In a load-balanced multi-middle-tier-server Cloud cluster, requests that take longer than about a minute (a slow search, a long-running report order) silently stop with no error message, even though the same request completes fine on a single, non-load-balanced server.

All FAQ questions / Developer Studio, deployment, and upgrades

Answer

Check the load balancer's own connection/timeout settings rather than IFS Cloud's own configuration first — a load balancer sitting in front of multiple middle-tier servers can have its own default idle-connection timeout and per-node reply-wait timeout that's shorter than some genuinely slow requests need, silently dropping the connection right around that threshold (commonly close to a minute by default) with no error surfaced back to the user; increasing the load balancer's virtual-server timeout and its pool-level max-reply-time setting to accommodate your slowest expected requests has resolved this exact class of silent one-minute cutoff.

Documentation-based answer. Confirm the product and release before applying it. This answer has not been verified in an IFS environment.

← Browse Developer Studio, deployment, and upgrades questions