Issue: Processing Storage Usage Retrieves Statically Coded Storage Values

[ross-spencer]

The function that retrieves the processing storage usage returns values for a statically coded location.

It is picking up the values from the dashboard django settings: here

A function more analogous to Sources(...) May be preferable as it talks to the Storage Service to retrieve its values dynamically.

Other impacts of this issue:

• Providing statistics to the end-user
• Providing the ability to clear a location via this interface. The folder will not exist, e.g.

A workaround for now is to use the operating system capabilities to allow continued maintenance on non-default locations that have been configured.

Credit for spotting this issue to Radim on the Google Group.

[ross-spencer]

Testing whether this has been considered with the disk-full work already merged into qa/1.x for 1.11, and we can see the issue is still there:

1, Modify this line. 2, Reload the administration/usage/ page, and observe the changes are picked up by the UI.

[sromkey]

@jraddaoui could this be addressed under the umbrella of Disk Full?

[jraddaoui]

Hi @ross-spencer and @sromkey,

This is what I was trying to explain with the phrase we removed from the disk full PR. It's a little hard to explain ...

The first problem we have is that we declare the same setting in three different places:

• Dashboard: https://github.com/artefactual/archivematica/blob/qa/1.x/src/dashboard/src/settings/base.py#L117
• MCPServer: https://github.com/artefactual/archivematica/blob/qa%2F1.x/src/MCPServer/lib/settings/common.py#L132
• MCPClient: https://github.com/artefactual/archivematica/blob/qa%2F1.x/src/MCPClient/lib/settings/common.py#L174

Changing this setting in the dashboard configuration will not break the entire pipeline (as I wrongly said in the disk full PR). However, this setting is passed to the SS when the pipeline is registered on it and that path is used in there to create the CP and other locations. Therefore, changing this setting after a pipeline is registered in the SS doesn't actually change where the CP location is located and it only breaks this page, so I'm not sure why would anyone change this only in the dashboard.

Nevertheless, if you change the setting in the MCPServer or MCPClient (after the pipeline is registered) the paths in the SS and the ones in these settings won't match and that's what will actually break the pipeline.

That being said, I think it's more important to unify the settings used by the three components and/or to document what is actually required to modify the watched directories configuration (three setting files or environment variables plus the changes required in the SS to make the locations associated with the pipeline match) than fixing an issue caused by modifying the configuration in the wrong manner.

Extra references: https://github.com/archivematica/Issues/issues/407 https://github.com/archivematica/Issues/issues/406 https://github.com/archivematica/Issues/issues/312#issuecomment-435979548 https://github.com/artefactual/archivematica/issues/1091

[ross-spencer]

Don't mind me Radda... I feel like it might be like this for you:

:smile:

Do you think we now have two or three variants of this same issue now, or if not, is there a way to unify them that's helpful? (Unify the issues, then unify the approach? :slightly_smiling_face: )

[jraddaoui]

A forum thread related to this issue: https://groups.google.com/forum/?utm_medium=email&utm_source=footer#!topic/archivematica/Xv3e1I35vN4