Project

Profile

Help

Issue #8582

Updated by ttereshc over 3 years ago

In Pulp 2, it was is possible to add N different files with the same filename (relative path) to the same repo.   
 Only one is published but according to the database there are N files with the same name in one repo.   
 For that reason, migration plugin tries to create a repo version with all N files with the same name.  
 It can happen only on the system which was upgraded from pulp <= 2.13.3. 
 Here is the failure: 

 ~~~ 
  pulp: rq.worker:ERROR: Traceback (most recent call last): 
  File "/usr/lib/python3.6/site-packages/pulpcore/plugin/repo_version_utils.py", line 124, in validate_version_paths 
  validate_file_paths(paths) 
  File "/usr/lib/python3.6/site-packages/pulpcore/app/files.py", line 133, in validate_file_paths 
  raise ValueError(_("Path is duplicated: {path}").format(path=path)) 
  ValueError: Path is duplicated: katello-installer-3.0.0.88.tar.gz 
  During handling of the above exception, another exception occurred: 
  Traceback (most recent call last): 
  File "/usr/lib/python3.6/site-packages/rq/worker.py", line 936, in perform_job 
  rv = job.perform() 
  File "/usr/lib/python3.6/site-packages/rq/job.py", line 684, in perform 
  self._result = self._execute() 
  File "/usr/lib/python3.6/site-packages/rq/job.py", line 690, in _execute 
  return self.func(*self.args, **self.kwargs) 
  File "/usr/lib/python3.6/site-packages/pulp_2to3_migration/app/migration.py", line 194, in complex_repo_migration 
  create_repo_version(progress_rv, pulp2_repo, pulp3_remote) 
  File "/usr/lib/python3.6/site-packages/pulp_2to3_migration/app/migration.py", line 344, in create_repo_version 
  new_version.remove_content(Content.objects.filter(pk__in=to_delete)) 
  File "/usr/lib/python3.6/site-packages/pulpcore/app/models/repository.py", line 795, in __exit__ 
  repository.finalize_new_version(self) 
  File "/usr/lib/python3.6/site-packages/pulp_file/app/models.py", line 79, in finalize_new_version 
  validate_repo_version(new_version) 
  File "/usr/lib/python3.6/site-packages/pulpcore/plugin/repo_version_utils.py", line 139, in validate_repo_version 
  validate_version_paths(version) 
  File "/usr/lib/python3.6/site-packages/pulpcore/plugin/repo_version_utils.py", line 126, in validate_version_paths 
  raise ValueError(_("Cannot create repository version. {err}.").format(err=e)) 
  ValueError: Cannot create repository version. Path is duplicated: katello-installer-3.0.0.88.tar.gz. 
 ~~~ 

 It's a Pulp 2 bug #2773 which was fixed in 2.13.3 but we should handle this case to improve customer experience. There is no easy workaround, no    easy way for a user to figure out which file is published and which needs to be removed from a Pulp 2 repo to fix the problem. 

 ### Proposed solution 

 Check pulp 2 logic and make choice in the same way if it's deterministic. That will give a user exactly the same result as they have in Pulp 2.   
 If it's somewhat random in Pulp 2, choose the latest to add to a repo version. 


 ### Observation 

 It seems like Pulp 2 picks the first one(the oldest one) to publish, I haven't checked the code but the system I see the error on has the older file published. 

Back