Open oivindoh opened 6 years ago
Sorry about this issue. I'd like to reproduce. Could you share your configuration, please? Yesterday, I update to the new version.
Ok. I'll have a look
I run my pipeline to set 2.6.0, then it works. I'd like to reproduce. Could you share your pipeline, please? What I did on the comment is like this.
curl -L https://storage.googleapis.com/kubernetes-helm/helm-v2.6.0-linux-amd64.tar.gz | tar xz
cp **/helm /opt/vsts/work/r1/a/.vstsbin
Do you have any **/helm directory?
The change I released yesterday is, it fails when I've got error when we execute a command. Before that I had a issue it doesn't error, even if it fails. In your environment, the script cause an error however, it was ignored. That is why, it passed. You might be able to use Continue on error on Control Options for currently, however, essentially, avoid **/helm directory might be the essential solution. I also, come up with the idea how to solve this issue.
Also having this issue. My setup is brand new, so I haven't yet seen it working, the error is the same:
UnhandledPromiseRejectionWarning: Unhandled promise rejection (rejection id: 1): Failed to exec the .helmdownloader.sh which is helm downloader.
I tried both the default helm version 2.5.0 and the one I have locally, 2.7.2 - no difference
@TsuyoshiUshio Any update on this? I'm also getting this on releases at v2.8.2.
Log:
2018-04-06T04:23:27.9849679Z ##[section]Starting: helm upgrade
2018-04-06T04:23:28.0675689Z ==============================================================================
2018-04-06T04:23:28.0846956Z Task : helm
2018-04-06T04:23:28.1035285Z Description : execute helm command.
2018-04-06T04:23:28.1219035Z Version : 1.0.3
2018-04-06T04:23:28.1387539Z Author : Tsuyoshi Ushio
2018-04-06T04:23:28.1558553Z Help : See [Kubernetes task](https://github.com/TsuyoshiUshio/KubernetesTask)
2018-04-06T04:23:28.1712435Z ==============================================================================
2018-04-06T04:23:30.2131897Z [command]/opt/vsts/work/r1/a/.vstsbin/helm upgrade canary . -i --dry-run --debug
2018-04-06T04:23:30.9326555Z Error: no Chart.yaml exists in directory "/opt/vsts/work/r1/a"
2018-04-06T04:23:30.9524394Z [debug] Created tunnel using local port: '37077'
2018-04-06T04:23:30.9721742Z (node:140) UnhandledPromiseRejectionWarning: Unhandled promise rejection (rejection id: 1): Failed to execute helm command.
2018-04-06T04:23:30.9798729Z
2018-04-06T04:23:30.9952231Z [debug] SERVER: "127.0.0.1:37077"
2018-04-06T04:23:31.0030267Z
2018-04-06T04:23:31.0200885Z Release "canary" does not exist. Installing it now.
2018-04-06T04:23:31.0367532Z [debug] CHART PATH: /opt/vsts/work/r1/a
2018-04-06T04:23:31.0450727Z
2018-04-06T04:23:31.1130523Z ##[section]Finishing: helm upgrade
Since yesterday we've had a bunch of failed builds, apparently because the helm downloader stage is failing on an unhandled promise rejection.
The helm version we've been targeting is 2.6.0, which is the one we started seeing the failure on yesterday. Attempting to specify 2.7.0 leads to the same - these versions appear to be available on the URI found in this project.
Here's a VSTS build log