The intent of this Operator is to deploy common workloads to establish a performance baseline of your provider.
| Workload | Use? | Status in Operator |
|---|---|---|
| UPerf | Network Performance | Working |
| FIO | Storage IO | Working |
| Couchbase + YCSB | Databse Performance | WIP |
# git clone https://github.com/jtaleric/benchmark-operator
# cd benchmark-operator
# export KUBECONFIG=<your_kube_config>
# oc new-project benchmark
# oc project benchmark
# oc create -f deploy/role.yaml
# oc create -f deploy/role_binding.yaml
# oc create -f deploy/service_account.yaml
# oc create -f deploy/crd/bench_v1alpha1_bench_crd.yaml
# oc create -f deploy/operator.yamlAt this point, you can modify `deploy/crd/bench_v1alpha1_bench_cr.yaml
apiVersion: bench.example.com/v1alpha1
kind: Bench
metadata:
name: example-bench
spec:
uperf:
# Server size must always be 1 or more
pairs: 1
proto: tcp
test_type: stream
nthr: 2
size: 16384
runtime: 60Deploying the above will result in :
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE
benchmark-operator-6ff5bf5db8-nzvl8 1/1 Running 0 2m12s 10.129.2.186 ip-10-0-152-138.us-west-2.compute.internal <none>
example-bench-uperf-client-bench-7f8fb9bc8-q2xbx 1/1 Running 0 101s 10.129.2.188 ip-10-0-152-138.us-west-2.compute.internal <none>
example-bench-uperf-server-bench-68779b4986-nz5w8 1/1 Running 0 117s 10.129.2.187 ip-10-0-152-138.us-west-2.compute.internal <none>
The first pod is our Operator orchestrating the UPerf workload.
To review the results, oc logs <client>, the top of the output is
the actual workload that was passed to UPerf (From the values in the custom resource)
<?xml version=1.0?>
<profile name="tcp-stream-16384B-2i">
<group nthreads="2">
<transaction iterations="1">
<flowop type="connect" options="remotehost=$h protocol=tcp"/>
</transaction>
<transaction duration="60">
<flowop type=write options="count=16 size=16384"/>
</transaction>
<transaction iterations="1">
<flowop type=disconnect />
</transaction>
</group>
</profile>
Starting 2 threads running profile:tcp-stream-16384b-2i ... 0.00 seconds
Txn1 0 / 0.00(s) = 0 0op/s
Txn1 0 / 1.00(s) = 0 2op/s
Txn2 0 / 0.00(s) = 0 0op/s
Txn2 1.23GB / 1.02(s) = 10.30Gb/s 78563op/s
Txn2 2.22GB / 2.03(s) = 9.39Gb/s 71623op/s
Txn2 3.06GB / 3.09(s) = 8.51Gb/s 64889op/s
Txn2 4.17GB / 4.09(s) = 8.75Gb/s 66770op/s
Txn2 4.92GB / 5.20(s) = 8.12Gb/s 61941op/s
Txn2 5.17GB / 6.21(s) = 7.16Gb/s 54644op/s
Txn2 5.71GB / 7.21(s) = 6.81Gb/s 51957op/s
Txn2 6.57GB / 8.21(s) = 6.87Gb/s 52436op/s
Txn2 7.43GB / 9.22(s) = 6.93Gb/s 52851op/s
Txn2 8.40GB / 10.22(s) = 7.06Gb/s 53896op/s
Txn2 9.46GB / 11.22(s) = 7.24Gb/s 55239op/s
... Trimmed ...
Txn2 54.98GB / 56.48(s) = 8.36Gb/s 63798op/s
Txn2 55.72GB / 57.48(s) = 8.33Gb/s 63524op/s
Txn2 56.34GB / 58.49(s) = 8.27Gb/s 63117op/s
Txn2 57.26GB / 59.70(s) = 8.24Gb/s 62863op/s
Txn3 0 / 0.00(s) = 0 0op/s
Txn3 0 / 0.00(s) = 0 0op/s
-------------------------------------------------------------------------------
Total 57.26GB / 61.80(s) = 7.96Gb/s 60722op/s
Netstat statistics for this run
-------------------------------------------------------------------------------
Nic opkts/s ipkts/s obits/s ibits/s
eth0 47663 2575 7.98Gb/s 1.37Mb/s
-------------------------------------------------------------------------------
Run Statistics
Hostname Time Data Throughput Operations Errors
-------------------------------------------------------------------------------
10.129.2.187 61.80s 57.26GB 7.96Gb/s 3752501 0.00
master 61.80s 57.26GB 7.96Gb/s 3752900 0.00
-------------------------------------------------------------------------------
Difference(%) -0.00% 0.01% 0.01% 0.01% 0.00%