We recently imported approximately 850 product images into our Design review workspace using the Agile Admin Console batch import feature. The import completed successfully according to the logs, but the thumbnails are still showing the old placeholder images in the viewer. I’ve checked the thumbnail generation job status in the Admin Console and it shows as ‘Completed’ but with warnings. The file permissions on our image repository look correct (755 for directories, 644 for files). We’re also seeing inconsistent behavior where some thumbnails update after a browser refresh while others remain stale. Has anyone dealt with thumbnail cache management issues after bulk imports? Our design review process is completely blocked as engineers can’t visually identify components.
Here’s the complete solution based on your symptoms:
1. Thumbnail Generation Job Status Analysis: First, review the detailed warnings in your completed job. Navigate to Admin Console > System Settings > Jobs and click on the thumbnail generation job to see the full log. Look for ‘IOException’ or ‘DiskFullException’ entries - these confirm the cache capacity issue you discovered.
2. File Permissions and Repository Verification:
While your base permissions are correct, verify the Agile service account has permissions on the ENTIRE path hierarchy. Run: namei -l /path/to/image/repository to check each directory level. Also confirm the thumbnail subdirectory isn’t mounted on a separate filesystem that’s full.
3. Thumbnail Cache Management - The Core Fix:
Immediate Actions:
-
Stop Agile application services
-
Clear the thumbnail cache directory completely (backup first if needed)
-
Increase cache size in agile.properties: thumbnail.max.cache.size=10GB (adjust based on your image volume)
thumbnail.cleanup.threshold=80
-
Verify disk space on the cache partition (you need at least 20% free)
-
Restart services
Post-Restart Steps:
- Run forced thumbnail regeneration: Admin Console > Content Management > Regenerate Thumbnails
- Select your 850 imported images by date range
- Monitor the job completion - should finish without warnings this time
- Verify thumbnails appear correctly in Design Review
Long-term Prevention:
- Set up a scheduled task to clean thumbnails older than 90 days
- Monitor cache usage with alerts at 75% capacity
- Consider implementing a CDN or separate image server for large deployments
- Review and archive old design images that are no longer actively used
For Multi-Server Environments: If you’re running clustered application servers, ensure the thumbnail cache is on shared storage (NFS/SAN) or implement cache synchronization. Each server trying to maintain its own cache with 850 new images will multiply your storage requirements.
The inconsistent behavior you saw (some thumbnails updating, others not) is the classic symptom of cache thrashing when near capacity. The system was successfully generating some thumbnails but immediately having to evict them due to space constraints, creating the random update pattern.
This draft is based on general Oracle Agile PLM knowledge. It has not been verified against your specific version and environment. Practitioners: verify the steps and share your experience below.
Check the thumbnail cache directory on your application server. The path is usually defined in agile.properties under thumbnail.cache.directory. Sometimes the batch import process doesn’t properly invalidate the cache, especially if you’re running multiple application servers. Try manually clearing the cache directory and restarting the Agile services.
“Confirmed this resolves the thumbnail generation failure — clearing the Agile PLM cache partition and granting recursive permissions on the full repository path via namei -l fixed our batch import issues immediately.”
I’ve seen this exact issue before. The warnings in the thumbnail generation job usually indicate file access problems. Even though your permissions look correct at the OS level, check if the Agile application server process owner actually has write access to the thumbnail directory. Also verify that the image repository path in the database (FILESERVER table) matches your actual file system path. Run this query to check: SELECT SERVER_NAME, SERVER_URL, FILE_PATH FROM agile.fileserver WHERE SERVER_TYPE=‘Image’. Sometimes after infrastructure changes, these paths get out of sync and the thumbnail generator can’t write the processed images.
Thanks both. I checked the thumbnail cache directory and found it was nearly at capacity (98% full). The file path in the FILESERVER table matches our actual repository location. The application server owner does have write permissions. Could the full cache be causing the inconsistent behavior?
Absolutely yes. When the thumbnail cache reaches capacity, Agile’s behavior becomes unpredictable. It tries to clean up old thumbnails but the cleanup process can fail silently if multiple users are accessing the same cached images. You need to increase the cache size or implement a regular cleanup schedule.
Adding to what others said - after batch imports, I always run a manual thumbnail regeneration for the affected items. You can do this through the Admin Console by selecting the specific items and forcing a thumbnail rebuild. This bypasses the normal cache logic and ensures fresh thumbnails are generated. Also check your thumbnail.max.cache.size property - default is usually too small for production environments with heavy image usage.