Tests cleanup #53

prathamesh · 2024-01-15T08:41:35Z

prathamesh commented

2024-01-15 08:41:35 +00:00

General test improvements.

Simplifies Dockerfile
Refactors tests so they can be run independently

General test improvements. - Simplifies Dockerfile - Refactors tests so they can be run independently

prathamesh added 3 commits 2024-01-15 08:41:36 +00:00

formatting 7eb37f894d

refactor tests to run independently 7e18b5a8cf

dockerfile cleanup

Tests / sdk_tests (pull_request) Successful in 19m25s

Details

913389ab9f

prathamesh changed title from ~~Tests cleanup~~ to [WIP] Tests cleanup

2024-01-15 08:51:07 +00:00

prathamesh added 1 commit 2024-01-16 09:04:55 +00:00

Add container build script and update workflow

Tests / sdk_tests (pull_request) Failing after 5m32s

Details

e5900b2e4a

prathamesh added 1 commit 2024-01-16 09:57:57 +00:00

Update container build script

Tests / sdk_tests (pull_request) Failing after 4m55s

Details

b7223c1c43

prathamesh added 1 commit 2024-01-16 10:49:59 +00:00

Update naming test timeout

Tests / sdk_tests (pull_request) Failing after 4m54s

Details

603fe5916a

prathamesh added 1 commit 2024-01-16 12:05:30 +00:00

Update naming test timeout

Tests / sdk_tests (pull_request) Failing after 5m10s

Details

972b878674

prathamesh changed title from ~~[WIP] Tests cleanup~~ to Tests cleanup

2024-01-16 12:17:22 +00:00

ashwin requested review from telackey 2024-01-16 12:22:38 +00:00

telackey requested changes 2024-01-16 18:22:12 +00:00

telackey left a comment

I like the look of the changes overall, but the CI job is failing in Gitea. I think we need to get that sorted out before merging.

telackey commented

2024-01-16 18:33:34 +00:00

I suspect the failure is related to the new script not using the correct DOCKER_HOST.

This may be corrected with our new task executor containers (I'll try relaunching once deployed), otherwise you'll need to do something like:

run: DOCKER_HOST=unix:///var/run/dind.sock  ./scripts/build-sdk-test-container.sh

I suspect the failure is related to the new script not using the correct DOCKER_HOST. This may be corrected with our new task executor containers (I'll try relaunching once deployed), otherwise you'll need to do something like: ``` run: DOCKER_HOST=unix:///var/run/dind.sock ./scripts/build-sdk-test-container.sh ```

prathamesh added 1 commit 2024-01-17 13:05:17 +00:00

Add debug logs

Tests / sdk_tests (pull_request) Failing after 11s

Details

0a99a1aea1

prathamesh added 1 commit 2024-01-17 13:10:10 +00:00

Update repo URL

Tests / sdk_tests (pull_request) Failing after 10m15s

Details

9535362887

prathamesh added 1 commit 2024-01-17 13:26:02 +00:00

Log docker versions in CI

Tests / sdk_tests (pull_request) Failing after 12s

Details

7db69c314b

prathamesh force-pushed pm-tests-cleanup from 7db69c314b to d7ca4730a2

2024-01-17 13:26:49 +00:00

Compare

prathamesh commented

2024-01-17 13:45:04 +00:00

but the CI job is failing in Gitea

The reason for failure earlier was that laconicd PR (cerc-io/laconicd#134) is yet to be merged and the CI in this PR has been updated to use changes from there.

However, even if I temporarily point it to use the PR branch, the sdk tests are still failing (naming tests). Have added some logs to investigate this, but needs more debugging.

> but the CI job is failing in Gitea The reason for failure earlier was that laconicd PR (https://git.vdb.to/cerc-io/laconicd/pulls/134) is yet to be merged and the CI in this PR has been updated to use changes from there. However, even if I temporarily point it to use the PR branch, the sdk tests are still failing (naming tests). Have added some logs to investigate this, but needs more debugging.

prathamesh commented

2024-01-18 03:42:28 +00:00

However, even if I temporarily point it to use the PR branch, the sdk tests are failing (naming tests). Have added some logs to investigate this, but needs more debugging.

The tests do pass locally though; both when laconicd is run on the host machine as well as in docker (as done in the workflow).

> However, even if I temporarily point it to use the PR branch, the sdk tests are failing (naming tests). Have added some logs to investigate this, but needs more debugging. The tests do pass locally though; both when laconicd is run on the host machine as well as in docker (as done in the workflow).

telackey commented

2024-01-18 18:35:24 +00:00

We actually redeployed the task executors with ones that automatically set DOCKER_HOST, so that probably corrected part of it.

As to the naming failures, if it works locally, is the idea that it is some sort of timing/performance issue in CI?

I am just a bit concerned in that, that unlike the laconid tests, which we need to get passing in Gitea, these tests actually were passing, although looking at the history they seem to have been pretty flakey.

I didn't worry about them failing in CI on the last PR, because it seemed to be that the failures were just from the laconicd/protobuf version mismatch and would sort themselves out, but since the PR is about cleaning up the tests, I think it is worth some investigation as to why they are failing.

We actually redeployed the task executors with ones that automatically set DOCKER_HOST, so that probably corrected part of it. As to the naming failures, if it works locally, is the idea that it is some sort of timing/performance issue in CI? I am just a bit concerned in that, that unlike the laconid tests, which we need to get passing in Gitea, these tests actually were passing, although looking at the history they seem to have been pretty flakey. I didn't worry about them failing in CI on the last PR, because it seemed to be that the failures were just from the laconicd/protobuf version mismatch and would sort themselves out, but since the PR is about cleaning up the tests, I think it is worth some investigation as to why they are failing.

telackey commented

2024-01-18 18:57:11 +00:00

One option (and the same goes for cerc-io/laconicd#134) is that if everything is passing for you locally, to go ahead and merge, but open an issue with all the details of the failures, and we will fix the CI as a separate task/PR.

Sound reasonable?

One option (and the same goes for https://git.vdb.to/cerc-io/laconicd/pulls/134) is that if everything is passing for you locally, to go ahead and merge, but open an issue with all the details of the failures, and we will fix the CI as a separate task/PR. Sound reasonable?

telackey referenced this pull request from cerc-io/laconicd

2024-01-18 18:57:28 +00:00

Clean up test suite #134

telackey requested review from telackey 2024-01-18 18:58:02 +00:00

telackey approved these changes 2024-01-18 18:58:23 +00:00

prathamesh referenced this pull request

2024-01-22 06:43:14 +00:00

SDK tests failing in CI #54

prathamesh added 1 commit 2024-01-22 06:56:08 +00:00

Remove debug logs

Tests / sdk_tests (pull_request) Failing after 3m44s

Details

f525e18ca4

prathamesh commented

2024-01-22 08:30:03 +00:00

One option (and the same goes for cerc-io/laconicd#134) is that if everything is passing for you locally, to go ahead and merge, but open an issue with all the details of the failures, and we will fix the CI as a separate task/PR.

Created an issue to track this: #54

> One option (and the same goes for cerc-io/laconicd#134) is that if everything is passing for you locally, to go ahead and merge, but open an issue with all the details of the failures, and we will fix the CI as a separate task/PR. Created an issue to track this: https://git.vdb.to/cerc-io/laconic-sdk/issues/54

ashwin merged commit fedf35d702 into main

2024-01-22 08:30:36 +00:00

ashwin deleted branch pm-tests-cleanup

2024-01-22 08:30:36 +00:00

ashwin referenced this issue from a commit

2024-01-22 08:30:38 +00:00

Tests cleanup (#53)

ashwin referenced this pull request

2024-01-22 08:31:07 +00:00

Tests cleanup #43

prathamesh referenced this pull request from cerc-io/laconicd

2024-01-22 09:08:15 +00:00