We also need to do this fix because authorities and clients can have different views of the Running relays.
Trac: Version: sbws: unspecified to sbws: 1.1.0 Severity: Major to Critical Summary: Should sbws vote for relays that aren't Running? to Make sbws vote for all measured relays, even if they are not Running Milestone: sbws: unspecified to sbws: 1.1.x-final Priority: High to Very High
sbws only reports bandwidths for Running relays, but torflow reports bandwidths for all relays it has recently measured.
So, is torflow reporting bandwidth for all the relays seen in the last 5 days even if they were not in the consensus?. What about the relays for which there's no descriptors?, just use the last descriptor seen?.
sbws only reports bandwidths for Running relays, but torflow reports bandwidths for all relays it has recently measured.
So, is torflow reporting bandwidth for all the relays seen in the last 5 days even if they were not in the consensus?
Yes.
What about the relays for which there's no descriptors?, just use the last descriptor seen?.
Why is sbws missing the descriptors for some relays?
The authorities and directory mirrors should have them?
See bug legacy/trac#30733 (moved).
But yes, if we can't get the descriptor in the consensus (or if the relay is not in the consensus), just use the most recent descriptor data we have.
Trac: Summary: Make sbws vote for all measured relays, even if they are not Running to Make sbws vote for all measured relays, even if they are not Running / not in the consensus