CLI Command Reference

Use float CLI commands to interact with the OpCenter; for example, to submit and manage jobs.

Version

The MMCloud CLI commands described here are consistent with the OpCenter release shown in the table.

MMCloud CLI	OpCenter Release	Date
FLOAT_v3.1.0-0293166-Jericho.bin	FLOAT_v3.1.0-0293166-Jericho.bin	2025-04-01T03:22:50Z

Usage

Use float in the following format.

float [global_flags] [command] [subcommand] [options]

Global Flags

Use global flags with any floatcommand or subcommand.

Flag	Usage	Definition
-a, --address ip_address	Connect to OpCenter server	IP address of OpCenter (default: localhost or last OpCenter IP address used)
-F, --format json | yaml | table	Specify format for output	Output format (default: yaml)
-h, --help	Display help	Help for MMCloud CLI
--logLevel log_level	Specify log level	Log level (default: info)
-p, --password password	Log in to OpCenter	Login password
--scroll	Enable scroll mode for multiple page output	Enable navigation (up, down, left, right) for displays that span multiple pages
-u, --username user_name	Log in to OpCenter	Login username
-v, --verbose on | off	Turn verbose mode on or off	Verbose mode setting (default: off)

Categories

The MMCloud CLI commands are grouped into categories as shown in the table. An alphabetical listing of the commands follows.

Category	Commands
Job Management	list, show, submit, cancel, migrate, modify, suspend, resume, snapshot, rerun, promote, hosts, cluster, project, queue, ptop
Job Status Monitoring	log, ps, top, df, mi
Authentication & Authorization	login, logout, secret, nis, ldap
User Management	user, group, quota-policy
OpCenter Core Services	image, gateway, template, config, license, report, storage, smtp, node
OpCenter Management Operations	status, version, release, restart
Other	queues, completion

float cancel (scancel)

Use the float cancel (or float scancel) command to cancel a job. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Cancel job	--filter filter	Filter to select jobs if --job not specified. A simple filter is [attribute][operator][value]. Use --filter multiple times to apply multiple filters combined with "and" operator. Create complex filters by combining simple filters using parentheses and operators (and, or). Use "? to match a single character and "*" to match multiple characters. Values are strings, datetimes, or numbers. Operators are: = for strings, datetimes, or numbers >, <, >=, <= for datetimes or numbers Attributes include: id name user status duration start lastUpdate update end tags envs exitCode output cpu memory dataVolumes publishes imageID cost category Examples: --filter status=Executing --filter timeRange=2010-10-22~ --filter "tags=kind:training,project:finance-*" --filter "((user=admin) and (status=Executing or status=Completed))"
		-f, --force	Automatically answer "yes" at confirmation prompt
		-j, --job job_id	Job to cancel if --filter not specified

Example

$ float cancel -j ctZLDo7OFG4BuJ8ytiTem
Warning: Are you sure you want to cancel this job? 

ID: ctZLDo7OFG4BuJ8ytiTem 
Name:   python-c5d.large 

All the related resources will be released. (yes/No): y
Request to cancel ctZLDo7OFG4BuJ8ytiTem has been submitted

float cluster

Use the float cluster command to manage High Performance Computing (HPC) clusters. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add (create, new)	Add new HPC cluster	-n cluster_name	Name of new HPC cluster
		-I, --appImage image1,image2,...	Container image(s) allowed for use in the cluster
		-A, --auth auth_provider_id	Identifier of authorization provider for cluster
		--clusterStatusCheckInterval duration	Interval (in minutes and seconds) between checks whether the cluster has the appropriate number of nodes
		-D, --dataVolume [size=vol_size,throughput=rate]:/data_dir or [mode=rw]vol_id :/container_mnt_pt/path or [mode=rw]nfs://ip_address/server_export_dir :/container_mnt_pt/path or [opts=xx,yy]lustre://dnsname/mountname/path :/container_mnt_pt/path	Data volume mounted by container. Use option multiple times to attach multiple data volumes. (New EBS volume) vol_size in GB, rate in Mbps (can omit), data_dir is data directory created (Existing EBS volume) vol_id specifies existing EBS volume, /container_mnt_pt/path is path to where EBS volume is mounted on container (NFS mount) ip_address is IP address of NFS server, /server_export_dir is directory exported by NFS server, and /container_mnt_pt/path is path to where NFS directory is mounted on container (Lustre FS) dnsname is the DNS name for the Lustre server and mountname is the name assigned to the Lustre file system
		--defaultQueue queue_name	Name of default queue
		--dns dns_ip	IP address of domain name server
		-H, --headNode stringArray	Configuration of head node in this format: "[mi=xxx,count=yyy,instType=zzz]"
		-m, --instType inst1,inst2,...	Allowed instance type(s) for any cluster nodes
		--jobCollectDeviation duration	Tolerance (in minutes and seconds) within which two timestamps are considered to be the same
		--jobCollectInterval duration	Interval (in minutes and seconds) between collection of information on jobs
		-l, --loginNode stringArray	Configuration of login node in this format: "[mi=xxx,count=yyy,instType=zzz,noPublicIP=true|false]"
		--metricsCollectInterval duration	Duration (in minutes and seconds) between collection of information regarding the operation of the cluster
		-M, --mi machine_id	Identifier of machine image to use in cluster nodes
		--noPublicIp	Flag to indicate that no public IP required for any cluster node. Confirm cloud services are accessible from these instances.
		--nodeReadyTimeout duration	Maximum time to wait for a cluster node to initialize
		--nodeThrottlingSize node_num	Maximum number of concurrent nodes to create to run one job
		--precheck	Run pre-check before adding node
		-Q, --qos cpu=cpu_num,mem=mem_num	CPU limit and memory limit (in GB) that each user can use for their jobs (default: no limits)
		-s, --scheduler scheduler_type	Scheduler type (currently, only slurm supported)
		-G, --securityGroup SG1,SG2,...	Security group(s) (or network tags in GCP) to add to the instance for the job
		--setupScript script_name	Setup script for cluster
		--smtp_id	SMTP server ID
		--storage storage_config	Storage configuration
		-N, --subnet subnet_id	Specify subnet (aws)or vSwitch (alicloud) to use (default: system chooses one for you)
		--timezone	Time zone of cluster (default "Local")
addNode (createNode, newNode)	Add new node, with specified parameters, to HPC cluster	-i, --id cluster_id	ID of HPC cluster
		-n, --name cluster_name	Name of HPC cluster
		--noPublicIp	Flag to indicate that no public IP required for this node. Confirm cloud services are accessible from this instance.
		-s, --nodeSpec node_spec	Node specification
		-q, --queue queue_name	If adding a compute node, name of queue associated with this node.
		-r, --role node_type	Type of node to add (default: login). Options are login or compute.
delete (del, rm, remove)	Delete an existing HPC cluster (by name or by ID)	-i, --id cluster_id	ID of HPC cluster to delete (if using cluster ID to identify cluster)
		-n, --name cluster_name	Name of HPC cluster to delete (if using cluster name to identify cluster)
deleteNode (delNode, rmNode, removeNode)	Delete an existing node (by name or by ID)	-H, --hosthost_id1,host_id2, ...	ID(s) of node(s) to delete
		-i, --id cluster_id	ID of HPC cluster that node(s) is(are) part of
		-n, --name cluster_name	Name of HPC cluster that node(s) is(are) part of
get	Get information on an existing HPC cluster (by name or by ID)	-i, --id cluster_id	ID of HPC cluster to query (if using cluster ID to identify cluster)
		-n, --name cluster_name	Name of HPC cluster to query (if using cluster name to identify cluster)
ls (list)	List available HPC clusters	-f, --filter filter_definition	Filter applied to list of all HPC clusters
		-s, --orderBy order_definition	Criterion for sorting filtered list
modify (update)	Modify an HPC cluster	-i, --id cluster_id	ID of HPC cluster to modify
		-n, --name cluster_name	Name of HPC cluster to modify
		-o, --overwrite	Overwrite existing fields (default: append to existing fields)
		-r, --reconfigure	Modify configuration, then reconfigure cluster immediately
		Any flag from float cluster add	Modify any flag used with float cluster add by including the flag, and its new value, with float cluster modify
nodes (nodels, nls)	List available nodes in an HPC cluster	-f, --filter filter_definition	Filter applied to list of all nodes in the cluster
		-i, --id cluster_id	ID of HPC cluster that nodes are part of
		-n, --name cluster_name	Name of HPC cluster that nodes are part of
		-s, --orderBy order_definition	Criterion for sorting filtered list
queues (queuels, qls)	List available queues in an HPC cluster	-f, --filter filter_definition	Filter applied to list of all queues in the cluster
		-i, --id cluster_id	ID of HPC cluster that queues are part of
		-n, --name cluster_name	Name of HPC cluster that queues are part of
		-s, --orderBy order_definition	Criterion for sorting filtered list
start (run, launch)	Start an HPC cluster	-i, --id cluster_id	ID of HPC cluster to start
		-n, --name cluster_name	Name of HPC cluster to start
stop (shutdown, suspend)	Stop an HPC cluster	-i, --id cluster_id	ID of HPC cluster to stop
		-n, --name cluster_name	Name of HPC cluster to reconfigure
reconfigure (refresh, apply)	Reconfigure an HPC cluster	-i, --id cluster_id	ID of HPC cluster to reconfigure
		-n, --name cluster_name	Name of HPC cluster to reconfigure

float completion

Use the float completion command to generate auto-completion script for use in current shell or in every shell started subsequently. Use a subcommand with the -h flag to display help on how to use the auto-completion script.

Subcommands	Usage	Option	Option Definition
bash	Generate auto-completion script for bash	--no-descriptions	Disable completion descriptions
fish	Generate auto-completion script for fish	--no-descriptions	Disable completion descriptions
powershell	Generate auto-completion script for powershell	--no-descriptions	Disable completion descriptions
zsh	Generate auto-completion script for zsh	--no-descriptions	Disable completion descriptions

float config

Use the float config command to view or change the OpCenter configuration. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
cert	Set server certificate and key (must supply both)	-c, --cert /path/to/cert	Path to certificate file
		-k, --key /path/to/key	Path to key file
get	Show runtime value of a single OpCenter configuration parameter	config_parameter	Configuration parameter to display, for example, sessionTimeout
ldap	Set values for LDAP configuration parameters	--addr ldap_server_address	IPv4 address of LDAP server
		--adminGroup ldap_admin_group	LDAP admin group
		--anonymous=true | false	LDAP anonymous bind (default false)
		--base base_DN	LDAP base Distinguished Name (DN)
		--bindDN bind_DN	LDAP bind DN
		--bindPW bind_pw	LDAP bind password
		--cert /path/to/cert	Path to LDAP certificate file
		--conf /path/to/conf_file	Path to LDAP configuration file
		--connTimeout duration	Duration (format: HhMmSs) until LDAP connection times out (default 10s)
		--enable=true | false	Enable LDAP (default true)
		--id ldap_id	Pre-configured LDAP configuration ID
		--key /path/to/key	Path to LDAP key file
		--network tcp | udp	LDAP connection protocol (default "tcp")
		-W, --passwordStdin	Flag to prompt user for bind password
		--peopleOU people_OU	LDAP people OU (default "People")
		--reset	Reset all LDAP parameters to empty (null)
		--tls=true | false	Use tls for LDAP connection (default: true)
list, ls	List configuration parameters for OpCenter (shows if a parameter can be changed and if a change requires a system restart)	-s, --scope filter	Condition(s) to filter the list of configuration parameters (default: list all parameters)
mset	Set runtime values of multiple configuration parameters	config_parm1=parm_value1 config_parm2=parm_value2...	Array of configuration parameters and the values to set them to. If the word "default" is used for `parm_value`, the configuration parameter is reset to its default value.
set	Set runtime value of a single configuration parameter	config_parm parm_value	Configuration parameter and the value to set it to. If the word "default" is used for `parm_value`, the configuration parameter is reset to its default value.
		--file value.txt	Text file containing value for `config_parameter` (see example)

Examples

$ float config get log.level
info

$ float config set sessionTimeout 48h
sessionTimeout is set to 48h0m0s
$ float config set sessionTimeout default
sessionTimeout is set to 1h0m0s

$ float config ls --scope "session"
 +----------------+----------+----------+--------------+
 |      KEY       |  VALUE   | EDITABLE | NEED RESTART |
 +----------------+----------+----------+--------------+
 | sessionTTL     | 168h0m0s | Y        |              |
 | sessionTimeout | 1h0m0s   | Y        |              |
 +----------------+----------+----------+--------------+

$ echo -n 48h > session.txt
$ float config set sessionTimeout --file session.txt
The value of sessionTimeout is set to '48h0m0s'.

float df

Use the float df command to display the file systems mounted by an executing job (includes used and available disk space). The linux command df must be installed in the container. This command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Display mounted file systems and associated disk space.	--args df_args	Arguments to pass to df command
		-j, --jobId job_id	Job to query

Example

$ float df --args "-h" -j WIY92p0jWyaCMP0CNQYjC
Filesystem                                                       Size  Used Avail Use% Mounted on
overlay                                                          6.0G  1.1G  5.0G  17% /
/dev/nvme3n1                                                      10G  105M  9.9G   2% /data
tmpfs                                                             64M     0   64M   0% /dev
172.31.81.17:/mnt/memverge/slurm/work/nzG6oM5DoLCysXAkoZFCA/app   50G   20G   31G  39% /mmce
/dev/nvme2n1                                                     6.0G  1.1G  5.0G  17% /etc/hosts
/dev/nvme0n1p2                                                    40G  7.5G   33G  19% /opt/aws
shm                                                               63M     0   63M   0% /dev/shm
devtmpfs                                                         1.7G     0  1.7G   0% /proc/key

float gateway

Use the float gateway command to create and manage a reverse proxy server. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
connect	Connect a running job to the gateway (reverse proxy)	-g, --gateway gw_id	ID of gateway to connect job to (format is g- followed by a fixed-length character string)
		-j,--job job_id	ID of job to connect to gateway
		--targetPort port_number	Port used to connect to job, for example, 8787 for RStudio. Include --targetPortmultiple times to connect multiple ports from a single server to the gateway.
create	Create a gateway	--bandwidth bw	Minimum gateway bandwidth in Mbps (default: 25). Only use with AliCloud.
		-c,--cpu min_cpu	Minimum number of virtual CPUs for gateway (default: 0)
		-t,--instType instance_type	VM instance type for gateway, for example, c5.xlarge for AWS. Do not combine with --cpu or --mem options.
		-m,--mem min_mem	Minimum memory capacity in GB for gateway (default: 0)
		-n,--name gw_name	Name to associate with gateway
		--noPublicIP	Create gateway with private IP address only. Ensure that gateway is reachable from the hosts that need to connect.
		--portRange min:max	Range of client-side ports opened on gateway (allowed range is between 1000 and 65535).
		--securityGroup sec_group	AWS security group (or tag in GCP) applied to gateway. Include option multiple times to apply multiple security groups or use comma-separated list.
		--subnet subnet_id	Subnet (aws) or vswitch (ali) in which to create gateway. Include option multiple times to apply multiple subnets or use a comma-separated list. Default: system chooses for you.
		-z, --zone availability_zone	Availability zone in which to create gateway
destroy	Destroy a gateway	-g, --gateway gw_id	ID of gateway to destroy (format is g- followed by a fixed-length character string)
		-f, --force	Automatically answer "yes" at confirmation prompt
disconnect	Disconnect a job from a gateway	-g, --gateway gw_id	ID of gateway to query (format is g- followed by a fixed-length character string)
		-j, --job job_id	ID of job to disconnect from gateway
		--port port_num	Server-side port to disconnect from gateway. If gateway connects to multiple ports on server, include --port for each port or use comma-separated list
info	Display information about a gateway (including IP address and connected jobs)	-g, --gateway gw_id	ID of gateway to query (format is g- followed by a fixed-length character string)
list	List all running gateways (optionally include stopped gateways)	-A, --showAll	Option to list stopped as well as running gateways (overrides filters)
		-f, --filter filter	Filter(s) to apply to list of gateways. Use option multiple times to apply multiple filters combined with "and" operator. See details in Working with OpCenter/Filters/Gateway Filters.
		-o, --orderBy attribute	Attribute used to order gateway listing (prepend with "-" to reverse order). Supported values are (default: lastUpdate): id name status cpu memory ip portStart portEnd start usedPorts cost lastUpdate
modify	Modify attributes of a gateway	--addSecurityGroup sec_group	AWS security group (or tag in GCP) added to gateway. Include option multiple times for multiple security groups or use comma-separated list.
		-g, --gateway gw_id	ID of gateway to modify
		--rmSecurityGroup sec_group	AWS security group (or tag in GCP) removed from gateway. Include option multiple times for multiple security groups or use comma-separated list

Example

$ float gateway create -n NewGateway --securityGroup sg-0fbb6a83983183364 --portRange 10000:10500
id: g-wsfthf3z8zeb0cyoc6dsb
name: NewGateway
status: Creating
configuration: ""
IPAddress: ""
portRange: 10000-10500
startTime: "2024-01-01T16:33:20Z"
usedPorts: 0
cost: ""
clientJobs: {}

float group

Use the float group command to manage OpCenter user groups. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add, create	Create a new group	new_group	Name of new group
		--admin user1,user2,...	User(s) given admin role in new group
		--gid GID	Group ID for this group (default is next available gid)
		--user user1,user2,...	User(s) added to new group
add --ldap	Associate an LDAP group with an LDAP directory	ldap_group_name	Name of group to associate with LDAP directory
delete, remove, rm	Delete a group	group_name | group_id	Name or id of group to delete
info, show	Display information about group including members	group_name | group_id	Name or id of group to query
list, ls	List groups that current user belongs to
update	Update attributes of a group	group_name | group_id	Name or id of group to update
		--add user1, user2,...	User(s) added to group
		--admin	Flag to indicate that any username listed after --add is given admin role in group and any username listed after --remove has admin role in group removed
		--gid GID	New group ID for this group
		--name group_name	New name for group
		--remove user1, user2,...	User(s) removed from group

Examples

$ float group add team --user tester
name: team
gid: 3
admins: ""
users: tester

float group update crops --admin --remove barley --add wheat
name: crops
gid: 2007
admins: wheat
users: wheat,barley
type: builtin

float hosts (sinfo)

Use the float hosts (or sinfo) command to show details and current status of current (or all) worker nodes. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Show current (or all) worker node details including status	-A, --all	Clear all filters to show all jobs
		-f, --filter filter	Filter(s) to apply to jobs whose associated hosts are displayed. Use option multiple times to apply multiple filters combined by "and" operator. See details in Working with OpCenter/Filters/Host Filters. The default filter is: -f "running=true or update<=1h" Example: -f status=executing -f timeRange=2010-10-22~
		-o, --orderBy attribute	Attribute used to order host listing (prepend with "-" to reverse order). Supported values are (default: start): entity status instanceType cpu memory payType publicIP start end
		-w, --windowSize num_hosts	Number of hosts displayed before header row is repeated (default: 512)

float image

Use the float image command to manage container images on OpCenter. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add	Add image pull information to OpCenter	image_name image_uri	Name to associate with image in the AppLibrary and URI to pull image from repository. Do not use image_uri if --link specified.
		--cachePath /path/to/cache	Location to cache image from a running job. Only valid if combined with -j flag identifying a running job. Choose a name different from image name in AppLibrary. Path choices are value of image.cachePath in OpCenter configuration snapshot://
		--import-all	Import all tags for this image from repository
		-j, --job job_id	ID of running job to identify container image
		--link file_link	Link to access AWS S3 or AliCloud OSS file, for example, s3://bucket_name/file_path Do not use if image_uri specified.
		--token repo_access_token	Token to pull image from private repository
		--user repo_access_user	Username to pull image from private repository
cache	Add (delete) image to (from) cache configured for OpCenter (NFS-mounted directory or S3 bucket)	image_name	Container image to add to or delete from cache. If cachePath flag not included, image is cached at location specified by image.cachePath in the OpCenter configuration.
		--cachePath /path/to/cache	Path to image cache. Choices are s3://bucketname nfs://server_ip/dir snapshot://
		-d, --delete	If option included, delete image from cache
		-f, --force	Overwrite existing cached version of image
		--imageSize image_size	Size of image to cache
		--tag tag_name	Tag to select specific container image (default: "latest")
delete, rm, remove	Delete container image(s) and one or more tags	image1 image2... or image1:tag1 image2:tag2...	Container image(s) to delete. If no tags are specified, then all tags are removed.
		-f, --force	Automatically answer "yes" at confirmation prompt
		--tag tag_name	Tag associated with image1 image2... to remove. If this is the only tag associated with images, images are removed.
list, ls	Show images available on OpCenter	-f, --filter filter_definition	Filter applied to all images, for example, -f "name=ubuntu". See Working with OpCenter/Filters for details.
tags, tag	Display tags associated with image and image status	image_name	Container image name
update	Update information associated with image	image_name	Container image name
		--addtag image_tag1,image_tag2...	Additional tags to associate with image
		--name image_name	New name to identify image
		--token repo_access_token	New token required to access private repository
		--user repo_access_user	New username associated with token required to access private repository
upload	Load image from local server	image_name	Name to identify image in App Library
		-i, --image local_name	Name of image in local repository (cannot use if --path included)
		--path /path/to/image	Path to where image is located ((cannot use if --image included))

Examples

$ float image list
+------------+-------------------------------+--------+-------------+
|    NAME    |              URI              | TAGS   | ACCESS USER |
+------------+-------------------------------+--------+-------------+
| python     | docker.io/bitnami/python      | latest |             |
| r-base     | docker.io/rocker/r-base       | latest |             |
.....(edited)
+------------+-------------------------------+--------+-------------+

$ float image delete r-base
Deleting r-base

$ float image add r-base docker.io/rocker/r-base
name: r-base
uri: docker.io/rocker/r-base
owner: admin
tags:
    latest:
        status: Available
        locked: false
        lastUpdated: 2023-06-15T16:12:41.739947963Z
        size: Unknown

$ float image add ubuntu docker.io/ubuntu --cachePath snapshot:// -j dit41mw0127mh5b4ozzmo
Error: Resource already exists, image already exists (code: 1018)
$ float image add ubu3 docker.io/ubuntu --cachePath snapshot:// -j dit41mw0127mh5b4ozzmo
name: ubu3
uri: docker.io/library/ubuntu
owner: admin
tags:
    latest:
        status: Ready
        uri: snapshot://snap-0c2f373876759e470
        locked: false
        lastUpdated: 2025-04-23T15:51:38.874421771Z
        lastPushed: 2025-04-23T15:51:38.874421871Z
        size: 6.00 GB

$ float image cache blast
Request to cache image blast (tag: latest) has been submitted
$ float image cache blast --delete
Request to clean image blast (tag: latest) has been submitted

$ float image cache postgres --cachePath snapshot://
Error: Resource not found (code: 1002)
$ float image add postgres docker.io/postgres
name: postgres
uri: docker.io/library/postgres
owner: admin
tags:
    latest:
        status: Available
        locked: false
        lastUpdated: 2025-04-23T16:04:12.482960127Z
        size: Unknown
$ float image cache postgres --cachePath snapshot://
Request to cache image postgres (tag: latest) has been submitted

$ float image tags blast
+--------------------------+--------+-----------+----------------------+
|           URI            |  TAG   |  STATUS   |     LAST UPDATED     |
+--------------------------+--------+-----------+----------------------+
| docker.io/memverge/blast | 2.14.0 | Available | 2023-06-26T05:15:02Z |
|                          | latest | Available | 2024-08-13T15:51:33Z |

$ float image upload test2 --path /tmp/hello-world-latest.tar
Start uploading image test2 (localhost/hello-world:latest) from /tmp/hello-world-latest.tar
Progress: |>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>| 100.00% Complete (ETA. 0s)
Uploaded image /tmp/hello-world-latest.tar, time spent: 0s

name: test2
uri: localhost/hello-world
owner: admin
tags:
    latest:
        status: Ready
        uri: file:///mnt/memverge/images/test2-latest.tar
        locked: false
        lastUpdated: 2023-04-17T19:26:38.584557866Z
        lastPushed: 2023-04-17T19:26:38.584557866Z
        size: Unknown

$ float image upload testinggzupload --path ./testinggzupload.tar.gz
Start uploading image testinggzupload (docker.io/library/testinggzupload:latest) from ./testinggzupload.tar.gz
Progress: |>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>| 100.00% Complete (ETA. 0s)
Uploaded image ./testinggzupload.tar.gz, time spent: 31m20s

name: testinggzupload
uri: docker.io/library/testinggzupload
owner: admin
tags:
    latest:
        status: Ready
        uri: s3://opcenter-bucket-594424d0-c760-11ee-8abd-128ad93e7ec9/images/ntltcroq2g.tar
        locked: false
        lastUpdated: 2024-04-11T15:39:29.686393328Z
        lastPushed: 2024-04-11T15:39:29.686393428Z
        size: 1.33 GB

Note

The user that queries the local repository for the IMAGE ID must be the same user that executes the float image upload command In this example, the user is root.

$ sudo podman images
REPOSITORY                  TAG         IMAGE ID      CREATED        SIZE
quay.io/podman/hello        latest      39ae24b9cabf  3 days ago     1.7 MB

$ sudo /opt/memverge/bin/float image upload testimage --image 39ae24b9cabf
Found image quay.io/podman/hello:latest, using it as repository and tag
Using /bin/podman to save image quay.io/podman/hello:latest to testimage.tar
Start uploading image testimage (quay.io/podman/hello:latest) from testimage.tar
...[edited]

float ldap

Use the float ldap command to manage LDAP configurations on the OpCenter. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add	Add a new LDAP configuration	--addr ldap_addr	LDAP server IP address
		--adminGroup group_name	LDAP admin group
		--anonymous	Use anonymous bind
		--base base_DN	Base DN used in ldap search
		--bindDN bind_DN	LDAP bind DN for authentication
		--bindPW bind_PW	LDAP bind password for authentication
		--cert /path/to/cert_file	Path to LDAP certificate file
		--connTimeout duration	LDAP connection timeout (default: 10s)
		--groupOU group_OU	LDAP group OU
		--key /path/to/key_file	Path to key file
		-n, --name ldap_conf	Name of the LDAP configuration
		--network tcp | udp	Transport protocol to connect to LDAP server
		--passwordStdin	Prompt for bindPassword
		--peopleOU people_OU	LDAP people OU
		--precheck	Run pre-check before adding LDAP config
		--useTLS	Use TLS for transport layer security
delete	Delete an LDAP configuration	-i, --id ldap_config_id	ID of LDAP configuration to delete
get	Get an LDAP configuration	-i, --id ldap_config_id	ID of LDAP configuration to get
list (ls)	List all LDAP configurations
update	Update an LDAP configuration	--addr ldap_addr	LDAP server IP address
		--adminGroup group_name	LDAP admin group
		--anonymous	Use anonymous bind
		--base base_DN	Base DN used in ldap search
		--bindDN bind_DN	LDAP bind DN for authentication
		--bindPW bind_PW	LDAP bind password for authentication
		--cert /path/to/cert_file	Path to LDAP certificate file
		--connTimeout duration	LDAP connection timeout (default: 10s)
		--groupOU group_OU	LDAP group OU
		-i, --id ldap_config_id	ID of LDAP configuration to update
		--key /path/to/key_file	Path to key file
		-n, --name ldap_conf	Name of the LDAP configuration
		--network tcp | udp	Transport protocol to connect to LDAP server
		--passwordStdin	Prompt for bindPassword
		--peopleOU people_OU	LDAP people OU
		--precheck	Run pre-check before adding LDAP config
		--useTLS	Use TLS for transport layer security

float license

Use the float license command to manage OpCenter licenses. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
acquire	Acquire license from MMCloud Portal	-A, --account acct_name	Username (email address) to access the MMCloud Portal
		-P, --password passwd	Password to access the MMCloud Portal
info	Display license information and status

float list (squeue, ls)

Use the float list (or squeue or ls) command to show a filtered list of queued jobs. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Show filtered list of queued jobs (default: list jobs from oldest to newest)	-A, --all	Clear all filters to show all jobs
		-f, --filter filter	Filter(s) to apply to list of queued jobs. Use option multiple times to apply multiple filters combined with "and" operator. See details in Working with OpCenter/Filters/Job Filters. Example: -f status=executing -f timeRange=2023-10-22~ -f tags=kind:training,project:finance-\*
		-o, --orderBy attribute	Attribute used to order job listing (prepend with "-" to reverse order). Supported values are (default: start): id name user status duration start cost cpu memory
		-w, --windowSize num_jobs	Number of jobs displayed before header row is repeated (default 10000)

Examples

$ float squeue -f "imageID=*blast*"
+-------+------------------+...+-------+-----------+----------+----------------------+------------+
|   ID  |       NAME       |...| USER  |  STATUS   | DURATION |     SUBMIT TIME      |    COST    |
+-------+------------------+...+-------+-----------+----------+----------------------+------------+
| m8t...| blast-c5.9xlarge |   | admin | Completed | 8h44m18s | 2024-02-14T04:07:38Z | 5.4851 USD |
| w70...| blast-c5.9xlarge |   | bean  | Completed | 9h3m16s  | 2024-02-14T04:22:45Z | 5.3502 USD |
+-------+------------------+...+-------+-----------+----------+----------------------+------------+
...[edited for clarity]

$ float squeue -f "imageID=*blast*" -f user=bean
+-------+------------------+...+-------+-----------+----------+----------------------+------------+
|   ID  |       NAME       |...| USER  |  STATUS   | DURATION |     SUBMIT TIME      |    COST    |
+-------+------------------+...+-------+-----------+----------+----------------------+------------+
| w70...| blast-c5.9xlarge |   | bean  | Completed | 9h3m16s  | 2024-02-14T04:22:45Z | 5.3502 USD |
+-------+------------------+...+-------+-----------+----------+----------------------+------------+
...[edited for clarity]

$ float squeue -o name
+--------+----------------------+------------------+-------+-----------+--------------+----------+------------+
| ID     |         NAME         |    WORKING HOST  | USER  |  STATUS   | SUBMIT TIME  | DURATION |    COST    |
+-------------------------------+------------------+-------+-----------+--------------+----------+------------+
| SZd... | apple                | 54.163.154.116...| admin | Executing | ...19:11:47Z | 1h35m1s  | 0.1143 USD |
| btC... | banana               | 34.234.64.157... | admin | Executing | ...19:12:19Z | 1h34m30s | 0.0926 USD |
| hnP... | camel                |                  | admin | Completed | ...20:43:01Z | 3m28s    | 0.0031 USD |
| 5d1... | cherry               | 54.89.203.124... | admin | Executing | ...19:12:27Z | 1h34m22s | 0.0924 USD |
| grd... | python-t3a.xlarge    | 34.229.17.82...  | admin | Completed | ...20:43:48Z | 2m51s    | 0.0026 USD |
| nVD... | tidyverse-t3a.xlarge | 3.89.224.246...  | admin | Executing | ...19:14:39Z | 1h32m10s | 0.0903 USD |
+--------+----------------------+------------------+-------+-----------+--------------+----------+------------+
...[edited for clarity]

$ float squeue -f "imageID=*python:latest and status=Executing"
+-----------------------+-------------------------+------------------------------+--------+-----------+...
|          ID           |          NAME           |         WORKING HOST         |  USER  |  STATUS   |...
+-----------------------+-------------------------+------------------------------+--------+-----------+...
| w2763vpgwuy9cksihxbcd | python-d72rur-t3.medium | 44.220.82.26 (2Core4GB/Spot) | banana | Executing |...
| q6r4ome7wyywk33b5jo9g | testpy                  | 3.239.38.179 (2Core4GB/Spot) | bean   | Executing |...
+-----------------------+-------------------------+------------------------------+--------+-----------+...
...[edited for clarity]

$ float squeue -f "imageID=*python* or imageID=*tidyverse*"
+-----------------------+-------------------------+------------------------------+--------+-----------+...
|          ID           |          NAME           |         WORKING HOST         |  USER  |  STATUS   |...
+-----------------------+-------------------------+------------------------------+--------+-----------+...
| w2763vpgwuy9cksihxbcd | python-d72rur-t3.medium | 44.220.82.26 (2Core4GB/Spot) | banana | Executing |...
| q6r4ome7wyywk33b5jo9g | testpy                  | 3.239.38.179 (2Core4GB/Spot) | bean   | Executing |...
| bxulaob02n1d91cmymvxk | Rtest                   | 44.212.52.61 (2Core4GB/Spot) | bean   | Executing |...
+-----------------------+-------------------------+------------------------------+--------+-----------+...
...[edited for clarity]

float log

Use the float log command to view and manage log files. Use subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
cat	Write log file contents to standard output	log_file	Log file whose contents are displayed
		-c, --cluster cluster_id	Cluster whose logs are displayed
		-i, --hid host_id	Host whose logs are displayed (default: OpCenter server)
		-j, --job job_id	Job whose logs are displayed (default: OpCenter)
download	Download a zip file of selected logs associated with job	-i, --include logs1,logs2,...	Logs included in zip file (default: "all"). Options are: "all": all logs "opCenter": OpCenter logs "app": application logs "host": host logs "job": job logs "ocStack": OpCenter stack trace "agentStack": agent stack trace "jobObject": job object
		-j, --job job_id	Job whose log files are included
		--path /path/to/dir	Path to save zip file (default "./")
list, ls	List all logs associated with target	--All	Display all log files
		-c, --cluster cluster_id	Cluster whose log files are listed
		-i, --hid host_id	Host whose log files are listed (default: OpCenter server)
		-j, --job job_id	Job whose log files are listed (default: OpCenter)
		-H, --readable	Display log size in human-readable format
rm	Remove all logs associated with target	-i, --hid host_id	Host whose log files are removed (default: OpCenter server)
		-j, --job job_id	Job whose log files are removed (default: OpCenter)
tail	Write last n lines of log file to standard output	log_file	Log file to display
		-f, --follow	Display new lines as they are appended to log file
		-c, --cluster cluster_id	Cluster whose log file lines are displayed
		-i, --hid host_id	Host whose log file lines are displayed (default: OpCenter server)
		-j, --job job_id	Job whose log file lines are displayed (default: OpCenter)
		-n, --num n	Number of lines to display (default: 100)

Examples

$ float log tail --follow output -j XGiUDRto7kwofWBNPkiW5
Ready to prepare source data
Ready to download pbmc_1k_v3_fastqs from s3
Ready to download refdata-gex-GRCh38-2020-A from s3
Ready to run test
...[output edited]

$ float log ls -H
+---------------------+---------------+----------------------+
|      LOG NAME       | READABLE SIZE |   LAST UPDATE TIME   |
+---------------------+---------------+----------------------+
| opcenter.access_log | 7.40 MB       | 2024-02-14T19:02:02Z |
| opcenter.log        | 712.73 KB     | 2024-02-14T18:55:28Z |
| upgrade.log         | 1.92 KB       | 2024-02-11T03:54:27Z |
| messages            | 220.56 KB     | 2024-02-14T18:48:48Z |
+---------------------+---------------+----------------------+

float login

Use the float login command, with valid username and password, to log in to OpCenter. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Log in to OpCenter	--info	Display login status

Examples

$ float -u admin -p secret123 -a 192.168.0.1 login

$ float login --info
address: 192.168.0.1
username: admin
role: admin
type: builtin

$ float login
Username: admin
Password: 
Login Succeeded!

float logout

Use the float logout command to log the current user out of the OpCenter and invalidate the authorization token. The command has no subcommands. Use the command with the -h flag to list the options.

Example

$ float logout 
Logout Succeeded!
$ float login --info
Error: cannot find any login session (code: 2004)

float mi

Use the float mi command to manage virtual machine images. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
list	Display information on available machine images	--source filter	Filter to select which source of machine images to show (default: "all")
update	Upload machine image file	-f, --file /path/to/file	Path to machine image file

Example

$ float mi list --source "official"
aws:
    us-east-1:
        - id: ami-01a1370c28ed96480
          oid: mi-xw3smh9ojzacdpko7uyf0
          name: rocky_94-nvidia_550
          description: 'os: Rocky9.4, gpu: NVIDIA 550'
          tags:
            gpu: NVIDIA
            os: rocky-9.4
          source: official
          cloudMachineImage:
            arch: x86_64
            id: ami-01a1370c28ed96480
            name: Rocky-9.4-x86_64-MMC-v3.0.0-NVIDIA_550
            description: ""
            tags: {}
          updateTime: 2025-04-02T23:01:59.139806718Z
        - id: ami-05fa59fc8129caa16
          oid: mi-ny39fy3h6d53lec7ka3fe
          name: rocky_94
          description: 'os: Rocky9.4'
          tags:
            os: rocky-9.4
          source: official
          cloudMachineImage:
            arch: x86_64
            id: ami-05fa59fc8129caa16
            name: FLOAT-release_v3.0-20240716014405-8a0e60e3-8e27-4221-8bee-69944e051d58
            description: FLOAT-release_v3.0-20240716014405
            tags: {}
          updateTime: 2025-04-02T23:01:59.139806718Z

float migrate

Use the float migrate command to move a job from one VM instance to another VM instance of the same type or of a different type. The command has no subcommands. Use the command with the -h flag to display the options.

Subcommands	Usage	Option	Option Definition
	Migrate job to a new VM instance. If no options used, migrate to identical instance in same availability zone.	-c, --cpu min_cpu:max_cpu	Range of number of virtual CPUs to select from for new VM instance. If max_cpu omitted, value of twice min_cpu used.
		-e, --env	Environment variables that apply when the migrated job resumes. Use format env_key=env_value. Include secrets in this format: "SECRET={secret:TEST_SECRET}"
		-f, --force	Automatically answer "yes" at confirmation prompt
		--gpu-count min_gpu:max_gpu	Range of number of GPUs to select from for new VM instance
		--gpu-mem min_gpumem:max_gpumem	Range of memory size (in GB) of GPUs to select from for new VM instance
		--gpu-vendor gpu_vendor	GPU vendor name (for example, nvidia or amd)
		-t, --instType instance_type	VM instance type to migrate to (e.g., c5.xlarge in AWS). Do not combine with --cpu or --mem options.
		-j, --job job_id	Job to migrate
		-m, --mem min_mem:max_mem	Range of memory size (in GB) to select from for new VM instance. If max_mem omitted, value of twice min_mem used.
		-P, --payType spot | ondemand	Pricing tier for VM instance (Spot or On-demand)
		--rerun	Ignore snapshot and restart job from the beginning on new VM instance (job must be running)
		--sync	Block all terminal input until job has migrated
		-z, --zone availability_zone	Availability zone in which to execute job

Examples

$ float migrate -t c5.xlarge -j NDt428IsJtJZsB9WNUhGH

$ float migrate -j lL07E84pQQpYqCQ88xeIQ
entity:
  id: i-0cfbd3f1f82087dd5
  type: host
  name: 192.168.0.2
status: normal
instanceType: c5.xlarge
startTime: "2022-09-23T15:09:09Z"
downTime: ""

$ float migrate -f --sync --payType ondemand -j tqrGc4Z6g18nkphzTeaxM
tqrGc4Z6g18nkphzTeaxM is now migrating...
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

tqrGc4Z6g18nkphzTeaxM has been migrated to 44.212.92.162 (4Core16GB/OnDemand).

float modify

Use the float modify command to change a subset of attributes associated with a running job. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Modify certain attributes associated with running job	--addCustomTag tagName:tagValue	Tag name and value to associate with running job. Separate multiple tags with "," or include option multiple times to add multiple tags.
		--addSecurityGroup sec_group	Security group (or tag in GCP) added to VM instance for this job. Include option multiple times to add multiple security groups.
		--errPolicy err_policy	The policy to use if the job fails. The allowed policies are the following. reclaimAll: change job state to failed and reclaim all cloud resources restart: retain VM and local storage volumes (file systems) and automatically restart job from the beginning on same VM retainVolumes: reclaim VM and retain local storage volumes (file systems); user must cancel job or resume job on new VM using float resume -j job_id autosuspend: if OOM protection triggers migration and no instance type found, suspend job; user must cancel job or resume manually
		--force	Automatically answer "yes" at confirmation prompt
		-j, --job job_id	Job to apply changes to
		-M, --migratePolicy migrate_policy	New migrate policy to apply. See float submit for format.
		-m, --monthly	Turn monthly subscription on (or off). Only used for Alicloud and Tencent cloud.
		-P, --period	Duration of monthly subscription in months (default: 1). Only used for Alicloud and Tencent cloud.
		--rmSecurityGroup sec_group	Security group (or tag in GCP) to remove from VM instance for this job. Include option multiple times to remove multiple security groups.
		--snapshotInterval snapshot_interval	New periodic snapshot interval for this job. Use "disable" or "0" to turn off periodic snapshots.
		-V, --vmPolicy vm_policy	New VM creation policy to apply. See float submit for format.

Example

$ float modify --vmPolicy [SpotOnly=true] -j PF0bgCvlpdJkog0RCBZPg
Warning: Are you sure you want to modify PF0bgCvlpdJkog0RCBZPg?
New vmPolicy may trigger migration.(yes/No): yes
Successfully modified PF0bgCvlpdJkog0RCBZPg:  --vmPolicy [SpotOnly=true]

$ float modify -j y78iefqeo8k6ng2rpvduj --addCustomTag run-name=task1
Successfully modified y78iefqeo8k6ng2rpvduj:  --addCustomTag run-name=task1
$ float show -j y78iefqeo8k6ng2rpvduj
...
customTags:
    run-name=task1: ""
...[edited]
$ float modify -j y78iefqeo8k6ng2rpvduj --addCustomTag project=RNA-sequencing,dept=Research
Successfully modified y78iefqeo8k6ng2rpvduj:  --addCustomTag project=RNA-sequencing,dept=Research
$ float modify -j y78iefqeo8k6ng2rpvduj --addCustomTag PI=jones --addCustomTag funding=NHS
Successfully modified y78iefqeo8k6ng2rpvduj:  --addCustomTag PI=jones --addCustomTag funding=NHS
$ float show -j y78iefqeo8k6ng2rpvduj
...
customTags:
    PI=jones: ""
    funding=NHS: ""
    project=RNA-sequencing,dept=Research: ""
    run-name=task1: ""
...[edited]

float nis

Use the float nis command to manage NIS configurations. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add	Update an SMTP configuration	-n, --name nis_config	Name to associate with the nis config
		-D, --domain nis_domain	NIS domain
		--precheck	Run precheck before adding NIS configuration
		-S, --server nis_addr	IP address of NIS server
delete	Delete an NIS configuration	nis_id	ID that identifies NIS configuration to delete
get	Get NIS configuration	nis_id	ID that identifies NIS configuration to get
list, ls	List all NIS configurations
update	Update an NIS configuration	-i , --id nis_id	ID to identify NIS config
		-D, --domain nis_domain	NIS domain
		-n, --name nis_name	Name to identify NIS config
		S, --server nis_ip_addr	IP address of NIS server

float node

Use the float node command to create and manage nodes (host or cloud instance). Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
create	Create a new node (host or cloud instance)	--cphLimit limit	Maximum allowed cost per hour for this node
		-c, --cpu min_cpu:max_cpu	Range of number of virtual CPUs to select from for new node. If max_cpu omitted, value of twice min_cpu used.
		--imageID mi_id	Machine image ID
		-t, --instType instance_type	VM instance type (e.g., c5.xlarge in AWS). Do not combine with --cpu or --mem options.
		--maxCpuCores max_cpu	Maximum number of CPU cores
		--maxMemGB max_mem	Maximum memory capacity in GB
		-m, --mem min_mem:max_mem	Range of memory capacity to select from for new node. If max_mem omitted, value of twice min_mem used.
		--miTag mi_tag1,mi_tag2,...	Machine image tag(s) required for node
		--minCpuCores min_cores	Minimum number of CPU cores (default: 1)
		--minCudaCores min_cuda	Minimum number of Cuda cores
		--minGpuMemoryGB min_gmem	Minimum GPU memory capacity in GB
		--minGpuCores min_gpu	Minimum number of GPU cores
		--minMemGB min_cmem	Minimum CPU memory capacity in GB
		--name node_name	Name of node (optional)
		--onDemand	Create on-demand instance
		--onDemandCphLimit od_limit	Maximum allowed cost per hour for on-demand instance
		--optional-miTag mi_tag1,mi_tag2, ...	Optional machine image tags for this node
		--region region	Region to create node in
		--rootVolSizeGB root_vol	Minimum size of root volume in GB (default: 32)
		--spotCphLimit sp_limit	Maximum allowed cost per hour for spot instance
		--spotFirst	Create a spot instance if possible (default: true)
		--spotOnly	Create a spot instance only
		--zone	Availability zone in which to create node
delete (del, destroy, rm)	Delete node by name	node_name	Name of node to delete
list	List the managed worker nodes	-f, --filter filter	Filter(s) to apply to list of queued jobs. Format is [attribute][operator][value]. Use option multiple times to apply multiple filters combined with "and" operator. See details in Working with OpCenter/Filters/Job Filters. Supported attributes are: cores currency interrupt memoryGB name onDemand spot vendor zone
		-l, --limit num_results	Limit to the number of results displayed (0 means no limit)
		-o, --orderBy attribute	Attribute used to order node listing (prepend with "-" to reverse order). Nest attributes by separating with a comma. Supported values are: cores currency interrupt memoryGB name onDemand spot vendor zoneid
		-s, --pos num_pos	Start position of the display of the results
info	Show detailed information about a node	node_id	Node ID

float project

Use the float project command to create and manage projects. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add (create, new)	Create new project with specified parameters	--dataVolume stringArray	Data volume(s) to run job. See float submit for syntax.
		-m, --instType instance_type	VM instance type to run job (e.g., c5.xlarge in AWS). Do not combine with --cpu or --mem options.
		-I, --maxIdleTime duration	Maximum idle time for nodes in the project
		-M, --maxNodes integer	Maximum number of nodes in project
		-mi mi_id	ID of machine image to use in project
		-n, --name string	Name of project
		--precheck	Run check before adding project
		-s, --subnet net1,net2,...	Subnet(s) in AWS or vSwitch in Alicloud to create nodes in.
		-t, --tag key1=value1,key2=value2,...	Tag(s) to associate with project
		-V, --vmPolicy stringArray	VM creation policy. See float submit for syntax.
delete (del, remove, rm)	Delete project by name or ID	i, --id project_id	ID of project to delete
		n, --name project_name	Name of project to delete
get	Get properties of project	i, --id project_id	ID of project to query
		n, --name project_name	Name of project to query
jobs (jl, joblist, jobls, jls)	Display jobs in project	i, --id project_id	ID of project to query
		n, --name project_name	Name of project to query
list (ls)	List all projects
modify (update)	Modify project parameters	--dataVolume stringArray	Data volume(s) to run job. See float submit for syntax.
		i, --id project_id	ID of project to modify
		-m, --instType instance_type	VM instance type to run job (e.g., c5.xlarge in AWS). Do not combine with --cpu or --mem options.
		-I, --maxIdleTime duration	Maximum idle time for nodes in the project
		-M, --maxNodes integer	Maximum number of nodes in project
		-mi mi_id	ID of machine image to use in project
		-n, --name string	Name of project
		-o, --overwrite	Overwrite existing parameter values
		-s, --status	New status of project
		-N, --subnet net1,net2,...	Subnet(s) in AWS or vSwitch in Alicloud to create nodes in.
		-t, --tag key1=value1,key2=value2,...	Tag(s) to associate with project

float promote (boost)

Use the float promote or float boostcommand (as an admin user) to move a job in the "Submitted" state to the front of the scheduling queue. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Promote a job in the "Submitted" state to the front of the scheduling queue. Must be admin user.	-j, --job job_id	Job to promote

float ps

Use the float ps command to show the complete process tree of a running job (the linux command ps must be installed in the container). The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Show the complete process tree of job	--args podman_args	Arguments passed to podman
		-j, --jobId job_id	Job to query

float ptop (qtop)

Use the float ptop or float qtop commands to show the complete process tree of a running job (the linux command ps must be installed in the container). The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Show top of pending jobs queue visible to the current user (index indicates the order of the job in the pending queue)	-l, --limit integer	Maximum number of pending jobs to return (default: 32)

float queue

Use the float queue command to manage queues in an HPC cluster. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add, create, new	Add a new queue to an existing cluster	-c, --cluster cluster_name	Cluster name
		-D, --dataVolume stringArray	Data volume for jobs in the queue. See float submit for syntax.
		-H, --hyperThread int	Hyperthread setting for queue. 0 means default, 1 means disable, 1 means enable
		-m, --instType i_type1,i_type2,...	VM instance type(s) (e.g., c5.xlarge in AWS)
		-I, --maxIdleTime duration	Maximum idle time for nodes in the project
		-M, --maxNodes max_nodes	Maximum number of nodes for the queue
		--imageID mi_id	Machine image ID for jobs in the queue
		-n, --name queue_name	Name of queue
		--precheck	Run check before adding queue
		-Q, --qos conf1,conf2,...	QoS configuration(s) for the jobs in the queue
		--reconfigure	Reconfigure queue after modifying
		--storage storage_def	Storage for the jobs in the queue. See float storage for syntax.
		-s, --subnet net1,net2,...	Subnet(s) in AWS or vSwitch in Alicloud for jobs in the queue
		--useClusterSg	Use cluster security groups directly
		--useClusterStorage	Use cluster storage directly
		--useClusterSubnet	Use cluster subnets directly
		-V, --vmPolicy policy_def	VM creation policy for jobs in queue. See float submit for syntax.
delete (del, rm, remove)	Delete a queue from an HPC cluster. Identify queue by name or ID.	-c, --clusterID cluster_id	ID of HPC cluster
		-i, --id queue_id	ID of queue to delete
		-n, --name queue_name	Name of queue to delete
get	Get properties of a queue. Identify queue by name or ID.	-c, --clusterID cluster_id	ID of HPC cluster
		-i, --id queue_id	ID of queue to get
		-n, --name queue_name	Name of queue to get
jobs (jl, joblist, jobls, jls)	List all jobs in a queue	-c, --clusterID cluster_id	ID of HPC cluster
		-f, --filter filter_def	Filter to apply to job listing
		-i, --id queue_id	ID of queue whose jobs are listed
		-n, --name queue_name	Name of queue whose jobs are listed
		-s, --orderBy order_def	Criteria to order job listing
		-v, --verbose	Show extended information of jobs in queue
ls	List available queues	-f, --filter filter_def	Filter to apply to queue listing
		-s, --orderBy order_def	Criteria to order job listing
		-v, --verbose	Show extended information of queues
modify (update)	Modify queue in an existing cluster	-c, --cluster cluster_name	Cluster name
		-D, --dataVolume stringArray	Data volume for jobs in the queue. See float submit for syntax.
		-H, --hyperThread int	Hyperthread setting for queue. 0 means default, 1 means disable, 1 means enable
		-i, --id queue_id	ID of queue whose jobs are listed
		-m, --instType i_type1,i_type2,...	VM instance type(s) (e.g., c5.xlarge in AWS)
		-I, --maxIdleTime duration	Maximum idle time for nodes in the project
		-M, --maxNodes max_nodes	Maximum number of nodes for the queue
		--imageID mi_id	Machine image ID for jobs in the queue
		-n, --name string	Name of queue
		--precheck	Run check before adding queue
		-Q, --qos conf1,conf2,...	QoS configuration(s) for the jobs in the queue
		--reconfigure	Reconfigure queue after modifying
		--storage storage_def	Storage for the jobs in the queue. See float storage for syntax.
		-s, --subnet net1,net2,...	Subnet(s) in AWS or vSwitch in Alicloud for jobs in the queue
		--useClusterSg	Use cluster security groups directly
		--useClusterStorage	Use cluster storage directly
		--useClusterSubnet	Use cluster subnets directly
		-V, --vmPolicy policy_def	VM creation policy for jobs in queue. See float submit for syntax.

float queues

Use the float queues command to get information about queues in an HPC cluster. The command has no subcommands. Use the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	List available queues	-f, --filter filter_def	Filter to apply to queue listing
		-s, --orderBy order_def	Criteria to order job listing
		-v, --verbose	Show extended information of queues

float quota-policy

Use the float quota-policy command to define and manage quota (SurfZone) policies. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add, create	Create quota policy	policy_name	Name to associate with the quota policy
		--limit budget_limit	Maximum amount (in $) allowed to spend in one month
		--action cancel | suspend	Action taken when quota limit reached (default: cancel)
		--auto-resume=true | false	Action taken on suspended job when quota replenished (default: true)
		--threshold threshold	Percentage of quota (budget) consumed to trigger alert (default: 80)
info, show	Display quota policy details	policy_id	ID that identifies quota policy
list	List available quota policies
update, modify	Update parameters in existing quota policy	policy_id	ID that identifies quota policy
		--name policy_name	New name to associate with the quota policy
		--limit budget_limit	Updated maximum amount (in $) allowed to spend in one month
		--action cancel | suspend	Updated action taken when quota limit reached
		--autoresume true | false	Updated action taken on suspended job when quota replenished
		--threshold threshold	Updated percentage of quota (budget) consumed to trigger alert
delete, rm, remove	Delete a quota policy	policy_id	ID to identify quota policy

Examples

$ float quota-policy list
id: u2w02ntshv61mxc7muaz0
  name: fruit
  metric: Cost
  overageAction: cancel
  autoResume: false
  threshold: 80%
  limit: 37
- id: 2f4qaesir4wxz71mazstq
  name: legume
  metric: Cost
  overageAction: suspend
  autoResume: true
  threshold: 80%
  limit: 35

$ float quota-policy update 2f4qaesir4wxz71mazstq --limit 50
id: 2f4qaesir4wxz71mazstq
name: legume
metric: Cost
overageAction: suspend
autoResume: true
threshold: 80%
limit: 50

float release

Use the float release command to manage the OpCenter software. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
info	Display information regarding new features and bug fixes in OpCenter release	-r, --release version	The release to display information about
list, ls	List available OpCenter releases
migrate	Migrate existing database to new database or new location	-d, --dbPath /path/to/db	Destination path of new DB or new DB location
sync	Sync CLI version with OpCenter version
upgrade	Upgrade OpCenter software	--force	Automatically answer "yes" at confirmation prompt
		-r, --release version	The release to upgrade to (default: "latest"). Only the admin user can upgrade software. Cannot upgrade using web CLI console.
		--sync	Wait for upgrade to complete and then sync the CLI to the new release

Examples

$ float release ls
+----------+--------------------------------------+----------------------+-----------+
| VERSION  |               RELEASE                |     RELEASE TIME     |   SIZE    |
+----------+--------------------------------------+----------------------+-----------+
| * v2.5.0 | FLOAT_v2.5.0-171088e-HalfMoonBay.bin | 2024-02-09T03:31:15Z | 220.65 MB |
| v2.4.1   | FLOAT_v2.4.1-0803674-Goa.bin         | 2024-01-09T18:29:33Z | 219.32 MB |
+----------+--------------------------------------+----------------------+-----------+

$ float release sync
downloading ...
Progress: |>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>| 100.00% Complete (ETA. 0s)
The float binary is synced up with opcenter 

Note

If you place the float binary in a directory that is only writable by the root user, use sudo float release sync

float report

Use the float report command to download reports from OpCenter. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
download	Download OpCenter usage history report	--path /path/to/dir	Path to save report on local computer
get	Generate and display usage report with filter(s) applied	report_name	Name of report, e.g., usage_report_by_job
		-A, --all	Compile report from all usage data retained by this OpCenter. If -A not used, default filter applied (different for each report).
		-d, --date date_string	Date used to filter reports (default: current report)
		--filter filter	Filter(s) to apply to usage report. Use option multiple times to apply multiple filters. Working with OpCenter/Filters/Job Filters for details. Example: timeRange=2010-10-22~
		-l, --limit num_records	Maximum number of records to display (default: 0 which means unlimited)
		-o, --orderBy field	Field to order by. Use 'cost' or 'jobs' (default). The column entitled 'JOB COUNT' is used to order entries when 'jobs' specified.
		-r, --refresh	Force the refresh of report
ls, list	List all available usage reports

Examples:

$ float report get usage_report_by_user -A -f status=Cancelled
+-----------+-----+-----------+------------+--------------+----------------+--------------------+...
| USER NAME | FEE | JOB COUNT | WALL TIME  | COMPUTE TIME | SPOT INSTANCES | ONDEMAND INSTANCES |...
+-----------+-----+-----------+------------+--------------+----------------+--------------------+...
| admin     |   0 |        69 | 617h57m42s | 615h42m7s    |             47 |                 11 |...
| apple     |   0 |         1 | 6h33m59s   | 6h53m31s     |              8 |                  0 |...
  ...
(edited)

$ float report download --path ./temp.gzip
Downloaded to ./temp.gzip
`file temp.gzip`
temp.gzip: gzip compressed data, original size modulo 2^32 14336

float rerun (requeue, resubmit)

Use the float rerun (or requeue or resubmit) command to re-submit a completed job, a canceled job, or a job that failed to complete. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Re-submit completed or failed job	-j, --job job_id	Job to re-submit

float restart (reboot)

Use the float restart (or reboot) command to restart OpCenter (terminates all login sessions). The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Restart OpCenter	-f, --force	Automatically answer "yes" at confirmation prompt

Example

$ float list
+-----+---------------------+...+-----------+----------------------+----------+------------+
| ID  |        NAME         |...|  STATUS   |     SUBMIT TIME      | DURATION |    COST    |
+-----+---------------------+...+-----------+----------------------+----------+------------+
| g...| tidyverse-t3.medium |...| Executing | 2023-12-11T23:22:17Z | 20m37s   | 0.0066 USD |
+---------------------------+...+-----------+----------------------+----------+------------+
$ float restart
Warning: There are running jobs in server, do you want to restart forcibly?
Some job may fail if you do that!(yes/No): yes
Warning: Are you sure you want to restart OpCenter? All active sessions will be inactivated.(yes/No): yes
OpCenter is now restarting and will be online soon.
Pause briefly and then log back in
$ float list
+-----+---------------------+...+-----------+----------------------+----------+------------+
| ID  |        NAME         |...|  STATUS   |     SUBMIT TIME      | DURATION |    COST    |
+-----+---------------------+...+-----------+----------------------+----------+------------+
| g...| tidyverse-t3.medium |...| Executing | 2023-12-11T23:22:17Z | 21m42s   | 0.0072 USD |
+---------------------------+...+-----------+----------------------+----------+------------+
(edited for clarity)

float resume (recover, restore)

Use the float resume (or recover or restore) command to resume a suspended job (see float suspend command). The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Resume a suspended job. If no options used, resume on identical instance as original..	-c, --cpu min_cpu:max_cpu	Range of number of virtual CPUs to select from for new VM instance (can omit max_cpu).
		-t, --instType instance_type	Instance type for new VM (e.g., c4.xlarge in AWS). Do not combine with --cpu or --mem options.
		-j, --job job_id	Job to resume
		-m, --mem min_mem:max_mem	Range of memory size (in GB) to select from for new VM instance (can omit max_mem).
		-P, --payType spot | ondemand	Pricing tier for VM instance (Spot or On-demand).

float sbatch (submit)

The float sbatch command is the same as float submit.

float scancel (cancel)

The float scancel command is the same as float cancel.

float secret

Use the float secret command to interact with the secret manager. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
get	Retrieve secret from the secret manager	secret_name	Name associated with secret
ls, list	List all secrets created by current user
set, put	Insert {name, value} pair into secret manager database	secret_name secret_value	{name, value} pair to insert
		--file /path/to/file	File containing secret_value. Use instead of specifying secret_value on command line.
		-f, --force	Overwrite existing secret value
unset, rm, del, remove, delete	Delete {name, value} pair from secret manager database	secret_name	Name associated with {secret_name, secret_value} pair to delete

Example

$ float secret set s3access 1234567
Set s3access successfully
$ float secret ls
+----------+
|   NAME   |
+----------+
| s3access |
+----------+
$ float secret unset s3access
unset secret s3access successfully

$ cat value.txt
secret123
$ float secret set usersecret --file ./value.txt
Set usersecret successfully

float show

Use the float show command to display the status of a job and content of job scripts. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Show current status of job or VM instance, or show contents of scripts used for job	-C, --cluster cluster_name | cluster_id	Cluster in which job is running
		--containerInitScript	Display contents of container init script
		-c, --content	Display contents of job script
		-i, --hid host_id	VM instance to query
		--hostInitScript	Display contents of host init script
		--hostTerminateScript	Display contents of host terminate script
		-j, --job job_id	Job to query
		-q, --queue queue_name | queue_id	Queue to query

Example

$ float show -j FOIW5Y16KZgJ6Tsd02QuS
id: ctZLDo7OFG4BuJ8ytiTem
name: python-c5d.large
workingHost: 52.7.123.178 (2Core4GB/Spot)
user: admin
imageID: docker.io/bitnami/python:latest
imageDigest: sha256:24c1d45bf41c396184bd9808b307c67267a809754cd176ac8d91cceb47d0f3ef
output: |-
    Getting image source signatures
    Copying blob sha256:1acb894a7ceb1ba5362fb85123b0248a064ed3195aaa24af74e8cb710ca1c5a4
    Copying config sha256:23bacce690702dac91557ef74ab312cb3db5a2b4bb54ada968d1352e7d9a110a
    Writing manifest to image destination
    Storing signatures
    Loaded image: docker.io/bitnami/python:latest
    First submit job ctZLDo7OFG4BuJ8ytiTem, call podman directly
    No cmd args provided, launch job directly
    4eda6b0d471fb31eecfd06b59e1d8c527310861009fa4baf967d834397934bac
status: Executing
.....[output edited]

$ float show -c -j S0JPnWd3a2hQmVmVWCMWc
#!/usr/bin/bash
LOG_PATH=$1
LOG_FILE=$LOG_PATH/output
touch $LOG_FILE
exec >$LOG_FILE 2>&1
echo "Congratulations! You have submitted your first job"
for(( c=1; c<3; c++))
do
        if [[ $(($c % 3)) == 1 ]] then
                echo "Hello World!"
        else
                echo "Your next job will be more interesting" >&2
        fi
                sleep 20s
done
echo "Job complete"

float sinfo (hosts)

The float sinfo command is the same as float hosts.

float smtp

Use the float smtp command to define and manage email (SMTP) configurations. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add	Update an SMTP configuration	-n, --name smtp_config	Name to associate with the SMTP config
		--from from_addr	Email address to use in From: field of email message
		-P, --password user_pw	Password for user
		--passwordStdin	Prompt user to enter password
		--port port_num	Port number for SMTP server
		S, --server ip_addr	IP address of SMTP server
		--useSSL	Use SSL for server connection
		--useTLS	Use TSL for server connection
		-U, --user user_name	Name of SMTP user
delete	Delete an SMTP configuration	smtp_id	ID that identifies SMTP configuration to delete
get	Get SMTP configuration	smtp_id	ID that identifies SMTP configuration to get
		--password show_pwd	Password to get SMTP configuration
list, ls	List all SMTP configurations
update	Update an SMTP configuration	-i , --id smtp_id	ID to identify SMTP config
		--from from_addr	Email address to use in From: field of email message
		-n, --name smtp_name	Name to identify SMTP config
		-P, --password user_pw	Password for user
		--passwordStdin	Prompt user to enter password
		--port port_num	Port number for SMTP server
		S, --server ip_addr	IP address of SMTP server
		--useSSL	Use SSL for server connection
		--useTLS	Use TSL for server connection
		-U, --user user_name	Name of SMTP user

float snapshot

Use the float snapshot command to display information about snapshots. Use subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
list	List available snapshots	-A, --all	Show all snapshots associated with job
		-f, --filter filter	Filter to apply to snapshots, for example: status=normal. The default filter is active=true. Use option multiple times to apply multiple filters combined with "and" operator.
		-j, --jobID job_id	Job to query
show, info	Display information about snapshot	-s, --snapID snapshot_id	ID of snapshot to query

Example:

$ float snapshot show -s jobSnap-Kp89d7qzJjiuhNeaOUfsA
id: jobSnap-Kp89d7qzJjiuhNeaOUfsA
jobID: h3xg4jpC8ydcgkVWBKj4S
volumeSnapshots:
    - zone: us-east-1b
      volumeId: vol-02786458cebdeca43
      volumeSize: 6
      status: completed
      snapshotId: snap-08e5b6ed8b7834b96
      createTime: "2023-04-18T15:51:17Z"
      mountPoint: /mnt/float-data
      cost: 0.0000 USD
    - zone: us-east-1b
      volumeId: vol-06dd42bf6b091466a
      volumeSize: 6
      status: completed
      snapshotId: snap-0f2833cdb639fa4b5
      createTime: "2023-04-18T15:51:17Z"
      mountPoint: /mnt/float-image
      cost: 0.0000 USD
    - zone: us-east-1b
      volumeId: vol-023c2e91e34f37335
      volumeSize: 10
      status: completed
      snapshotId: snap-025a733db664e9054
      createTime: "2023-04-18T15:51:17Z"
      mountPoint: /data
      cost: 0.0000 USD
status: normal
createTime: 2023-04-18T15:51:17.434161661Z

float squeue (list)

The float squeue command is the same as float list.

float status

Use the float status command to show current status of the OpCenter. The command has no subcommands. Use the command with the -h flag to list the options.

Example

$ float status
id: b39a9d89-cf9e-4b55-9399-f792c2bec020
Server Status: normal
API Request Status: normal
License Status: valid
Init Time: "2025-04-02T22:59:51Z"
Up Time: "2025-04-02T23:01:56Z"
Current Time: "2025-04-26T22:52:27Z"

float storage

Use the float storage command to "register" (that is, pre-configure) storage services for use when submitting jobs. Use subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
delete, rm, remove	Delete a registered storage service	storage_id	Identifier associated with the storage service to delete
info, show	Display information about a registered storage service	storage_id	Identifier associated with the storage service to query
list	Display filtered list of registered storage services	-f, --filter filter	Filter to apply to registered storage services. Use option multiple times to apply multiple filters combined with "and" operator. Simple filters consist of an attribute, operator, and value. Supported attributes include: id name storageType url zone path createdTime
		-o, --orderBy attribute	Attribute used to order listing (prepend with "-" to reverse order). Use any filter attribute (default: createdTime). Create nested ordering by connecting multiple attributes with ",".
register (add) volume	Register an existing volume as a storage service	-n, --name storage_name	Name to associate with this storage service
		--id volume_id	Volume ID that identifies volume (for example, an identifier like vol-0da677a1f4967350a in AWS)
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
register (add) nfs	Register NFS-exported directory as a storage service	-n, --name storage_name	Name to associate with this storage service
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
		--url nfs://nfs_server_ip/exported_dir	IP address of NFS server and exported directory, for example 192.168.1.1/home
register (add) lustre	Register Lustre file system as a storage service	-n, --name storage_name	Name to associate with this storage service
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--options string	Mount options for lustre file system, for example, [opts=rw, noexec, abort_recov, recovery_time_hard=120s]
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
		--url lustre://dns-name/mountname/exported_dir	DNS name of lustre server, file system's mount name, and exported directory, respectively. The mount name is the name you assign the file system when configuring FSx for Lustre on AWS. Naming the file system makes it easier to find.
register (add) s3	Register S3 bucket as a storage service	-n, --name storage_name	Name to associate with this storage service
		--bucket s3://bucket_name/folder	S3 bucket name and folder (optional) to use for storage service
		-c, --credential [accessKey=key_value, secret=secret_value,]	Access key string and access secret value, respectively
		--endpoint s3_endpoint	Endpoint for S3 bucket, for example, s3.us-east-1.amazonaws.com
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
register (add) gs	Register Google Cloud Storage as a storage service	-n, --name storage_name	Name to associate with this storage service
		--bucket gs://bucket_name/folder	Google Cloud Storage bucket name and folder (optional) to use for storage service
		-c, --credential [accessKey=key_value, secret=secret_value,]	Access key string and access secret value, respectively
		--endpoint gs_endpoint	Endpoint for GCS bucket, for example, storage.googleapis.com
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
register (add) omics	Register a new omics sequence readSet	-n, --name storage_name	Name to associate with this storage service
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--options string	Mount options for this storage service
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
		readSetId string	readSet id for the omics storage
		--seqId string	Sequence id for the omics storage
register	Register storage service using dataVolume format (instead of register volume|nfs|lustre|s3|gs).	--dataVolume [string]	Parameters that define the storage service. Use the following format. volume (EBS): [mode=rw]volume-id[:/mountpoint] or [mode=rw]volume://volume-id[:/mountpoint] NFS: [mode=rw]nfs://172.31.2.120/path[:/mountpoint] S3: [accesskey=xxx,secret=yyy,endpoint=zzz,mode=rw]s3://bucketname/path[:/mountpoint] GCS: [accesskey=xxx,secret=yyy,endpoint=zzz,mode=rw]gs://bucketname/path[:/mountpoint] Lustre: [opts=xxx,yyy,zzz,mode=rw]lustre://dnsname/mountname[:/mountpoint]
update volume	Update an existing volume as a storage service	storage_id	Identifier associated with this storage service
		--id volume_id	Volume ID that identifies volume (for example, an identifier like vol-0da677a1f4967350a in AWS)
		-n, --name storage_name	Name to associate with this storage service
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
update nfs	Update NFS-exported directory as a storage service	storage_id	Identifier associated with this storage service
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		-n, --name storage_name	Name to associate with this storage service
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
		--url nfs://nfs_server_ip/exported_dir	IP address of NFS server and exported directory, for example 192.168.1.1/home
update lustre	Update Lustre file system as a storage service	storage_id	Identifier associated with this storage service
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		-n, --name storage_name	Name to associate with this storage service
		--options string	Mount options for lustre file system, for example, [opts=rw, noexec, abort_recov, recovery_time_hard=120s]
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
		--url lustre://dns-name/mountname/exported_dir	DNS name of lustre server, file system's mount name, and exported directory, respectively. The mount name is the name you assign the file system when configuring FSx for Lustre on AWS. Naming the file system makes it easier to find.
update s3	Update S3 bucket as a storage service	storage_id	Identifier associated with this storage service
		--bucket s3://bucket_name/folder	S3 bucket name and folder (optional) to use for storage service
		-c, --credential [accessKey=key_value, secret=secret_value,]	Access key string and access secret value, respectively
		--endpoint s3_endpoint	Endpoint for S3 bucket, for example, s3.us-east-1.amazonaws.com
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		-n, --name storage_name	Name to associate with this storage service
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)
update gs	Update Google Cloud Storage as a storage service	storage_id	Identifier associated with this storage service
		--bucket gs://bucket_name/folder	Google Cloud Storage bucket name and folder (optional) to use for storage service
		-c, --credential [accessKey=key_value, secret=secret_value,]	Access key string and access secret value, respectively
		--endpoint gs_endpoint	Endpoint for GCS bucket, for example, storage.googleapis.com
		--mode ro | rw	Access mode ro: read only rw (default): read and write
		-m, --mountPoint path	Path to directory where volume is mounted by container
		-n, --name storage_name	Name to associate with this storage service
		--permision normal | public	Storage service access permission normal (default): storage accessible by all users in group (editable by admin and owner) public: storage accessible by all users (editable by admin and owner)

float submit

Use the float submit (or float sbatch) command to submit a job. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Submit job for execution	--agentPort port_num	Agent port (default:443). If not using default, confirm OpCenter can access port.
		--allowList instance_type	Allowed instance type(s). Use format like "c5*" which allows any VM type in the c5 family. Include --allowList multiple times to add other VM instance types to the allow list. Default is [*] which allows all types.
		--bandwidth bandwidth	Bandwidth (in Mbps) required for workload (only use in AliCloud)
		-A, --cmdArgs command_string	Commands that are executed immediately after container starts (can be used with or without a job script)
		-c, --cpu min_cpu:max_cpu	Range of number of virtual CPUs to select from to run job (can omit max_cpu). If mac_cpu omitted, value set to twice min_cpu.
		--cpuVendor cpu_vendor_name	CPU vendor of allowed VM instances. Allowed values are amd, intel, or " ". Default is " " which allows any vendor. Use with AWS only.
		--customTag tagName:tagValue	Tag customized by each MMCloud subscriber, for example, to generate customized usage reports. Multiple custom tags can be applied by using --customTag multiple times in a single command.
		-D, --dataVolume [size=vol_size,throughput=rate]:/data_dir or [mode=rw]vol_id :/container_mnt_pt/path or [mode=rw]nfs://ip_address/server_export_dir :/container_mnt_pt/path or [accesskey=x,secret=y,token=t,endpoint=z,mode=rw]s3://bucketname/path :/container_mnt_pt/path or [accesskey={secret:S3_BUCKET_ACCESS_KEY},secret={secret:S3_BUCKET_SECRET_KEY},endpoint=z,mode=rw]s3://bucketname/path :/container_mnt_pt/path or [credfile=aws.conf,profile=user1,endpoint=z,mode=rw]s3://bucketname/path :/container_mnt_pt/path or [opts=xx,yy]lustre://dnsname/mountname/path :/container_mnt_pt/path or [accesskey=xxx,secret=yyy]jfss3://bucketname :/container_mnt_pt/path	Data volume mounted by container. Use option multiple times to attach multiple data volumes. (New EBS volume) vol_size in GB, rate in Mbps (can omit), data_dir is data directory created (Existing EBS volume) vol_id specifies existing EBS volume, /container_mnt_pt/path is path to where EBS volume is mounted on container (NFS mount) ip_address is IP address of NFS server, /server_export_dir is directory exported by NFS server, and /container_mnt_pt/path is path to where NFS directory is mounted on container (S3|GCS|OSS bucket) bucketname is S3 bucket (access credentials optional if action allowed by IAM role). Replace s3 with gs for GCP or oss for AliCloud. (Lustre FS) dnsname is the DNS name for the Lustre server and mountname is the name assigned to the Lustre file system (JuiceFS on S3) bucketname is the name of the S3 bucket used as the physical data store for the JuiceFS file system
		-d, --def definition_file	Path to file if using definition file (yaml or json format) to provide input parameters to submit command
		--denyList instance_type	Excluded instance type(s). Use format like "c5*" which excludes all VM types in the c5 family. Include --denyList multiple times to exclude other VM instance types. Default is [ ] which does not exclude any types.
		--disableImageVol	Disable float-image volume creation for job. Only use if job migration disabled.
		--dumpMode full | incremental	Setting for snapshot type. Full means complete memory snapshot taken every time. Incremental means only incremental changes captured after initial snapshot. Default is full.
		-e, --env env_key=env_value	Environment variable setting for the job. Include --env multiple times to add multiple variables.
		-E, --errPolicy err_policy	Policy (AWS only) to use if the job fails. The choices are: reclaimAll: reclaim virtual machine and all EBS volumes (default) retainVolumes: reclaim virtual machine and retain all EBS volumes. User must manually resume or cancel job. Retained EBS volumes mounted on new virtual machine. restart: retain virtual machine and all EBS volumes; resubmit the job on same virtual machine with same EBS volumes (job may run without stopping if the application does not limit the number of restarts) autosuspend: if OOM protection triggers migration and no instance type found, suspend job; user must cancel job or resume manually
		--extraContainerOpts	Extra options passed directly to container (enclose in quotes)
		-f, --force	Automatically answer "yes" at confirmation prompt
		--gateway gateway_id	ID of gateway to connect job to. Replacing gateway_id with the word "auto" directs the OpCenter to select a gateway automatically.
		--gpu-count min_gpu : max_gpu	Range of number of GPUs to select from to run job
		--gpu-disable	Disallow the use of GPUs for this job
		--gpu-mem min_gpu_mem : max_gpu_mem	Range of size of GPU memory to select from to run job
		--gpu-name GPU_NAME	Specific GPU to use for this job, for example, m60 or h100 or a100, and so on.
		--gpu-vendor GPU_VENDOR	Allowed GPU vendor to use for this job, for example, nvidia or amd.
		--hyperThread thread_setting	Hyper-thread setting for job (0 means use the setting configured by the cloud provider, 1 means disable, 2 means enable)
		-i, --image image_name | image_URI	Image name or image URI to pull container image for job
		--imageVolSize image_vol_size	Size of volume (in GB) to act as root volume for image (default: 6)
		--imageVolType image_vol_type	(AWS only) Type of EBS volume to use for image volume, for example, gp2 or gp3.
		-t, --instType instance_type	VM instance type, for example, c5.4xlarge for AWS. Overrides --cpu or --mem options.
		-j, --job job_script	Job script to run workload. Use format: path to local file, s3 file, OSS file or https | http file.
		-k, --keepDumpLog	Save dump log during job migration
		-K, --keepPrivateIP	Keep private IP address during job migration
		--mem min_mem:max_mem	Range of memory size (in GB) to select from to run job. If max_mem omitted, value set to twice min_mem.
		--metricsInterval metrics_int	Time in seconds between queries to obtain container metrics (default: 10s)
		-M, --migratePolicy [migrate_policy]	Policy to determine auto-migration behavior. Format is [option1=value1,option2=value2...]. The available options with their default values are (if no units are attached, the value is a percentage): disable=false evadeOOM=true stepAuto=true cpu.disable=true cpu.upperBoundRatio=90 cpu.lowerBoundRatio=5 cpu.upperBoundDuration=120s cpu.lowerBoundDuration=5m0s cpu.step=50 cpu.limit=0 (no limit) cpu.lowerLimit=0 (no limit) mem.disable=true mem.upperBoundRatio=90 mem.lowerBoundRatio=5 mem.upperBoundDuration=120s mem.lowerBoundDuration=5m0s mem.step=50 mem.limit=0 (no limit) mem.lowerLimit=0 (no limit) cost.disable=true cost.threshold=0.9 cost.crossInstType=true cost.crossZone=true cost.optimizeSpot=true
		--miTag [tag1:value1 tag2:value2...]	Tag(s) to select virtual machine image. Use option multiple times to submit multiple tag,value pairs.
		-n, --name job_name	Name to associate with job
		--noPublicIP	No public IP address assigned to container host. Ensure that host can reach AWS services. AliCloud users must include --endpoint option in their job scripts.
		-o, --output /path/to/dir	Folder to save stdout and stderr as files with names: stdout.autosave.$jobid and stderr.autosave.$jobid. Options for path are: nfs://nfs_server_ip/remote_folder file:///local_folder /local_folder (local_folder must be mounted by worker node)
		--outputFlag	If included (set to true), job status is included in job output folder.
		-P, --publish host_port:container_port	Rule for publishing container port to container host port, for example, 8080:80. Include option multiple times to publish multiple ports. Confirm security group allows access to port.
		--rootVolSize root_vol_size	Root volume size in GB to load base OS (default: 40)
		--securityGroup sec_group	AWS security group (or tag in GCP) added to VM instance for this job. Use option multiple times to apply multiple security groups.
		--shmSize shm_size	Size of /dev/shm in format nu where n is a number and u is b, k, m, or g for bytes, KiB, MiB or GiB, respectively (default: 64m)
		--snapLocation local | s3://bucketname	Location to save snapshot image and metadata. Choices are: local: volume automatically mounted by OpCenter (EBS for AWS, Persistent Disk for Google Cloud, Block Storage for AliCloud) s3://bucketname: JuiceFS with bucket type dependent on available cloud object storage service. Replace s3 with gs in GCP. Replace s3 with oss in AliCloud.
		--snapSkipOpenFileLis file1,file2,...	List of files to skip during snapshot. The format is: *.log,*.tmp
		-I, --snapshotInterval snap_int	Time between periodic snapshots in format HhMmSs where H, M, and S are numbers and h, m, and s are hours, minutes, and seconds, respectively. Default is 0 (disable). Minimum is 10m.
		--storage id | name:/container_mnt_pt/path/to/dir	Registered storage service included with job. Specify storage service by id or name. Specify path to where storage service is mounted on container.
		-s, --subnet subnet_id	AWS subnet ID or AliCloud vSwitch ID in which to execute the job. Default is that the OpCenter automatically selects the subnet.
		--swapDurationOnOOM duration	Duration for which swap memory is used to trigger OOM evasion (format: Nm). Default is 0 which means use system level setting.
		--swapSize swap_size	Swap size in GB to use for job. Defaul is 0 which means use system level swap size setting.
		--swapUsageOnOOM swap_usage	Size of swap memory used (measured as a fraction of total swap size) to trigger OOM evasion (format: 0.5). Default is 0 which means use system level setting.
		--tag tag_name	Tag to select image version for job. Default is "latest" or only tag.
		--targetPort port_num	Port to connect job to on gateway. Use option multiple times to connect multiple ports.
		-T, --template template_name	Template in format name:tag to use for submitting this job. Default tag is only tag or "latest".
		-l, --timeLimit max_time	Maximum time that a job is allowed to run. Use format hhmmss where h, m, and s are numbers and h, m, and s are hours, minutes, and seconds, respectively. Default is unlimited (use 0).
		-V, --vmPolicy [vm_policy]	VM creation policy. Format is [key1=value1,key2=value2...]. The allowed keys (and values) are: spotFirst (true | false): Try to start spot instance. If none available, start on-demand instance. spotOnly (true | false): Try to start spot instance. If none available, stop. onDemand (true | false): Start on-demand instance retryLimit (n): Number of cycles of attempts to start spot instance retryInterval (ns): Time to wait between cycles of attempts priceLimit (number in $/hour): Maximum price to bid for spot instance optimize (true | false): Migrate job from on-demand instance to spot instance when spot instances are available. mode (loose | strict | same): CPU compatibility check when migrating to new compute instance. Loose: same model and same or newer version. Strict: CPU metadata compatible with new instance. Same: same model and same version. maxSpotReclaim (integer): When this number of spot reclaim events reached, job migrates to on-demand instance The default VM creation policy is: spotFirst=true retryLimit=3 optimize=true retryInterval=10m0s priceLimit: no upper bound on price imposed mode=loose maxSpotReclaim=0 (no limit on number of spot reclaim events)
		--withRoot	Run job with root privileges
		-z, --zone availability_zone	Availability zone in which to execute job

Examples

$ float submit -i tidyverse -j run_genericr.sh --cpu 4 --mem 8 --dataVolume [size=10]:/data
id: pw6nupnmbej0qedlvz3dx
name: tidyverse
user: admin
imageID: docker.io/rocker/tidyverse:latest
status: Submitted
submitTime: "2023-12-13T19:08:27Z"
duration: 0s
queueTime: 0s
cost: 0.0000 USD
inputArgs: -i tidyverse -c 2 -m 4 --dataVolume [size=10]:/data
cpu: 2
memGB: 4
vmPolicy:
    policy: spotFirst
    retryLimit: 3
    optimize: true
    retryInterval: 10m0s
migratePolicy:
    evadeOOM: true
    stepAuto: true
    cpu:
        upperBoundRatio: 90
        lowerBoundRatio: 5
        upperBoundDuration: 2m0s
        lowerBoundDuration: 5m0s
        step: 50
        disable: true
    mem:
        upperBoundRatio: 90
        lowerBoundRatio: 5
        upperBoundDuration: 2m0s
        lowerBoundDuration: 5m0s
        step: 50
        disable: true
actualEmission: 0.0000 g
baselineEmission: 0.0000 g

$ float submit -i docker.io/centos -j helloworld.sh -c 2 -m 4 --dataVolume [size=10]:/data  --name hw -o nfs://172.31.81.17/mnt/memverge/shared

Log in to NFS server and check that output files are populated.

$ ls -l /mnt/memverge/shared/*
-rw-r--r-- 1 5001 5001   0 Feb 15 20:02 /mnt/memverge/shared/hw.08ogctot3ufqnyf9k3wg8.stderr.autosave
-rw-r--r-- 1 5001 5001 116 Feb 15 20:02 /mnt/memverge/shared/hw.08ogctot3ufqnyf9k3wg8.stdout.autosave

float suspend (hibernate)

Use the float suspend (or hibernate) command to temporarily suspend a job. In-memory state and relevant files are saved so that the job can resume executing at a later time. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Suspend a running job	--cold	With this option, storage resources are reclaimed, which may save cost, but it slows the suspend operation (and the associated resume operation). Use if suspending for extended period.
		-f, --force	Automatically answer "yes" to confirmation prompts
		-j, --job job_id	ID of job to suspend

float template

Use the float template command to use a template or to display information about templates. Use subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
delete, rm, remove	Delete a template	-f	Force deletion of template
		-T, --template name:tag	Template name and tag. If tag omitted, default is "latest."
deploy	Submit a job using a template	-c, --cpu min_cpu:max_cpu	Range of number of virtual CPUs to select from to run job (can omit max_cpu, in which case max_cpu is set to twice the value of min_cpu).
		--gateway gw_id	ID of gateway to connect job to (if applicable). Use `auto` to have the system select a gateway automatically.
		-t, --instType instance_type	VM instance type (e.g., c5.xlarge for AWS). If used, this option overrides -c and -m options.
		-m, --mem min_mem:max_mem	Range of memory size (in GB) to select from to run job (can omit max_mem , in which case max_mem is set to twice the value of min_mem).
		--noPublicIP	No public IP address assigned to container host. Make sure host can reach internal cloud services.
		--securityGroup SG1,SG2,...	AWS security group(s) (or tag(s) in GCP) applied to VM instance for this job. Include option multiple times to apply multiple security groups (or tags). Use [ ] to clear all security groups or tags.
		--targetPort port_num1,port_num2,...	Port(s) used to connect to job, for example, 8787 for RStudio. Include multiple port numbers to connect multiple ports from a single server to the gateway. Use [ ] to clear all ports.
		-T, --template name:tag	Template (in format name:tag) to use for submitting this job. Default tag is only tag or "latest".
info, show	Display information about a template	-T, --template name:tag	Template (in format name:tag) to query. Default tag is only tag or "latest".
list, ls	Display available templates. Templates with source labeled "official" are from the MemVerge template repository.
save	Save submit string from an earlier job as a template	-f, --force	Force template save
		-j,--job job_id	ID of job whose submit string is saved as template
		--overwrite	Overwrite existing template
		-T, --template name:tag	Name and tag to save template as. Default tag is "latest".
sync	Sync OpCenter template library with the MemVerge template repository

float top

Use the float top command to show a sorted list of information (including CPU and memory utilization) about current jobs. The display is updated at regular intervals. Enter q to stop the display. The command has no subcommands. Use the command with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
	Show utilization and other information about running jobs. Display updates continually. Enter q to exit.	--interval top_int	Time in seconds (3 or greater) between queries to obtain container metrics (default: 3s). Format is ns where n is integer.
		-j, --job_Id job_id	Job to query

float user

Use the float user command to manage OpCenter users. Use a subcommand with the -h flag to list the options.

Subcommands	Usage	Option	Option Definition
add, create	Add a new user	new_username	Username for new user
		--create	Create group if the group does not exist (default: false)
		--email email_address	Email address for new user
		-gid gid	Group ID to use for group. If value not provided, gid automatically assigned.
		-g, --groups group_name	Group associated with new user (default: no group). Use option multiple times to include multiple groups.
		--ldap	Associate username with LDAP directory (default: false)
		--passwd password	Password for new user (default: "memverge")
		--quota-policy policy_name	Name of quota (SurfZone) policy to associate with this user
		--uid uint32	User Identifier (uid) for new user. For regular user accounts, uid normally starts at 1000.
delete, remove, rm	Delete a user	user_name | user_id	Username or uid of user to delete
		-f, --force	Remove all active tokens belonging to user and delete user immediately
disable	Disable a user's account without deleting it	user_name | user_id	Username or id of user to query
enable	Enable a user's account	user_name | user_id	Username or id of user to query
info, show	Display information about user	user_name | user_id	Username or id of user to query
list, ls	List groups that current user belongs to
passwd	Reset user's password	user_name | user_id	Username or id of user to update
		--passwd new_password	New password for user
update	Update information associated with a user	user_name | user_id	Username or id of user to update
		--email email_address	New email address for user
		-gid gid	New group ID to use for group.
		-g, --group group_name	New group associated with user
		--name user_name	New username for user
		--passwd password	New password for user
		--quota-policy policy_name	New quota (SurfZone) policy to associate with this user
		--uid uint32	New uid for user

Examples

$ float user add testcase2 --passwd secret123 --groups crops
username: testcase2
uid: 5007
gid: 5007
role: normal
group: crops
email: ""
type: builtin
enabled: true
ownGroup: ""

$ float group update crops --add testcase2 --admin

$ float user info testcase2
username: testcase2
uid: 5007
gid: 5007
role: normal
group: crops
email: ""
type: builtin
enabled: true
ownGroup: crops

$ float user update testcase2 --group default
username: testcase2
uid: 5007
gid: 5007
role: normal
group: default,crops
email: ""
type: builtin
enabled: true
ownGroup: crops

$ float user info user-five
username: user-five
uid: 2012
gid: 2010
role: normal
group: group-five
email: ""
type: ldap
enabled: true
ownGroup: ""

float version

Use the float version command to display the version of the float CLI client and the version of the OpCenter it is connected to. The command has no subcommands. Use the command with the -h flag to list the options.

Example

$ float version
float version: v3.0.0-69ce0c9-Imperia
OpCenter version: v3.0.0-69ce0c9-Imperia