Five FlexCache volumes are deployed across three clusters to accelerate reads from a single origin volume. A site reports stale reads after an application updated files at the origin. What is the most probable cause?

Five FlexCache volumes are deployed across three clusters to accelerate reads from a single origin volume. A site reports stale reads after an application updated files at the origin. What is the most probable cause?

• A. FlexCache volumes do not automatically invalidate stale data after origin changes.
• B. The inter-cluster peering network was interrupted, delaying invalidation messages.
• C. Each FlexCache volume must manually synchronize metadata nightly.
• D. FlexCache volumes cannot detect origin volume updates across clusters.

Answer: (B)

When origin data changes, FlexCache relies on invalidation messages sent across the inter-cluster peering network to tell every cache that its locally stored copy is no longer valid. If that peering pathway is disrupted, the invalidations don’t reach the caches promptly, so the caches keep serving the stale data until the changes are eventually revalidated or the cached items expire. That makes disruption of the inter-cluster peering network the most likely cause in this scenario: the caches across the clusters aren’t immediately invalidated after the origin update, leading to stale reads being served. The idea that invalidation doesn’t happen automatically isn’t accurate—FlexCache is designed to invalidate cached data automatically when the origin changes. Manual nightly synchronization isn’t how FlexCache operates, and the system can detect origin updates across clusters provided the invalidation messages can be delivered; if the network blocks those messages, detection is effectively delayed.

When origin data changes, FlexCache relies on invalidation messages sent across the inter-cluster peering network to tell every cache that its locally stored copy is no longer valid. If that peering pathway is disrupted, the invalidations don’t reach the caches promptly, so the caches keep serving the stale data until the changes are eventually revalidated or the cached items expire.

That makes disruption of the inter-cluster peering network the most likely cause in this scenario: the caches across the clusters aren’t immediately invalidated after the origin update, leading to stale reads being served.

The idea that invalidation doesn’t happen automatically isn’t accurate—FlexCache is designed to invalidate cached data automatically when the origin changes. Manual nightly synchronization isn’t how FlexCache operates, and the system can detect origin updates across clusters provided the invalidation messages can be delivered; if the network blocks those messages, detection is effectively delayed.