What is the best MPIO policy to avoid iSCSI session interruptions during ONTAP node takeover or upgrade
Applies to
- ONTAP 9
- Hyper-V Hosts using Failover Clustering
- iSCSI
Answer
- Recommended MPIO Policy: Failover Only
- Behavior:
- Maintains a single primary path until a failure occurs.
- Backup paths are only activated during a failover event.
- Advantages:
- Reduces the likelihood of path re-negotiation during a takeover, ensuring session stability.
- Ideal for environments requiring uninterrupted iSCSI sessions.
- Use Case:
Best suited for environments where session hiccups or interruptions are not tolerable.
- Alternative MPIO Policy: Low Queue Depth (LQD)
- Behavior:
Operates in active/active mode, distributing I/O across multiple paths with low queue depth.
- Advantages:
Balances load across active paths, improving performance in some scenarios.
- Disadvantages:
May still experience path re-negotiation during takeover or failover events, leading to brief hiccups.
- Use Case:
Suitable for environments where minor interruptions are acceptable and performance optimization is a priority.
- Avoid Using Round Robin with Subset (RRWS)
- Behavior:
Focuses on distributing I/O across a subset of optimal paths.
- Disadvantages:
More prone to session interruptions during takeover or failover events.
- Recommendation:
Not recommended for environments requiring high session stability.
NOTE:Before deploying the chosen MPIO policy in production, perform thorough testing in your environment to ensure it meets your stability and performance.
