AWS S3 deploy failing due to log length

Deployment
1

Hi all,

I have a build (https://travis-ci.org/github/supercollider/supercollider/jobs/720071451) which is currently failing during deployment (it seems) because the AWS S3 step prints the following log line many, many times:

/Users/travis/.rvm/gems/ruby-2.6.3/gems/aws-sdk-core-2.11.571/lib/aws-sdk-core/eager_loader.rb:22: warning: constant Struct::Tms is deprecated

until eventually The job exceeded the maximum log length, and has been terminated.

is there anything i can do configuration-wise to get this build to pass, or anything travis folks can do to solve it? this is for a release, so i’d really like to avoid having to modify my .travis.yml or other committed scripts. a similar build under the same configuration deployed fine about 2 weeks ago (https://travis-ci.org/github/supercollider/supercollider/jobs/715363990).

thanks in advance!

2

I’ve composed a fix. Please test it by specifying the following under the deploy entry:

edge:
  source: 'native-api/dpl'
  branch: 'silence_s3_warnings'

This also seems to be fixed in deployment v2, so a workaround is:

edge: true
3

that seems to have fixed it in a test build: https://travis-ci.org/github/supercollider/supercollider/jobs/720175189, thanks!

1 Like
4

what’s the next step? will that be merged and deployed on existing machines or will i have to add something extra to my configs, or both? this is my first time reporting something like this here.

5

Once this PR is merged and gets into a tag, you can remove the above clause.

6

I’ve pointed out @native-api’s patch in the upstream issue:


Hopefully this patch will be incorporated and, there will be a v2.x release soon. This way, everyone can benefit from it.
1 Like
7

AWS have made a fix upstream (https://github.com/aws/aws-sdk-ruby/pull/2413) and released it in https://rubygems.org/gems/aws-sdk/versions/2.11.596 .

@VIRTUALDOG as such, the workaround should no longer be needed.