Problem: Full reingest fails when Storage Service and Pipeline are on separate servers (under some system configuration conditions)

Expected behaviour

Full reingest should work irrespective of the system configuration (provided the configuration is correct)

Current behaviour

Attempting a full reingest fails under the following configuration:

Storage Service and the pipeline (Dashboard) are on separate servers (i.e., connected together via a pipeline FS space)
The pipeline FS staging path value is different from the Storage Service internal processing location

Steps to reproduce

On a system where the Storage Service and the pipeline are on separate servers (i.e., connected together via a pipeline FS space), configure the pipeline local filesystem staging path to be different from the Storage Service internal processing location (e.g., if the Storage Service Internal Processing location path is "/var/archivematica/storage_service" , set the pipeline local filesystem staging path to "/mnt/pfs_staging" )
Run an ingest and store an AIP (the ingest should complete without failures)
Trigger a full reingest of the ingested AIP (e.g., from the Archival Storage tab), and approve the transfer from the dashboard
During the transfer processing the job "Verify transfer compliance" (Microservice: Verify transfer compliance) fails

Your environment (version of Archivematica, operating system, other relevant details)

AM version: stable/1.15.x , SS version: stable/0.21.x (although the same problem may have also occurred in previous versions, ref issue https://github.com/archivematica/Issues/issues/1456 ) OS: Rocky 9/RHEL 9 (the issue seems to be unrelated to the OS)

Additional Information/Analysis

After triggering a reingest, and before approving the transfer, it can be noted that the transfer files are "enclosed" under a duplicated directory structure in the watched directory, for example:

watchedDirectories/activeTransfers/standardTransfer/
└── [   60]  oneimage2-8e8ff0f8-eef5-4401-bc2a-352f45b80b45
    └── [  112]  oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387
        ├── [  195]  bag-info.txt
        ├── [   55]  bagit.txt
        ├── [   86]  data
        │   ├── [  107]  logs
        │   │   ├── [   36]  fileFormatIdentification.log
        │   │   ├── [ 1.2K]  filenameChanges.log
        │   │   ├── [  124]  FileUUIDs.log
        │   │   └── [   60]  transfers
        │   │       └── [   18]  oneimage2-d4a5c1ca-3393-4da8-bb7d-e2cad03b8175
        │   │           └── [   90]  logs
        │   │               ├── [  599]  fileFormatIdentification.log
        │   │               ├── [  143]  filenameChanges.log
        │   │               └── [   61]  FileUUIDs.log
        │   ├── [  55K]  METS.74cad4f0-d1d4-467c-88f8-f29f008d6387.xml
        │   └── [  122]  objects
        │       ├── [  23K]  lion-2ffb671c-c541-48d5-bd61-19a16c2461ee.svg
        │       ├── [  18K]  lion.svg
        │       ├── [   23]  metadata
        │       │   └── [   60]  transfers
        │       │       └── [   32]  oneimage2-d4a5c1ca-3393-4da8-bb7d-e2cad03b8175
        │       │           └── [  153]  directory_tree.txt
        │       └── [   69]  submissionDocumentation
        │           └── [   22]  transfer-oneimage2-d4a5c1ca-3393-4da8-bb7d-e2cad03b8175
        │               └── [  16K]  METS.xml
        ├── [ 1.4K]  manifest-sha256.txt
        └── [  238]  tagmanifest-sha256.txt

Note there are two "oneimage2-xxxxx" subdirectory levels (one with the original AIP uuid, and one with the uuid assigned for reingest). The double directory level is what will cause the reingest process to fail.

When a full reingest is triggered, the AIP files are copied/moved among locations following roughly the following flow (assuming a compressed AIP) ( ref. SS code for locations.models.package:package:start_reingest() here and AM code for archivematica.dashboard:views:reingest() here ):

1) The SS copies the AIP from the aipstore and extracts it to the Storage Service internal processing location 2) The SS moves the extracted AIP files to the staging path of the respective pipeline FS 3) The SS moves the extracted AIP files to the "currently processing" location of the corresponding pipeline 4) A reingest API call is made to the pipeline. The pipeline moves the extracted AIP files from the currently processing location to the pipeline watched directory

It looks like the problematic step is 2) above. By taking a look at the logs, the move is implemented via rsync, which seems to be creating the problematic double directory level structure (note the source parameter of rsync does not have a "/" at the end, meaning that it will make a copy of the source files including the enclosing directory):

DEBUG   2024-02-09 11:23:48  locations.models.package:package:start_reingest:2257:  Reingest: extracted to /var/archivematica/storage_service/tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387
INFO    2024-02-09 11:23:48  locations.models.package:package:start_reingest:2331:  Reingest: files: ['tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387']
DEBUG   2024-02-09 11:23:48  locations.models.package:package:start_reingest:2347:  Reingest: Current location: 413d6e6f-dcad-4937-8179-544a2020c28d: var/archivematica/storage_service (Storage Service Internal Processing)
DEBUG   2024-02-09 11:23:48  locations.models.space:space:move_to_storage_service:341:  TO: src: var/archivematica/storage_service/tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387
DEBUG   2024-02-09 11:23:48  locations.models.space:space:move_to_storage_service:342:  TO: dst: tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387
DEBUG   2024-02-09 11:23:48  locations.models.space:space:move_to_storage_service:343:  TO: staging: /mnt/pfs_staging
INFO    2024-02-09 11:23:48  locations.models.space:space:move_rsync:545:  Moving from /var/archivematica/storage_service/tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387 to /mnt/pfs_staging/tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387
INFO    2024-02-09 11:23:48  locations.models.space:space:move_rsync:587:  rsync command: ['rsync', '-t', '-O', '--protect-args', '-vv', '--chmod=Fug+rw,o-rwx,Dug+rwx,o-rwx', '-r', '/var/archivematica/storage_service/tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387', '/mnt/pfs_staging/tmpztdld58u/oneimage2-74cad4f0-d1d4-467c-88f8-f29f008d6387']

In some installations, the SS internal processing location matches the value for fhe pipeline FS staging path (e.g., both are set to /var/archivematica/storage_service), in this case step 2) above does not trigger a move (i.e., rsync not run) and the problem is avoided. This (i.e., setting the value for fhe pipeline FS staging path to be the same as the SS internal processing location) could be the simplest workaround for this issue until the bug is fixed.

For Artefactual use:

Before you close this issue, you must check off the following:

[ ] All pull requests related to this issue are properly linked
[ ] All pull requests related to this issue have been merged
[ ] A testing plan for this issue has been implemented and passed (testing plan information should be included in the issue body or comments)
[ ] Documentation regarding this issue has been written and merged (if applicable)
[ ] Details about this issue have been added to the release notes (if applicable)

archivematica / Issues

Problem: Full reingest fails when Storage Service and Pipeline are on separate servers (under some system configuration conditions) #1663