Task #12023 (closed)
Opened 10 years ago
Closed 9 years ago
Bug: lost file annotations and instrument when moving images
Reported by: | sbesson | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | 5.1.1 |
Component: | General | Version: | 4.4.8 |
Keywords: | n.a. | Cc: | ux@… |
Resources: | n.a. | Referenced By: | n.a. |
References: | n.a. | Remaining Time: | n.a. |
Sprint: | n.a. |
Description
Tested on a 4.4.8p1 production server, user moved a dataset with LIF images containing file annotations created by himself and another user to another group (where both users were belonging to).
Images and image owner's annotations were successfully moved. However, other user's annotations were not moved and ended up orphaned in the source group. Additionally, some of the instrument components were not chgrped ending up in
serverStackTrace = "ome.conditions.SecurityViolation: Cannot read ome.model.acquisition.Instrument:Id_1522
errors when loading the image in the target group.
Chgrping the dataset back to the original group restored the image-instrument link.
Change History (12)
comment:1 Changed 9 years ago by jburel
- Cc ux@… added
comment:2 Changed 9 years ago by mtbcarroll
It would be interesting to try to reproduce this with 5.1.0-m2. I assume the images are being moved among read-annotate groups?
comment:3 Changed 9 years ago by sbesson
Double checked. I don't have the production logs anymore but I can confirm the source and target groups were read-annotate.
comment:4 Changed 9 years ago by mtbcarroll
Which image is this? You just import it, you both add file annotations, then try moving it?
comment:5 Changed 9 years ago by sbesson
Pretty sure this is a DV file, would you have a query for retrieving an image ID given an instrument ID?
comment:6 Changed 9 years ago by mtbcarroll
Perhaps,
bin/omero hql 'select id, name from Image where instrument.id = 123'
comment:7 Changed 9 years ago by mtbcarroll
Regarding annotations, https://github.com/openmicroscopy/openmicroscopy/pull/3292#issuecomment-69050602 may well be the same problem.
Petr, lately have you seen anything like the horrible problem with the instrument components? Perhaps that is fixed in the graphs reimplementation.
comment:8 Changed 9 years ago by mtbcarroll
- Milestone changed from 5.1.0 to 5.1.1
Barring scary regressions, or continued ability to separate images from instruments, not expecting further work on this for 5.1.0.
comment:9 Changed 9 years ago by pwalczysko
I have not seen any such instrument separations. Note that our permissions testing in Dec 2014 was performed mainly on Leica-lif files (MIF) with many metadata and instrument items. I have rerun one such move from ra1 to ra2 with annotated lif just now with following result:
- my own annotation persisted
- the annotation of the member who was not a member of the target group vanished from the moved image (expected)
- all the items in the Acquisition pane of this metadata-rich MIF member were preserved after the move
comment:10 Changed 9 years ago by jamoore
Referencing ticket #11752 has changed sprint.
comment:11 Changed 9 years ago by jamoore
Referencing ticket #11752 has changed sprint.
comment:12 Changed 9 years ago by mtbcarroll
- Resolution set to duplicate
- Status changed from new to closed
This ticket is a mix of unreproducible and things very much on our radar in Trello (especially object-sharing / deep copy). Closing to clear clutter.
Petr is currently preparing an improved permission sheet. cc ux, Mark and Josh are on that list