Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

make temporary path for some file operations configurable #7664

Closed
donsizemore opened this issue Mar 9, 2021 · 6 comments
Closed

make temporary path for some file operations configurable #7664

donsizemore opened this issue Mar 9, 2021 · 6 comments

Comments

@donsizemore
Copy link
Contributor

Related to #3818:

Dataverse relies on tmpcleaner / systemd-tempfiles.timer to clean up temporal detritus.

Suppose a system were set up as a virtual machine with disk partitions but not LVM. It would be a nice feature for administrators of such a system to be able to configure Dataverse to use $files.dir/temp instead, to avoid the possibility of exhausting disk space in the root partition during the upload and processing of large files.

@poikilotherm
Copy link
Contributor

@donsizemore Is this a part of #6656, related only or sth. completely different?

@qqmyers
Copy link
Member

qqmyers commented Mar 9, 2021

Currently, Dataverse uses temp file space directed by two jvm options:
java.io.tmpdir and dataverse.files.directory. If you set both of those, I think you can redirect all temp files. (TDL does this to put temp files on an EBS volume).

@poikilotherm
Copy link
Contributor

Jim @qqmyers while I agree on the directory from <dataverse.files.directory>/temp holding data during ingest, #6656 was about the issues experienced during upload before ingest.

I recently looked into the root cause for this and you cannot set this via java.io.tmpdir (but lets keep that conversation in the other issue).

@qqmyers
Copy link
Member

qqmyers commented Mar 9, 2021

@poikilotherm - you're right of course. java.io.tmpdir is useful to redirect as it is used to store temp files related to ingest, thumbnails and libraries doing full-text indexing, etc., but it doesn't catch that first upload step. the dataverse temp dir should not be accumulating many files these days (i.e. if someone uploads files and doesn't hit save or cancel for the upload - most other use cases we manage to delete the temp files now). java.io.tmpdir is never cleaned out by Dataverse afaik, so it does accumulate unless periodically cleaned.

@poikilotherm
Copy link
Contributor

So I still wonder if @donsizemore is asking for a cleaning job inside the Dataverse codebase that might take care of this.

@donsizemore
Copy link
Contributor Author

@poikilotherm I was planning/fretting ahead for a hypothetical situation, but your refactoring proposal sounds like the real solution.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants