Dir auths using 2x bandwidth in last week
https://atlas.torproject.org/#details/9695DFC35FFEB861329B9F1AB04C46397020CE... https://atlas.torproject.org/#details/F2044413DAC2E02E3D6BCF4735A19BCA1DE972... https://atlas.torproject.org/#details/BD6A829255CB08E66FBE7D3748363586E46B38... https://atlas.torproject.org/#details/74A910646BCEEFBCD2E874FC1DC997430F9681... https://atlas.torproject.org/#details/7EA6EAD6FD83083C538F44038BBFA077587DD7... all show a big increase in sent bytes starting at the end of July. It isn't growth in Tor users, since those have stayed relatively flat in the last two weeks. And the new rate seems to be the new normal -- it's showing no signs of going back to the old rate. I would assume it's outgoing directory stuff, since that's most of what dir auths do. Any guesses? --Roger
In July, Tor 0.3.0 became the most common relay version in the network, growing at quite a rapid rate: https://metrics.torproject.org/versions.html There doesn't seem to be any corresponding Tor Browser release in that timeframe: 3 July: https://blog.torproject.org/blog/tor-browser-702-released 8 August: https://blog.torproject.org/blog/tor-browser-704-released (Our estimates suggest that half the load on directory authorities is from relays, and half is from clients.) I wonder if the new guard selection algorithm, or some other relay change, is causing relays to download more descriptors from more directory authorities? T -- Tim Wilson-Brown (teor) teor2345 at gmail dot com PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B ricochet:ekmygaiu4rzgsk6n ------------------------------------------------------------------------
On Mon, Aug 21, 2017 at 01:56:16PM +1000, teor wrote:
2 July is when deb.torproject.org switched to 0.3.0. nusenu noted it here: https://twitter.com/nusenu_/status/884128686764687361 I added a line already to https://trac.torproject.org/projects/tor/wiki/doc/MetricsTimeline
We merged a fallback directory mirror change to 0.2.8 and later in mid-May, but I think that's too far back. (The actual releases were up to a month later.) https://trac.torproject.org/projects/tor/ticket/21564 If we want to reduce the load on authorities, we can backport this change, which makes clients try fallbacks before authorities: https://trac.torproject.org/projects/tor/ticket/17750 It's been in master (0.3.2) for almost 2 months now. T -- Tim Wilson-Brown (teor) teor2345 at gmail dot com PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B ricochet:ekmygaiu4rzgsk6n ------------------------------------------------------------------------
In July, we went from 1% of fallback directory mirrors being down, to about 10% of fallback directory mirrors being down. See "Fallback Directories Running, Past 90 Days" on: https://consensus-health.torproject.org/graphs.html (Needs JS, SVG) This means that 10% of client requests go straight to directory authorities. For the other 90%, the first fallback and first directory authority request happen simultaneously, and the authority often wins (and gets to supply the client's first consensus). Deploying #17750 in 0.3.2 (or backporting it to earlier releases) will fix both these issues:
T -- Tim Wilson-Brown (teor) teor2345 at gmail dot com PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B ricochet:ekmygaiu4rzgsk6n ------------------------------------------------------------------------
On Wed, Aug 09, 2017 at 11:36:27PM -0400, Roger Dingledine wrote:
Well, whatever it was, it settled out -- back to normal -- once August ended. https://atlas.torproject.org/#details/9695DFC35FFEB861329B9F1AB04C46397020CE... --Roger
On Sun, Sep 17, 2017 at 07:32:13PM -0400, Roger Dingledine wrote:
The start and end of the doubling of dirauth bandwidth coincide with a temporary doubling or tripling of reported users from the Netherlands: https://people.torproject.org/~dcf/metrics-country.html?start=2017-07-01&end... There is a somewhat similar pattern in the Seychelles, Lithuania, and Romania, all of which had unexplained increases at the same time as the Netherlands. https://lists.torproject.org/pipermail/metrics-team/2017-August/000428.html https://people.torproject.org/~dcf/metrics-country.html?start=2017-07-01&end... https://people.torproject.org/~dcf/metrics-country.html?start=2017-07-01&end... https://people.torproject.org/~dcf/metrics-country.html?start=2017-07-01&end...
participants (3)
-
David Fifield -
Roger Dingledine -
teor