If StrictExitNode is set to 1 tor start using 100% cpu after a few minutes

As the title says, I have Tor 0.2.1.30 installed on my system along with a correctly configured polipo. I've noticed that when I manually restrict tor to use a certain set of exit nodes (using StrictExitNodes 1 and endnodes <...>) and restart it, after a few minute it reaches 100% cpu usage. Although tor seems to be still running fine (i can still open websites etc...), I prefer to avoid such a massive load. I then comment those two line out, restart tor, and everything is back to normal.

Is this a bug or just some predictable behaviour?

Thank you.

Trac:
Username: Phoenix87

changed milestone to %Tor: unspecified

added component::core tor/tor milestone::Tor: unspecified priority::medium reporter::Phoenix87 resolution::user disappeared severity::normal status::closed tor-client type::defect version::tor 0.2.2.35 labels

That indicates a bug, but Tor 0.2.1.30 is obsolete. Please upgrade to an actually supported version (atm 0.2.2.35 or 0.2.3.12-alpha) and reopen this bug if it still occurs.

Trac:
Resolution: N/A to invalid
Status: new to closed

Trac:
Username: Phoenix87
Status: closed to reopened
Resolution: invalid to N/A

Trac:
Username: Phoenix87
Milestone: Tor: 0.2.1.x-final to Tor: 0.2.2.x-final

Please provide more information about your configuration, OS, etc.

Updated to Version: 0.2.2.35-1~oneiric+1 but problem is still there. I'm using StrictNodes instead of StrictExitNodes, for it seems that the latter has become deprecated with the new release.

Trac:
Username: Phoenix87

Trac:
Username: Phoenix87

torrc

my tor configuration. replace <country_code> with a valid country code.

Trac:
Username: Phoenix87

Trac:
Username: Phoenix87

config

the standard polipo configuration file

Trac:
Username: Phoenix87

Replying to Sebastian:

Please provide more information about your configuration, OS, etc.

Please find the configuration files attached to this ticket. I'm running ubuntu oneiric and used the official tor repositories for oneiric.

Trac:
Username: Phoenix87
Version: N/A to Tor: 0.2.2.35

hrm ok, if it really is UK then there should be a few nodes available. I'll see if I can reproduce it, otherwise you should get ready to profile your Tor to see where it spends its time

Replying to Sebastian:

hrm ok, if it really is UK then there should be a few nodes available. I'll see if I can reproduce it, otherwise you should get ready to profile your Tor to see where it spends its time

Well actually I've tried a few other country codes other than UK, and all seem to lead to the very same issue. I've also tried to manually specify some node nicknames with again the same result.

Trac:
Username: Phoenix87

Oh hey, if you're on Linux, this should be pretty easy to track down with profiling: just run Tor under 'perf', and see where the cycles are spent. To do this, just run "perf record tor (options)" instead of "tor (options)". Once the CPU has spiked for a while, exit Tor, then use "perf report > output" to get a breakdown of where Tor is spending its time. (I'd do this myself, but I can't make the bug happen: I just tried for about 20 minutes to reproduce with the current maint-0.2.2 branch on my computer, and didn't get any weird CPU issues.)

(BTW, as far as I can tell, "UK" isn't likely to work... the maxmind geoip database thinks that the country code in question is "GB".)

Also, is this

Trac:
Status: reopened to needs_information

Correct, {UK} is not going to get you what you want.

See also #3296 (moved) -- I wonder if some info-level logs would help here (in addition to, or instead of, the perf results above).

Trac:
Milestone: Tor: 0.2.2.x-final to Tor: unspecified

Trac:
Keywords: N/A deleted, tor-client added

Trac:
Component: Tor Client to Tor

Trac:
Sponsor: N/A to N/A
Status: needs_information to closed
Reviewer: N/A to N/A
Severity: N/A to Normal
Resolution: N/A to user disappeared

closed

moved to tpo/core/tor#5362 (closed)