Dell Lab Deployment

Use this guide when running StrataBench on Dell lab VMs (static IPs, shared storage cluster).

Full automation: see LAB-BOOTSTRAP.md — stratabench lab bootstrap installs warp/fio/agent, deploys MinIO, and runs benchmarks from lab.yaml.

Topology

Jump host (coordinator)
  ├── stratabench CLI
  └── stratabench-api (optional, :8080)

Client VMs (3+)
  └── stratabench-agent (:7777)

Storage targets
  ├── Block: /dev/nvme0n1 or array LUN
  ├── File: NFS/Lustre mount
  └── Object: MinIO/S3 endpoint

Ports

Service Port Purpose
stratabench-agent 7777 Remote benchmark execution
stratabench-api 8080 REST API + /metrics
Prometheus scrape 8080 GET /metrics
MinIO/S3 9000 Warp profiles

Open firewall rules between coordinator ↔ clients on 7777 and 8080 (if using API).

Setup

1. Build on Linux (or cross-compile)

make build

Copy bin/stratabench, bin/stratabench-agent, and profiles/ to each VM.

2. Start agents on clients

export STRATABENCH_AGENT_LISTEN=:7777
./bin/stratabench-agent

Verify from coordinator:

curl http://10.0.1.1:7777/v1/health

3. Run distributed benchmark

./bin/stratabench run \
  --profile nvme-random-oltp \
  --target /dev/nvme0n1 \
  --clients 10.0.1.1:7777,10.0.1.2:7777,10.0.1.3:7777

4. S3 cluster (Warp)

export WARP_ACCESS_KEY=...
export WARP_SECRET_KEY=...
./bin/stratabench run \
  --profile s3-cluster-put-get \
  --target 10.0.1.10:9000 \
  --clients 10.0.1.1:7777,10.0.1.2:7777

5. API + Prometheus (optional)

./bin/stratabench-api
# GET  http://<host>:8080/api/v1/health
# POST http://<host>:8080/api/v1/runs  {"profile":"ssd-random-4k","target":"/dev/sdb","mock":false}
# GET  http://<host>:8080/metrics

Import Grafana dashboard from deploy/grafana/stratabench-dashboard.json.

Profile Layer Engine Notes
nvme-random-oltp block fio OLTP-style 16K random
nvme-max-stress block mock/stratabench Heavy QD stress
spdk-nvme-peak block SPDK Userspace peak IOPS
vm-disk-random vm-block fio Guest disk on HCI
s3-cluster-put-get object warp Multi-node S3
s3-cluster-rdma object warp S3 with RDMA (--rdma=cpu)
afa-multi-lun block vdbench Multi-LUN AFA random read
file-parallel-read file elbencho NFS/Lustre parallel read

Cross-layer analysis

Run multiple layers and compare bottlenecks:

./bin/stratabench cross-layer \
  --profiles nvme-random-oltp,s3-put-throughput \
  --target /dev/nvme0n1

SBK import

Import legacy SBK CSV results:

./bin/stratabench import sbk /path/to/sbk-results.csv

Troubleshooting