Question
TankError: Could not find core configuration for environment – Core Cache Inconsistency After Failed Upgrade
Kavi Dey
0 reputation · 23 Dec 2020, 09:26 UTC
36.6K views
Loading…
0 reputation · 23 Dec 2020, 09:26 UTC
The goal is to understand how to restore normal tank operation after a partially applied tk-core upgrade leaves the core descriptor cache in an inconsistent state, causing the error "Could not find core configuration for environment" on subsequent launches.
Constraints include the persistence of mismatched version descriptors in ~/.shotgun/tank/cache/core/ after a rollback via pip install, the absence of a documented tank subcommand to validate or rebuild this cache, and the potential for concurrent upgrades on shared NFS home directories to further corrupt bootstrap state. It is unclear whether any automated recovery procedure exists that avoids manual cache deletion.
Is there a supported tank subcommand to validate or rebuild the core descriptor cache? If not, what is the recommended procedure to safely recover without manually removing the cache directory? How does concurrent upgrade activity on NFS affect cache integrity, and what precautions should be taken to prevent recurrence?
The error \"Could not find core configuration for environment\" appears when Tank\'s internal core descriptor cache at ~/.shotgun/tank/cache/core/ contains files referencing a tk-core version that no longer matches the installed Shotgun environment after a partial upgrade.
During a tk-core upgrade, descriptor files may be written or updated before the full version migration completes. If the process is interrupted or rolled back via pip install, the cache directory retains the older or mismatched descriptors, and Tank reads these on startup, producing the configuration-not-found error.
shotgridtank subcommand exists to validate or rebuild the core descriptor cache; help output and release notes do not include a cache-clear or cache-rebuild option.~/.shotgun/tank/cache/core/ after a pip install rollback because pip does not automatically purge pre-existing core cache files during version downgrade.pip show tk-core.~/.shotgun/tank/cache/core/ and compare descriptor filenames or any version-index JSON present.~/.shotgun/tank/cache/core/ only after confirming no custom pipeline configurations are stored exclusively in that directory.When Tank home directories reside on NFS, schedule upgrade windows so only one user or process modifies tk-core and the core cache at a time. Use a local temporary copy of the tk-core package during upgrades, then propagate the cleaned cache directory to the shared location after verification.
\nTo determine whether partial descriptor regeneration is possible without full cache deletion, please provide the Tank build number and a listing of ~/.shotgun/tank/cache/core/ showing whether a version-index JSON file or specific descriptor filenames are present. This single detail will clarify if a targeted rebuild can be attempted or if cache clearing is required.
Use comments to ask for clarification. Post a solution as an answer.
No question comments on this page.