Hello everyone!
Some of you might have read about relay operators being affected by a drop of bandwidth on their relays recently or might even have hit this problem themselves.
Bug 33775[1] has more details about what is going on (in short: sbws, our new bandwidth scanner, has still release critical bugs but the instances running it got recently the voting majority as Torflow is showing its age and the duct tape around it is wearing off). Roger forwarded a mail to this list yesterday as well about short term hacks to solve this problem[2].
Things are looking good at that front so far. We are optimistic that the situation is getting better substantially rather soon as some of our directory authorities are either switching right now to using Torflow data again or making progress in resolving their Torflow related issues that prevented them from running it.
We thought it would be helpful to give an update about our remaining sbws work as well. So, the good news here is that we aim to have all the critical bugs fixed by end of April (yes, this month :)). Right now we know of the following five:
#33350[3] #33775[4] #30735[5] #33009[6] #30719[7]
We have the last two in `needs_review` (the last one via #30905[8]) and are currently distributing the first three among developers working on sbws, so we can work in parallel on them.
There is the risk that behind those bugs some other critical bugs lurk, which we have not found yet. To mitigate that we'll switch one bandwidth authority to running sbws's git master branch so we detect those bugs as early as possible should they exist.
Finally, we are working on a proper transition plan away from Torflow to avoid mistakes we've made so far and provide as little friction as possible for all parties involved in the whole process.
Let us know if you have any questions. Either way, I hope I could get the point out that the current situation is worked on and temporal. There is no need to shut down relays as the result of the current issues. :)
Georg
[1] https://trac.torproject.org/projects/tor/ticket/33775 [2] https://lists.torproject.org/pipermail/tor-relays/2020-April/018337.html [3] https://trac.torproject.org/projects/tor/ticket/33350 [4] https://trac.torproject.org/projects/tor/ticket/33775 [5] https://trac.torproject.org/projects/tor/ticket/30735 [6] https://trac.torproject.org/projects/tor/ticket/33009 [7] https://trac.torproject.org/projects/tor/ticket/30719 [8] https://trac.torproject.org/projects/tor/ticket/30905
tor-relays@lists.torproject.org