#21969: We're missing descriptors for some of our primary entry guards
--------------------------+------------------------------------
Reporter: asn | Owner:
Type: defect | Status: new
Priority: Medium | Milestone: Tor: 0.3.1.x-final
Component: Core Tor/Tor | Version:
Severity: Normal | Resolution:
Keywords: tor-guard | Actual Points:
Parent ID: | Points: 1.5
Reviewer: | Sponsor: SponsorU
--------------------------+------------------------------------
Comment (by s7r):
Right. I am getting it every 2 days or so. The server is 1gbps at an ISP
datacenter, so it's out of the question if it's on a slow connection
network.
I am running `0.3.1.0-alpha-dev (git-b1c7e5d8c0a648b1+f7afd80)` with
default guard context (no bridges).
This Tor instance hosts a hidden service, and when this happens it can no
longer build new rend circuits. However, let's think about it from another
perspective maybe the bug is actually fixed in terms of logic of requested
descriptors, but it's a race condition in how they are processed after
being downloaded / refreshed. I am saying this because, based on the
timestamp, at exactly the same time or immediately after (~1 second) it
complains that it is missing descriptors for some of our primary entry
guards, it comes back to reality and says We now have enough directory
information to build circuits. And after more time (minutes - is this
normal?) it confirms that it successfully opened a circuit. This is logged
two times, every time, with the very same timestamp (is this also normal?)
Check timestamps in these last events:
{{{
Apr 12 04:26:10 vm16 Tor[978]: Our directory information is no longer up-
to-date enough to build circuits: We're missing descriptors for some of
our primary entry guards
Apr 12 04:26:10 vm16 Tor[978]: I learned some more directory information,
but not enough to build a circuit: We're missing descriptors for some of
our primary entry guards
Apr 12 04:26:10 vm16 Tor[978]: We now have enough directory information to
build circuits.
Apr 12 04:27:19 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
Apr 12 04:27:19 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
}}}
{{{
Apr 17 12:04:16 vm16 Tor[978]: Our directory information is no longer up-
to-date enough to build circuits: We're missing descriptors for some of
our primary entry guards
Apr 17 12:04:16 vm16 Tor[978]: I learned some more directory information,
but not enough to build a circuit: We're missing descriptors for some of
our primary entry guards
Apr 17 12:04:16 vm16 Tor[978]: We now have enough directory information to
build circuits.
Apr 17 12:05:24 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
Apr 17 12:05:24 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
}}}
{{{
Apr 19 05:38:18 vm16 Tor[978]: Our directory information is no longer up-
to-date enough to build circuits: We're missing descriptors for some of
our primary entry guards
Apr 19 05:38:18 vm16 Tor[978]: I learned some more directory information,
but not enough to build a circuit: We're missing descriptors for some of
our primary entry guards
Apr 19 05:38:18 vm16 Tor[978]: We now have enough directory information to
build circuits.
Apr 19 05:42:12 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
Apr 19 05:42:12 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
}}}
{{{
Apr 21 00:48:20 vm16 Tor[978]: Our directory information is no longer up-
to-date enough to build circuits: We're missing descriptors for some of
our primary entry guards
Apr 21 00:48:20 vm16 Tor[978]: I learned some more directory information,
but not enough to build a circuit: We're missing descriptors for some of
our primary entry guards
Apr 21 00:48:21 vm16 Tor[978]: We now have enough directory information to
build circuits.
Apr 21 00:49:26 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
Apr 21 00:49:26 vm16 Tor[978]: Tor has successfully opened a circuit.
Looks like client functionality is working.
}}}
asn should we allow a time period for Tor to process new descriptors upon
refresh? In terms of seconds, so slower computers or computers on slower
connections don't get these warnings? During that time, previous
descriptors can be used. This also requires to ensure we try to refresh
descriptors for guards with N minutes before the ones we already have
expire, so we have this time period to handle it.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/21969#comment:1>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online