, MIGHTYWANG which is an IP4/6 guard, dedicated hardware running on a 1Gb line uncontended. It is usually one of the top 5 relays by consensus weight but on the morning of 14th October it lost Guard status on account of losing the stable flag.
I checked logs, connectivity and server health - nothing unusual, everything is generally pretty bullet proof in and around the relay and it had been running for well over a year without a reboot - just the very occasional Tor daemon restart following upgrades but no such activity prior to the 14th.
So next I checked the consensus and I see that around half of the directory authorities seem to be not assigning the stable flag. See attached screenshot showing current consensus.
The peering to each of those relays seems OK from what I can see (IP4 and IP6) so any idea what gives?
I've got a MIGHTYWANG sitting here twiddling it's thumbs because have the directory authorities don't want to use it. Bit of a waste.
I had similar things happen a few years ago with one of my old relays; again no obvious reason, just seemed to be the a random whim of the directory authorities.
I've noticed a couple of other long term relays are in a similar position - is this some time of attack, deliberate action or just Tor magic?
Wang
I operate gabelmoo and your relay seems to be unreachable via IPv6 from here. Here's a traceroute:
I have one pretty large relay, MIGHTYWANG which is an IP4/6 guard, dedicated hardware running on a 1Gb line uncontended. It is usually one of the top 5 relays by consensus weight but on the morning of 14th October it lost Guard status on account of losing the stable flag.
I checked logs, connectivity and server health - nothing unusual, everything is generally pretty bullet proof in and around the relay and it had been running for well over a year without a reboot - just the very occasional Tor daemon restart following upgrades but no such activity prior to the 14th.
So next I checked the consensus and I see that around half of the directory authorities seem to be not assigning the stable flag. See attached screenshot showing current consensus.
The peering to each of those relays seems OK from what I can see (IP4 and IP6) so any idea what gives?
I’ve got a MIGHTYWANG sitting here twiddling it’s thumbs because have the directory authorities don’t want to use it. Bit of a waste.
I had similar things happen a few years ago with one of my old relays; again no obvious reason, just seemed to be the a random whim of the directory authorities.
I’ve noticed a couple of other long term relays are in a similar position - is this some time of attack, deliberate action or just Tor magic?
I operate gabelmoo and your relay seems to be unreachable via IPv6 from here. Here's a traceroute:
Ping and traceroute to that IP don't reach for me either, from anywhere*, but
TCP connection to port 443 works. Perhaps you could recheck that too, on your
end?
I have one pretty large relay, MIGHTYWANG which is an IP4/6 guard, dedicated hardware running on a 1Gb line uncontended. It is usually one of the top 5 relays by consensus weight but on the morning of 14th October it lost Guard status on account of losing the stable flag.
I checked logs, connectivity and server health - nothing unusual, everything is generally pretty bullet proof in and around the relay and it had been running for well over a year without a reboot - just the very occasional Tor daemon restart following upgrades but no such activity prior to the 14th.
So next I checked the consensus and I see that around half of the directory authorities seem to be not assigning the stable flag. See attached screenshot showing current consensus.
The peering to each of those relays seems OK from what I can see (IP4 and IP6) so any idea what gives?
I've got a MIGHTYWANG sitting here twiddling it's thumbs because have the directory authorities don't want to use it. Bit of a waste.
I had similar things happen a few years ago with one of my old relays; again no obvious reason, just seemed to be the a random whim of the directory authorities.
I've noticed a couple of other long term relays are in a similar position - is this some time of attack, deliberate action or just Tor magic?
Wang
I operate gabelmoo and your relay seems to be unreachable via IPv6 from here. Here's a traceroute:
I think your UDP based traceroute is hitting my firewall and getting dropped but you do have a route to me - in fact your relay has a long term active connection to mine via IP6 right now:
tcp6 0 0 2a02:29d0:8008:c0de:bad:beef:::443 2001:638:a000:4140::ffff:189:41011 ESTABLISHED
So it isn't an IP6 issue from what I can see (although that was an issue about 18 months ago as a result of some temporary peering issues).
I checked all the DA relays on IP6 and IP4 and all have active connections to me via IP6 (where they support it) or IP4 so if it is a connectivity issue it must be transient and so far undetectable.
There is something else happening here but I don't know what yet.
thanks
Wang
···
On 29/10/21 19:04, Sebastian Hahn wrote:
On 29. Oct 2021, at 18:10, Mighty Wang <wang@mighty.wang> wrote:
Yes I saw your post on the day it happened and guessed that we are suffering from exactly the same issue that started at exactly the same time.
I couldn’t correlate the loss of stable flag with anything in recent Tor server releases but I am going to recheck those; I am currently working my way back through the consensus voting lists in the run-up to the 14th October to try and understand where the problem started,
I’ll report back here.
thanks
W
···
On 29/10/21 18:57, Eddie wrote:
Welcome to the club: Since Georg opened that (on my behalf) I too have lost the Stable flag. Cheers.
I have one pretty large relay, MIGHTYWANG which is an IP4/6 guard, dedicated hardware running on a 1Gb line uncontended. It is usually one of the top 5 relays by consensus weight but on the morning of 14th October it lost Guard status on account of losing the stable flag.
I checked logs, connectivity and server health - nothing unusual, everything is generally pretty bullet proof in and around the relay and it had been running for well over a year without a reboot - just the very occasional Tor daemon restart following upgrades but no such activity prior to the 14th.
So next I checked the consensus and I see that around half of the directory authorities seem to be not assigning the stable flag. See attached screenshot showing current consensus.
The peering to each of those relays seems OK from what I can see (IP4 and IP6) so any idea what gives?
I’ve got a MIGHTYWANG sitting here twiddling it’s thumbs because have the directory authorities don’t want to use it. Bit of a waste.
I had similar things happen a few years ago with one of my old relays; again no obvious reason, just seemed to be the a random whim of the directory authorities.
I’ve noticed a couple of other long term relays are in a similar position - is this some time of attack, deliberate action or just Tor magic?
39.577 ms 39.414 ms 39.363 ms