#23817: Tor re-tries directory mirrors that it knows are missing microdescriptors ----------------------------------------+---------------------------------- Reporter: teor | Owner: (none) Type: defect | Status: new Priority: Medium | Milestone: Tor: | 0.3.3.x-final Component: Core Tor/Tor | Version: Severity: Normal | Resolution: Keywords: tor-guard, tor-hs, prop224 | Actual Points: Parent ID: #21969 | Points: Reviewer: | Sponsor: ----------------------------------------+----------------------------------
Comment (by asn): Here is a draft of a suggested technical solution here which involves a cache tracking failed md downloads: 1) Introduce an `md_failure_cache` cache which tracks failed or missing mds downloads per relays. That is a cache with entries as follows: "We failed to fetch md <hash> from relays X, Y, Z". engineering XXX: how to figure out missing mds from received batches 2) When we want to fetch a batch of mds from relay X, check `md_failure_cache` to see if any of the batched mds is included in the cache for relay X. if that's the case, add relay X to the excluded list of nodes for dirservers, so that our directory subsystems skips relay X and moves to the next one in line. engineering XXX: is there an exclude list for dirserver downloads? 3) When we succeed in fetching an md, remove its entries from the cache (so that we retry in the future from relay X which is probably our top primary guard). 4) Finally, to address #23863, when we detect that we are missing mds from our primary guards, order an immediate download of those missing mds, so that the previous steps get done. -- Ticket URL: <https://trac.torproject.org/projects/tor/ticket/23817#comment:4> Tor Bug Tracker & Wiki <https://trac.torproject.org/> The Tor Project: anonymity online
_______________________________________________ tor-bugs mailing list tor-bugs@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs