Bug #11560

Investigate if Chutney's network defaults fits the test suite needs

Added by bertagaz 2016-07-08 05:06:30 . Updated 2019-08-10 17:27:51 .

Status:
Resolved
Priority:
Normal
Assignee:
bertagaz
Category:
Test suite
Target version:
Start date:
2016-07-08
Due date:
% Done:

0%

Feature Branch:
Type of work:
Research
Blueprint:

Starter:
Affected tool:
Deliverable for:

Description

While testing online scenarios and watching Tails’ Tor client logs, I often see that it has troubles to have enough directory informations to build a circuit. Its logs are filled with lines like:

[notice] I learned some more directory information, but not enough to build a circuit: We have no recent usable consensus.

Even after retry_tor(), this problem is not solved, and the scenario fails.

If one watch at the chutney Tor nodes logs, they are filled with auth directory votes and stuffs like that. At the moment, the consensus is set to last 20 seconds only (if I get it right), so the network load with this is pretty high, and it may be difficult for Tails’ Tor client to get a valid consensus.

The chutney settings are quite tight, and done so to stress test the fake Tor network and accelerate it. But for Tails tests, we want a stable Tor network, so we might want to have different settings.

I’ve had good results by setting V3AuthVotingInterval to 600 (in place of 10), but there may be other changes we may want to consider.

Subtasks

History

#1 Updated by intrigeri 2016-07-08 07:42:25

  • Assignee set to bertagaz

(You’ve set a “Target version” in a few weeks, so I assume it means you’re going to take care of it.)

#2 Updated by anonym 2016-07-08 09:32:46

Note that this might be Bug #10238, which has plagued us for years.

#3 Updated by bertagaz 2016-07-09 07:48:46

  • Assignee deleted (bertagaz)

anonym wrote:
> Note that this might be Bug #10238, which has plagued us for years.

I’m not sure, given what I describe happens after Tor has bootstrapped (or said so).

I’ll run the same kind of tests you did on Bug #10238, and gather stats to see if changing Chutney’s config has any impact. Btw I’ve added a patch on the Bug #10238 branch that also save the chutney data directory when wait_until_tor_is_working fails, do you think it’s worth pushing?

#4 Updated by intrigeri 2016-07-13 12:22:15

  • Assignee set to anonym
  • Target version deleted (Tails_2.5)
  • QA Check set to Info Needed

(Dropping the target version since I see nobody volunteering to do it this month.)

#5 Updated by anonym 2016-09-28 08:09:44

bertagaz wrote:
> anonym wrote:
> > Note that this might be Bug #10238, which has plagued us for years.
>
> I’m not sure, given what I describe happens after Tor has bootstrapped (or said so).
>
> I’ll run the same kind of tests you did on Bug #10238, and gather stats to see if changing Chutney’s config has any impact. Btw I’ve added a patch on the Bug #10238 branch that also save the chutney data directory when wait_until_tor_is_working fails, do you think it’s worth pushing?

Yes, please! And sorry for the delay! :S

#6 Updated by anonym 2016-09-28 08:10:05

  • Assignee changed from anonym to bertagaz
  • QA Check changed from Info Needed to Dev Needed

#7 Updated by intrigeri 2019-08-10 17:27:51

  • Status changed from Confirmed to Resolved

Given:

  • We’ve bumped V3AuthVotingInterval to 180 a year ago.
  • On Bug #16792, anonym is reviewing and improving our Chutney config.

I think we can close this ticket. If there are still problems or other specific suggestions, let’s file new tickets about them :)