crazy-max / artifactory-cleanup Goto Github PK

Behaviour

Running on version 1.7.0 on Windows 10
Command ".\artifactory-cleanup.exe --dry-run" produces

Steps to reproduce this issue

1. Version 1.7.0 artifactory-cleanup.exe on Windows 10 using Windows Powershell
2. Command ".\artifactory-cleanup.exe --dry-run"
3. Using example artifactory-cleanup.yml from download.
4. Get error panic: runtime error: invalid memory address or nil pointer dereference
   [signal 0xc0000005 code=0x0 addr=0x0 pc=0xd6bed5]

Expected behaviour

Should return results from Dryrun

Actual behaviour

Get no results and an error (attached)

Configuration

• Artifactory Cleanup version (1.7.0):

• Platform (windows 10):

• System info (type uname -a): Device name USBLPF36S803
  Full device name USBLPF36S803.bct.gambro.net
  Processor 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz 2.80 GHz
  Installed RAM 32.0 GB (31.7 GB usable)
  Device ID 8AF03655-E747-40E2-8FDF-F82ED5B92155
  Product ID 00330-80000-00000-AA324
  System type 64-bit operating system, x64-based processor
  Pen and touch No pen or touch input is available for this display

• JFrog Artifactory version: Jfrog Cloud (v2023.2.414)

Logs

Artifactory Cleanup logs (set --log-level flag to debug) No logs as it never gets that far.

Is possible to use a subfolder?. I'm using artifactory-cleanup without any issue to clean a repository called x which is Package Type: | RPM

If I try to use cleanup_repository= x/y or y I'm getting: ERR Cannot retrieve repo configuration for sgd error="400 Bad Request" policy=cleanup-policy

I duplicated the project which is working well and for the new one I only changed the parameters cleanup_repository as I described before, plus I set cleanup_include_names=* and cleanup_dryrun=true for my testing.

Hi Max,

Loving the tool! Is there a plan for the tool to not require a CRON expression to run?
I'm using it in my CI setup, so I just put * * * * * there and it complains every minute, but it runs.
However, after it's done, the CI pipeline keeps running as the next run is automatically scheduled in another minute.

Would be great to have a option to simply run it immediately.
If it's already available, I must've missed it in the docs, so please point me to a correct section.

Thank you!

Behaviour

The setting lastDownloaded: true is not taken into account because the last_downloaded for docker images is always 1970-01-01T01:00:00+01:00.

Configuration

• Artifactory Cleanup version (0.1.0): 1.5.0
• Platform (windows/linux/freebsd/macos): Tested on MacOS and Linux
• JFrog Artifactory version: 7.35.2

Logs

Wed, 27 Apr 2022 11:30:20 CEST TRC Artifactory docker tag last modified response last_modified={"lastModified":"2022-04-23T00:22:07.230+0000","uri":"https://example/artifactory/api/storage/local-docker/foo/bar/manifest.json"}
Wed, 27 Apr 2022 11:30:20 CEST TRC Artifactory artifact file stats response file_stats={"downloadCount":0,"lastDownloaded":0,"remoteDownloadCount":0,"remoteLastDownloaded":0,"uri":"https://example/artifactory/local-docker/foo/bar"}
Wed, 27 Apr 2022 11:30:20 CEST INF Image will be removed docker_repo=foo/bar dry_run=true last_downloaded=1970-01-01T01:00:00+01:00 last_modified=2022-04-23T00:22:07Z policy=policy_docker registry=local-docker size=892MB tag=bar

Solution

My suggested solution is to add /manifest.json also to the last downloaded call in

 return c.LastDownloaded(registry, path.Join(repo, tag, "manifest.json")) 

Or is there any reason you added not the manifest.json? I can reproduce it via the Artifactory API. When calling the file stats call to the docker folder it's returning "lastDownloaded: 0", when querying the manifest.json it has a proper value set.

As JFrog is going to deprecate APIKey, what will be valuable option to keep using your awesome tool?
https://www.jfrog.com/confluence/display/JFROG/User+Profile#UserProfile-JFrogAPIKeyDeprecationProcess