RFR: 8343190: GHA: Try building JTReg several times

Aleksey Shipilev shade at openjdk.org
Tue Oct 29 14:19:19 UTC 2024


On Tue, 29 Oct 2024 13:39:08 GMT, Magnus Ihse Bursie <ihse at openjdk.org> wrote:

>> We still sometimes have problems checking out JTReg build dependencies. [JDK-8342988](https://bugs.openjdk.org/browse/JDK-8342988) makes it less pronounced, but it still happens. We might need to do a retries for JTReg builds to make them more reliable.
>> 
>> Additional testing:
>>  - [x] GHA (seeing retries with deliberately broken JTReg build)
>>  - [x] GHA (passing JTReg build)
>
> .github/actions/build-jtreg/action.yml line 56:
> 
>> 54:       run: |
>> 55:         # Try building JTReg several times, backing off exponentially on failure.
>> 56:         # ~500 seconds in total should be enough to capture most of the transient
> 
> Since building jtreg will block building the code (not just running tests), this feels like a potentially quite long waiting time. 
> 
> Are we sure waiting even helps? What kind of issues is it we are encoutering? Downloading resources from the net?

Well, we are seeing transient network errors like:


[build.sh][INFO] Downloading https://archive.apache.org/dist/ant/binaries/apache-ant-1.10.8-bin.zip to /home/runner/work/jdk/jdk/jtreg/src/make/../build/deps/apache-ant-1.10.8-bin.zip
Error: sh][ERROR] wget exited with exit code 4 


wget exit code 4 means "Network failure". We "know" waiting helps in the sense that job restarts help.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/21758#discussion_r1820895426


More information about the build-dev mailing list