See
<https://builds.apache.org/job/beam_PerformanceTests_JDBC/10/display/redirect>
------------------------------------------
[...truncated 53.54 KB...]
containers in a privileged mode. Note that some benchmarks execute commands
which are only allowed in privileged mode.
(default: 'true')
--kubeconfig: Path to kubeconfig to be used by kubectl
(default: '')
--kubectl: Path to kubectl tool
(default: 'kubectl')
--kubernetes_nodes: IP addresses of Kubernetes Nodes. These need to be
accessible from the machine running Perfkit benchmarker. Example:
"10.20.30.40,10.20.30.41"
(default: '')
(a comma separated list)
--rbd_pool: Name of RBD pool for Ceph volumes.
(default: 'rbd')
--rbd_user: Name of RADOS user.
(default: 'admin')
--username: User name that Perfkit will attempt to use in order to SSH into
Docker instance.
(default: 'root')
perfkitbenchmarker.providers.mesos.flags:
--docker_cpus: CPU limit for docker containers.
(default: '1.0')
(a number)
--docker_memory_mb: Memory limit for docker containers.
(default: '2048')
(an integer)
--marathon_address: Marathon IP address and port.
(default: 'localhost:8080')
--[no]mesos_privileged_docker: If set to True, will attempt to create Docker
containers in a privileged mode. Note that some benchmarks execute commands
which are only allowed in privileged mode.
(default: 'false')
perfkitbenchmarker.providers.openstack.flags:
--openstack_additional_flags: Additional flags to pass to every OpenStack CLI
command. See "openstack --help" for more.
(default: '[]')
--[no]openstack_boot_from_volume: Boot from volume instead of an image
(default: 'false')
--openstack_cli_path: The path to the OpenStack CLI binary.
(default: 'openstack')
--[no]openstack_config_drive: Add possibilities to get metadata from external
drive
(default: 'false')
--openstack_floating_ip_pool: Name of OpenStack floating IP-address pool. If
set, a floating-ip address from this pool will be associatedto each instance
and will be used for communicating with it. To use this flag, an internally
routable network must also be specified via the openstack_network flag.
--openstack_image_username: Ssh username for cloud image
(default: 'ubuntu')
--openstack_network: Name of OpenStack network. This network provides
automatically allocated fixed-IP addresses to attached instances. Typically,
this network is used for internal communication between instances. If
openstack_floating_ip_pool is not set then this network will be used to
communicate with the instance.
(default: 'private')
--openstack_neutron_path: The path to the Neutron CLI binary.
(default: 'neutron')
--openstack_nova_path: The path to the Nova CLI binary.
(default: 'nova')
--openstack_private_network: (DEPRECATED: Use openstack_network) Name of
OpenStack private network.
(default: 'private')
--openstack_public_network: (DEPRECATED: Use openstack_floating_ip_pool) Name
of OpenStack public network.
--openstack_scheduler_policy: <None|affinity|anti-affinity>: Add possibility
to use affinity or anti-affinity policy in scheduling process
(default: 'None')
--openstack_volume_size: Size of the volume (GB)
(an integer)
--openstack_volume_type: Optional Cinder volume type to use.
perfkitbenchmarker.providers.profitbricks.flags:
--profitbricks_boot_volume_size: Choose the boot volume size in GB.
(default: '10')
(an integer)
--profitbricks_boot_volume_type: <HDD|SSD>: Choose between HDD or SSD boot
volume types.
(default: 'HDD')
--profitbricks_config: Path to config file containing your email and password.
Can also be set via $PROFITBRICKS_CONFIG environment variable.
(File format: email:password)
(default: '~/.config/profitbricks-auth.cfg')
--profitbricks_location: <us/las|de/fkb|de/fra>: Location of data center to be
provisioned (us/las, de/fkb, de/fra)
(default: 'us/las')
perfkitbenchmarker.providers.rackspace.flags:
--additional_rackspace_flags: Additional global flags to pass to every RackCLI
command. See "rack --help" for more.
(default: '')
(a comma separated list)
--rack_path: The path for the rack CLI binary.
(default: 'rack')
--rack_profile: A string indicating which RackCLI profile to use. If none is
specified default profile is used (see https://developer.rackspace.com/docs
/rack-cli/configuration/#config-file)
--[no]rackspace_boot_from_cbs_volume: When flag is included the instance will
use a remote disk as its boot disk, if machine_type supports it.
(default: 'false')
--rackspace_network_id: (EXPERIMENTAL) The ID of an already created network to
use instead of creating a new one. Must have a subnet already associated
with the network.
--rackspace_region: A string indicating which Rackspace region to use.
(default: 'IAD')
--[no]rackspace_use_security_group: (EXPERIMENTAL) A boolean indicating
whether or not to create a security group for the new instance. Applies
default security group rules (e.g. allow ingress TCP, and UDP traffic
through port 22). If no security group is used, all incoming and outgoing
traffic through TCP, UDP and ICMP is allowed, this is the default.
(default: 'false')
perfkitbenchmarker.publisher:
--bigquery_table: The BigQuery table to publish results to. This should be of
the form "[project_id:]dataset_name.table_name".
--bq_path: Path to the "bq" executable.
(default: 'bq')
--bq_project: Project to use for authenticating with BigQuery.
--cloud_storage_bucket: GCS bucket to upload records to. Bucket must exist.
--[no]collapse_labels: Collapse entries in labels in JSON output.
(default: 'true')
--csv_path: A path to write CSV-format results
--es_index: Elasticsearch index name to store documents
(default: 'perfkit')
--es_type: Elasticsearch document type
(default: 'result')
--es_uri: The Elasticsearch address and port. e.g. http://localhost:9200
--gsutil_path: path to the "gsutil" executable
(default: 'gsutil')
--[no]hostname_metadata: A boolean indicating whether to publish VM hostnames
as part of sample metadata.
(default: 'false')
--json_path: A path to write newline-delimited JSON results Default: write to
a run-specific temporary directory
--metadata: A colon separated key-value pair that will be added to the labels
field of all samples as metadata. Multiple key-value pairs may be specified
by separating each pair by commas.;
repeat this option to specify a list of values
(default: '[]')
--[no]official: A boolean indicating whether results are official or not. The
default is False. Official test results are treated and queried differently
from non-official test results.
(default: 'false')
--product_name: The product name to use when publishing results.
(default: 'PerfKitBenchmarker')
--service_account: Service account to use to authenticate with BQ.
--service_account_private_key: Service private key for authenticating with BQ.
perfkitbenchmarker.spark_service:
--spark_static_cluster_id: If set, the name of the Spark cluster, assumed to
be ready.
perfkitbenchmarker.stages:
--run_stage: The stage or stages of perfkitbenchmarker to run.
(default: 'provision,prepare,run,cleanup,teardown')
(A complete benchmark execution consists of 5 stages: provision, prepare,
run, cleanup, teardown. Possible flag values include an individual stage, a
comma-separated list of stages, or 'all'. If a list of stages is provided,
they must be in order without skipping any stage.)
perfkitbenchmarker.static_virtual_machine:
--static_vm_tags: The tags of static VMs for PKB to run with. Even if other
VMs are specified in a config, if they aren't in this list they will be
skipped during VM creation.
(a comma separated list)
perfkitbenchmarker.timing_util:
--timing_measurements: Comma-separated list of values from
<none|end_to_end_runtime|runtimes|timestamps> that selects which timing
measurements to enable. Measurements will be included as samples in the
benchmark results. none: No measurements included (same as providing an
empty list, and cannot be combined with other options). end_to_end_runtime:
Includes an end-to-end runtime measurement. runtimes: Includes runtimes of
all measured intervals, including the end-to-end runtime, the time taken by
the benchmark module Prepare, Run, and Cleanup functions, and other
important intervals. timestamps: Includes start and stop timestamps of all
measured intervals.
(default: 'end_to_end_runtime')
(a comma separated list)
perfkitbenchmarker.traces.collectd:
--[no]collectd: Install and run collectd on the guest.
(default: 'false')
--collectd_output: Path to store collectd results.
perfkitbenchmarker.traces.dstat:
--[no]dstat: Run dstat (http://dag.wiee.rs/home-made/dstat/) on each VM to
collect system performance metrics during each benchmark run.
(default: 'false')
--dstat_interval: dstat sample collection frequency, in seconds. Only
applicable when --dstat is specified.
(an integer)
--dstat_output: Output directory for dstat output. Only applicable when
--dstat is specified. Default: run temporary directory.
--[no]dstat_publish: Whether or not publish dstat statistics.
(default: 'false')
perfkitbenchmarker.virtual_machine:
--vm_metadata: Metadata to add to the vm via the provider's AddMetadata
function. It expectskey:value pairs
(default: '')
(a comma separated list)
perfkitbenchmarker.vm_util:
--background_cpu_threads: Number of threads of background cpu usage while
running a benchmark
(an integer)
--background_network_ip_type: <INTERNAL|EXTERNAL>: IP address type to use when
generating background network traffic
(default: 'EXTERNAL')
--background_network_mbits_per_sec: Number of megabits per second of
background network traffic to generate during the run phase of the benchmark
(an integer)
--burn_cpu_seconds: Amount of time in seconds to burn cpu on vm before
starting benchmark
(default: '0')
(an integer)
--burn_cpu_threads: Number of threads to use to burn cpu before starting
benchmark.
(default: '1')
(an integer)
--default_timeout: The default timeout for retryable commands in seconds.
(default: '1200')
(an integer)
--ip_addresses: <REACHABLE|BOTH|INTERNAL|EXTERNAL>: For networking tests: use
both internal and external IP addresses (BOTH), external and internal only
if the receiving VM is reachable by internal IP (REACHABLE), external IP
only (EXTERNAL) or internal IP only (INTERNAL)
(default: 'REACHABLE')
perfkitbenchmarker.windows_packages.ntttcp:
--ntttcp_threads: The number of client and server threads for NTttcp to run
with.
(default: '1')
(an integer)
--ntttcp_time: The number of seconds for NTttcp to run.
(default: '60')
(an integer)
gflags:
--flagfile: Insert flag definitions from the given file into the command line.
(default: '')
--undefok: comma-separated list of flag names that it is okay to specify on
the command line even if the program does not define a flag with that name.
IMPORTANT: flags in this list that have arguments MUST use the --flag=value
format.
(default: '')
Build step 'Execute shell' marked build as failure
