fedimint / fedimint

Federated E-Cash Mint
https://fedimint.org/
MIT License
536 stars 209 forks source link

Backward compat againg `v0.3.0-rc.1` is failing due to unavailable dkg cli #5145

Closed dpc closed 2 weeks ago

dpc commented 2 weeks ago

Backward compat againg v0.3.0-rc.1 is failing, as it seems to try to use new dkg cli, which is not in this version. The condition in the code code is:

            if fedimint_cli_version >= *VERSION_0_3_0_ALPHA {
                run_cli_dkg(params, endpoints).await?;
            } else {
                // TODO(support:v0.2): old fedimint-cli can't do DKG commands. keep this old DKG
                // setup while fedimint-cli <= v0.2.x is supported
                run_client_dkg(admin_clients, params).await?;
            }

But that doesn't seem right.

00:00:02 2024-04-28T07:00:06.209208Z DEBUG spawn{task_name="JitTry alloc::sync::Arc<devimint::federation::Federation> value"}: fm::devimint: stderr output=error: unrecognized subcommand 'dkg'
00:00:02
00:00:02 Usage: fedimint-cli admin <COMMAND>
00:00:02
00:00:02 For more information, try '--help'.
00:17:24 dpc in 🌐 ren in …lab/fedimint/fedimint-wrk-1  24-04-27-0.2-was-supa-slow [$!?] is 📦 v0.4.0-alpha 🦀v1.77.2 ❄️
> /nix/store/6m5nkg73a1ziqhk8k3imixvz2wrbdh9y-fedimint-cli/bin/fedimint-cli --version
fedimint-cli 0.3.0-rc.1
00:17:28 dpc in 🌐 ren in …lab/fedimint/fedimint-wrk-1  24-04-27-0.2-was-supa-slow [$!?] is 📦 v0.4.0-alpha 🦀v1.77.2 ❄️
> /nix/store/6m5nkg73a1ziqhk8k3imixvz2wrbdh9y-fedimint-cli/bin/fedimint-cli admin
Usage: fedimint-cli admin <COMMAND>

Commands:
  status                  Show the status according to the `status` endpoint
  audit                   Show an audit across all modules
  guardian-config-backup  Download guardian config to back it up
  help                    Print this message or the help of the given subcommand(s)

Options:
  -h, --help  Print help
bradleystachurski commented 2 weeks ago

What is this failing against? It looks like the PR that bumped to v0.3.1-rc.1 passed back-compat.

https://github.com/fedimint/fedimint/pull/5030 https://github.com/fedimint/fedimint/actions/runs/8757485989/job/24036190539

Also, the output from above shows 0.3.0-rc.1 for the version

00:17:24 dpc in 🌐 ren in …lab/fedimint/fedimint-wrk-1  24-04-27-0.2-was-supa-slow [$!?] is 📦 v0.4.0-alpha 🦀v1.77.2 ❄️
> /nix/store/6m5nkg73a1ziqhk8k3imixvz2wrbdh9y-fedimint-cli/bin/fedimint-cli --version
fedimint-cli 0.3.0-rc.1
bradleystachurski commented 2 weeks ago

I verified locally just test-compatibility passes against v0.3.1-rc.1 (https://github.com/fedimint/fedimint/commit/a12af373eca15f62eff90cd17521aa827bf3bfa2)

dpc commented 2 weeks ago

@bradleystachurski In the output I did v0.3.0-rc.1, so I guess I misstyped it.

Can you try

./scripts/tests/test-ci-all-backcompat.sh v0.3.0-rc.1

?

Fail reliably on all

ETA: 243s Left: 30 AVG: 8.08s  local:0/36/100%/8.1s
Some tests failed:
Seq     Host    Starttime       JobRuntime      Send    Receive Exitval Signal  Command

29      :       1714355005.102      70.368      0       395111  1       0       run_test_for_versions latency_test_reissue FM: current CLI: v0.3.0-rc.1 GW: current

30      :       1714355008.313      70.085      0       395710  1       0       run_test_for_versions latency_test_ln_send FM: current CLI: v0.3.0-rc.1 GW: current

31      :       1714355048.447      67.957      0       327969  1       0       run_test_for_versions latency_test_ln_receive FM: current CLI: v0.3.0-rc.1 GW: current

32      :       1714355051.679      68.127      0       327732  1       0       run_test_for_versions latency_test_fm_pay FM: current CLI: v0.3.0-rc.1 GW: current

33      :       1714355055.722      69.944      0       395111  1       0       run_test_for_versions latency_test_restore FM: current CLI: v0.3.0-rc.1 GW: current

34      :       1714355059.957      68.226      0       423237  1       0       run_test_for_versions reconnect_test FM: current CLI: v0.3.0-rc.1 GW: current

35      :       1714355065.240      70.125      0       393070  1       0       run_test_for_versions lightning_reconnect_test FM: current CLI: v0.3.0-rc.1 GW: current

36      :       1714355068.421      68.020      0       328059  1       0       run_test_for_versions gateway_reboot_test FM: current CLI: v0.3.0-rc.1 GW: current

Search for '## FAIL' to find the end of the failing test
dpc commented 2 weeks ago

v0.3.0 passes, so I guess it might not be worth to fix.

bradleystachurski commented 2 weeks ago

Ahh yep, I'm also getting failures vs v0.3.0-rc.1

bradleystachurski commented 2 weeks ago

Shoot, I'm getting a failure running just test-compatibility v0.3.0 against master (https://github.com/fedimint/fedimint/commit/0d95d32bec03bcc578debcd34e63422809ca1f99). The failure is the same as https://github.com/fedimint/fedimint/issues/4999, so perhaps the root issue is the same.

bradleystachurski commented 2 weeks ago

Tracking in https://github.com/fedimint/fedimint/issues/5156

dpc commented 2 weeks ago

Shoot, I'm getting a failure running just test-compatibility v0.3.0 against master (https://github.com/fedimint/fedimint/commit/0d95d32bec03bcc578debcd34e63422809ca1f99). The failure is the same as https://github.com/fedimint/fedimint/issues/4999, so perhaps the root issue is the same.

I thought I tried it and it passed but I was distracted.

bradleystachurski commented 2 weeks ago

I thought I tried it and it passed but I was distracted.

No worries! I observed flakiness on the failures with just test-upgrades "v0.3.0 current", so it's possible it was successful on your end but flaked on my end.