Details
Description
ATM if we have cluster of YES node we may have outdated data for prices and inventory.
Therefore the approach is to keep those caches at minimal time frame to keep information fresh but at the same time if we do get out of synch on nodes we can forcefully flush inventory caches when we encounter out of stock exceptions during checkout.
Alternative would be to have each YES node broadcast WS onCacheableEntity event which may provide better synchronisation but at the cost of WS traffic, which seems unreasonable at this point.
Attachments
Issue Links
- links to
-
YC-296 Yes cache flush and reindex should propagate to other Yes nodes
- Closed