#1034352 golang-github-azure-go-autorest: autopkgtest regression on arm64: request header doesn't match #1034352
- Package:
- src:golang-github-azure-go-autorest
- Submitter:
- Paul Gevers
- Date:
- 2023-04-22 11:45:02 UTC
- Severity:
- normal
- Tags:
Dear maintainer(s), Your package has an autopkgtest, great. However, it fails on arm64 only since July 2022. Can you please investigate the situation and fix it? I copied some of the output at the bottom of this report. I'd like to note that our arm64 workers are all hosted in China, which *might* be of relevance as this test seems to be testing against the internet (I could be wrong). If it is testing against the internet, it requires the needs-internet restriction (although that currently doesn't change anything). The release team has announced [1] that failing autopkgtest on amd64 and arm64 are considered RC in testing. [Release Team member hat on] Because we're currently in the hard freeze for bookworm, I have marked this bug as bookworm-ignore. Targeted fixes are still welcome. More information about this bug and the reason for filing it can be found on https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation Paul [1] https://lists.debian.org/debian-devel-announce/2019/07/msg00002.html https://ci.debian.net/data/autopkgtest/testing/arm64/g/golang-github-azure-go-autorest/32818576/log.gz === RUN TestLogReqRespNoBody logger_test.go:90: request header doesn't match: (2023-04-12T19:18:59.1355402+08:00) INFO: REQUEST: GET https://fakething/dot/com --- FAIL: TestLogReqRespNoBody (0.01s) === RUN TestLogReqRespWithBody logger_test.go:155: request header doesn't match: (2023-04-12T19:18:59.1420497+08:00) INFO: REQUEST: GET https://fakething/dot/com --- FAIL: TestLogReqRespWithBody (0.01s) FAIL FAIL github.com/Azure/go-autorest/logger 0.017s
Looks like it was caused by the UTC+8 timezone in the testbed. I agree it's nice to have the tests not depending on the system timezone. But do we really want to bother with that? Could you just ensure all testbeds have the same timezone?
Hi Shengjing, In my opinion we should do both. You don't want people to go figure this out if they try to run the tests themselves in their own timezone. Also, we are only in control of the hosts that Debian uses. Any downstream may run into the same problems. Paul
Hi, I have changed the timezone on all our hosts, but your test still fails. https://ci.debian.net/data/autopkgtest/testing/arm64/g/golang-github-azure-go-autorest/32859920/log.gz is the test I just scheduled. It failed in the same way. Paul
Maybe you haven't rebuilt the lxc image with the new timezone, so it still uses UTC+8. Anyway, I've staged the change in git https://salsa.debian.org/go-team/packages/golang-github-azure-go-autorest/-/commit/73b09d0
Hi, That's what I thought so later too. Let's see what tomorrow brings. Great. Paul
Hi, So, the testsing container was rebuild on Apr 14 06:29 (UTC) and I just ran the golang-github-azure-go-autorest test: https://ci.debian.net/data/autopkgtest/testing/arm64/g/golang-github-azure-go-autorest/32898818/log.gz It still fails. Paul
Hi, Turns out that /etc/timezone is also used in some places and that although the host was running UTC, the containers were *still* being created in the timezone mentioned in /etc/timezone. I have now made all our hosts run on UTC with *both* /etc/timezone and /etc/localtime fixed [1]. I hope that tomorrow will see this resolved from the infrastructure side of things too. Paul [1] https://salsa.debian.org/ci-team/debian-ci-config/-/commit/42162fb473f9ee3ca0af2be882fb0f1e2bd8917b
Hi, The tests now pass. I still appreciate it if you future upload also fixes this on the package side as discussed earlier, as it's not only Debian's infrastructure that's affected. Paul