A customer experiences a sudden loss in write performance on the HP 3PAR StoreServ array on the primary site. Reads are not affected, but writes slowed down significantly. What should you advise the…

Host multi-pathing issues typically affect overall I/O access or specific paths, not usually causing a sudden, specific degradation only in writes while reads remain unaffected.

Changing from asynchronous periodic to synchronous replication would increase the latency and performance burden on primary site writes, potentially worsening the problem rather than solving it.

If the array is configured for synchronous replication, all writes on the primary site must be committed to the secondary site before the primary host receives an acknowledgment. Any latency, congestion, or issues on the replication link or the remote array can directly slow down primary site write performance. Switching to asynchronous periodic mode decouples the primary site writes from the replication latency, allowing writes to be acknowledged locally first, which would immediately alleviate the write performance bottleneck.

While Fibre Channel infrastructure issues can cause performance problems, if reads are unaffected and only writes are significantly slowed, the issue points more specifically to a write-dependent process like synchronous replication rather than a general hardware malfunction affecting all I/O paths.