IBM AIX 7.3 – Common Problem 147 – Diagnosis and Fix

📖 ~1 min read

Table of contents

1. Symptom & Impact
2. Environment & Reproduction
3. Root Cause Analysis
4. Quick Triage
5. Step-by-Step Diagnosis
6. Solution – Primary Fix
7. Solution – Alternative Approaches
8. Verification & Acceptance Criteria
9. Rollback Plan
10. Prevention & Hardening
11. Related Errors & Cross-Refs
12. References & Further Reading

Symptom & Impact

Client LPAR sees `iostat` latency >50ms on vSCSI disks while the SAN side looks healthy.

Environment & Reproduction

AIX 7.3 client LPAR using vSCSI through a single VIOS with backing storage on hdisk LUNs.

Root Cause Analysis

VIOS `vhost` mapping is bottlenecked on a single CPU, or `reserve_policy=single_path`.

Quick Triage

On VIOS run `lsmap -all`, `viostat -RDTl`, and `topas -C` to inspect host workload.

Step-by-Step Diagnosis

Capture `errlog` on the VIOS and `lsattr -El vscsi0` on the client.

Solution – Primary Fix

Set `reserve_policy=no_reserve` on backing hdisks: `chdev -l hdiskN -a reserve_policy=no_reserve`.

Solution – Alternative Approaches

Add a second VIOS path: `mkvdev -vdev hdiskN -vadapter vhostX` from VIOS2 and `cfgmgr` the client.

Verification & Acceptance Criteria

Client `iostat -DR` latency drops below 10ms and `vscsi1` shows active throughput.

Rollback Plan

Roll back vSCSI mapping with `rmvdev -vtd vtscsiY` from VIOS2.

Prevention & Hardening

Always deploy dual VIOS, monitor with `nmon -A`, and pin VIOS CPU entitlement.

Related Errors & Cross-Refs

Pairs with `VSCSI_HARD_ERR` and high `vhost` queue depth in `viostat`.

Related tutorial: View the step-by-step tutorial for aix-7.3.

View all aix-7.3 tutorials on the Tutorials Hub →

Browse all common problems & solutions on the Tutorials Hub.

References & Further Reading

IBM Docs: PowerVM VIOS performance tuning, `viostat`, dual-VIOS design.