probe-lab / zikade

A Go implementation of the libp2p Kademlia DHT specification
Other
11 stars 3 forks source link

Flaking test: TestDHT_SearchValue_quorum_test_suite/TestQuorumReachedPrematurely #48

Closed iand closed 11 months ago

iand commented 1 year ago

This test run failed on a PR with no code changes

https://github.com/plprobelab/zikade/actions/runs/6408785740/job/17398549960?pr=47

--- FAIL: TestDHT_SearchValue_quorum_test_suite (10.79s)
    --- PASS: TestDHT_SearchValue_quorum_test_suite/TestQuorumReachedAfterDiscoveryOfBetter (0.27s)
    --- FAIL: TestDHT_SearchValue_quorum_test_suite/TestQuorumReachedPrematurely (10.15s)
    --- PASS: TestDHT_SearchValue_quorum_test_suite/TestQuorumUnspecified (0.18s)
    --- PASS: TestDHT_SearchValue_quorum_test_suite/TestQuorumZero (0.17s)
dennis-tra commented 1 year ago

Still relates to https://github.com/libp2p/go-libp2p/issues/2589 if we extended the test timeout this will work

iand commented 1 year ago

I will try that

dennis-tra commented 1 year ago

I'm just experimenting with https://github.com/plprobelab/zikade/pull/50

iand commented 1 year ago

I've been unable to reproduce this locally.

I ran go test -count 1000 -failfast -run TestDHT_SearchValue_quorum_test_suite without any failures.

My system:

14:48 $ uname -a
Linux astrolith 6.1.0-10-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.38-1 (2023-07-14) x86_64 GNU/Linux
dennis-tra commented 1 year ago

Yeah, it's also working fine on my machine... I tried to change the local dial timeout in #50. Tests are running currently. Let's see if that helped.

dennis-tra commented 1 year ago

was auto-closed because I mentioned it in the pull request. Can reopen if the problem persists

iand commented 1 year ago

Flaked again on this documentation-only PR: https://github.com/plprobelab/zikade/actions/runs/6423207647/job/17441233868?pr=52

dennis-tra commented 12 months ago

Happened again: https://github.com/plprobelab/zikade/actions/runs/6459054917/job/17534078617?pr=60

This time it's probably a different reason

iand commented 11 months ago

This hasn't reoccured since we fixed the deadlock in #63