Can't upload single file using wildcard

[Orrison]

Unable to upload a single file generated in the job using a wildcard. I have input it into SOURCE_DIR and though it would work as the documentation for SOURCE_DIR states "The local directory (or file) you wish to sync/upload to S3."

I have to use a wildcard for a part of the file name as I can't always be sure what the version name is. E.X.:

SOURCE_DIR: 'package-name-*.zip'

But it always fails with the error:

warning: Skipping file /github/workspace/package-name-1.0.0.zip/. File does not exist.

Obviously it sees the file or else it wouldn't have been able to grab that 1.0.0 version number so it is strange that can't find it. I do see that it has /. at the end so I wonder if it is thinking it is a folder? Is there some way I can make it understand it is a single file and not a folder?

[Orrison]

I was able to get this to work by not included SOURCE_DIR and instead running it with these arguments:

with:
        args: --exclude '*' --include '/package-name-*.zip'

Leaving this open to get confirmation this is the best way to do it.

[karolszafranski]

I run into the same issue even when not using a wildcard. I just want to upload a single file, not a directory. I expect the following workflow to be working:

on:
  push:
    branches:
      - s3upload

jobs:
  test-s3-upload:
    runs-on: ubuntu-latest
    name: Upload artifact to S3
    steps:
      - name: Prepare artifact
        run: |
          echo "Hello World" > README.txt
          touch "$(date).txt"
          printenv | grep GITHUB_ > env.txt
          tar cf archive.tar *.txt
          ls -l
          pwd
      
      - name: Upload to S3
        uses: jakejarvis/s3-sync-action@7ed8b11
        with:
          args: --acl public-read
        env:
          AWS_S3_BUCKET: ${{ secrets.TEST_AWS_S3_BUCKET }}
          AWS_ACCESS_KEY_ID: ${{ secrets.TEST_AWS_ACCESS_KEY_ID }}
          AWS_SECRET_ACCESS_KEY: ${{ secrets.TEST_AWS_SECRET_ACCESS_KEY }}
          AWS_REGION: 'eu-central-1'
          SOURCE_DIR: 'archive.tar'
          DEST_DIR: 'test'

It fails with the following error:

warning: Skipping file /github/workspace/archive.tar/. File does not exist.

It says "warning" but the step is marked with ❌ and the job fails.

I get an expected result with workaround proposed by @Orrison:

      - name: Upload to S3
        uses: jakejarvis/s3-sync-action@7ed8b11
        with:
          args: --acl public-read --exclude '*' --include 'archive.tar'
        env:
          AWS_S3_BUCKET: ${{ secrets.TEST_AWS_S3_BUCKET }}
          AWS_ACCESS_KEY_ID: ${{ secrets.TEST_AWS_ACCESS_KEY_ID }}
          AWS_SECRET_ACCESS_KEY: ${{ secrets.TEST_AWS_SECRET_ACCESS_KEY }}
          AWS_REGION: 'eu-central-1'
          DEST_DIR: 'test'

[yelizariev]

I've faced the same issue.

@jakejarvis maybe add new environment value SOURCE_FILE and when it's set, use the suggested command #26 (comment)
?

[kaiomagalhaes]

I faced the same issue. Right now I'm solving it by using this solution: #26 (comment)

[sedkis]

Ditto.. doesn't work for me either

[andylehti]

Here is the console message, bucket name changed to generic name.

root@processor:~# aws s3 sync /root/data-processed/*.csv s3://bucket/sub
warning: Skipping file /root/data-processed/2023-04-22T07:51:48.595750.out.csv/. File does not exist.

So, it clearly knows the file exists but doesn't do anything about it.