Opened 8 weeks ago

Last modified 7 weeks ago

#33775 new defect

sbws measures some relays 100x lower than Torflow

Reported by: teor Owned by:
Priority: High Milestone: sbws: 1.1.x-final
Component: Core Tor/sbws Version:
Severity: Major Keywords: sbws-majority-blocker, sbws-roadmap
Cc: juga Actual Points:
Parent ID: #33121 Points: 2
Reviewer: Sponsor:

Description

Hi,

I've received multiple reports from relay operators who are concerned their relays are being measured too low.

At the moment, we only have 2 Torflow instances in the network, and 3 sbws instances. So bad sbws measurements can lower any relay's bandwidth, regardless of the Torflow measurement.

Here is one report:
https://lists.torproject.org/pipermail/tor-relays/2020-March/018321.html

At the moment, the bandwidths are:

moria1 494
maatuska 4
longclaw 5

https://consensus-health.torproject.org/consensus-health-2020-03-31-21-00.html#6E1DA4C0B0C05FB721B42329C47A20DA22908AEB

I have also received a similar report privately from a large relay operator.

This bug could be related to #33009 or #30733.

Child Tickets

TicketStatusOwnerSummaryComponent
#33824newrelay measures low / high on same network and same serverCore Tor/sbws
#33871needs_reviewjugaScale exactly as torflow does?Core Tor/sbws

Change History (5)

comment:1 Changed 8 weeks ago by teor

Summary: sbws measures some relays 100x less than Torflowsbws measures some relays 100x lower than Torflow

comment:2 Changed 7 weeks ago by teor

This bug could also be related to #30735.

Also note that:

moria1 and faravahar measure 300-1000 more relays:
https://consensus-health.torproject.org/graphs.html#bwauthstatus

And give higher bandwidths for 3700-4200 more relays:
https://consensus-health.torproject.org/graphs.html#bwauthgraphs

comment:3 Changed 7 weeks ago by teor

Priority: MediumHigh

These operators have complained about low bandwidths:

6E1DA4C0B0C05FB721B42329C47A20DA22908AEB
https://lists.torproject.org/pipermail/tor-relays/2020-March/018321.html

7AC93EC4B3725ABF7E09A903BC4A5C5C1E36517F
https://lists.torproject.org/pipermail/tor-relays/2020-April/018326.html

0C0388BC552D7AC903F49390A40244FD3FFF44A2
https://lists.torproject.org/pipermail/tor-relays/2020-April/018327.html

https://metrics.torproject.org/rs.html#aggregate/all/as:62744
5-10% of exit weight

https://metrics.torproject.org/rs.html#aggregate/all/as:AS205100%20nifty
3-25% of exit weight

Some operators are shutting down their relays because of this issue. It's getting really serious.

comment:4 Changed 7 weeks ago by teor

#33824 has a list of relays with huge measurement differences.

But these relays are all in the same data center:
https://metrics.torproject.org/rs.html#aggregate/all/as:AS205100%20nifty

comment:5 Changed 7 weeks ago by juga

Parent ID: #33121

Make it child of the ticket to deploy sbws in all bwauths since we can't until this is fixed.

Note: See TracTickets for help on using tickets.