Write the import report somewhere the importer can write - #48
Merged
Conversation
The Linux verification failed where macOS had not. After reading all 3.4GB the import died with AccessDeniedException on /import/import.report, about a directory that plainly belongs to the person who ran it. The export is a bind mount owned by whoever downloaded it and the neo4j image runs as its own user, so on a rootful docker host those are two different users and the report is the one file in the whole import that cannot be written. Rootless podman on macOS maps the two together, which is why the same command worked there and hid this for a day. So the report goes on the volume instead, which the image owns by definition, and the export is mounted read only now that nothing writes to it. Both import scripts pass their own arguments through to neo4j-admin, where a repeated option takes the last value, which is how the container run moves the report without the export format knowing anything about containers. The published dataset is republished as 2026.08.1, because the scripts inside it are part of it.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Found by running the installer on a real Linux host, which is the only place it could have been found.
macOS passed. Linux read all 3.4GB and then died:
about a directory that plainly belongs to the person who ran the command.
The export is a bind mount owned by whoever downloaded it, and the neo4j image runs as its own user.
On rootless podman, which is what macOS uses, those two are mapped together and everything works.
On a rootful docker host they are different users, and
import.reportis the one file in the whole import that cannot be written.Everything else the import produces goes into the volume, which the image owns.
So the report goes into the volume too, and the export is mounted read only now that nothing writes to it.
3.4GB somebody waited to download is worth not being able to damage.
Moving the report needs a flag, and the flag has to come from the caller rather than from the export, because the export format should not know that anybody runs it in a container.
Both
import.shandimport.cmdnow end by passing their own arguments through toneo4j-admin, where a repeated option takes the last value given.The container run appends
--report-file=/data/import.reportand nothing else changes.Verified by hand on server3 before the change was written: same image, same CSV files, read only mount, report on the volume, 8,175,346 nodes and 9,119,011 relationships in 1m 47s.
The published dataset goes to
2026.08.1, because the scripts inside it are part of it.That is what the version being separate from the code version is for.