This tool is intended to be an easy-to-use command-line interface that enables you to do the following:
- Quickly install a fully functional interface to the JSON-RPC using pip.
- Get help on how to use the commands by accessing the inline help.
- Execute any command supported by NetApp SolidFire Element OS.
- Utilize autofill on MacOS and Linux distributions to make constructing commands easier.
- Get responses in three different output formats: tree format, JSON format, and pickle format.
- Execute any non-public commands by using "sfcli sfapi invoke" with the necessary method name and parameters.
- Achieve the same functionality as the SDKs without ever opening an editor.
- Establish and store connection information locally.
The tool is for users who do not want to use PowerShell. It integrates as you would expect with bash. The main advantages of using this tool are the following:
- Objects are returned as strings, which allows for expeditious grepping.
- It is very easy to install and use.
Note: If you are on a Linux OS, you will have to set up a virtual environment first because the native linux python does not support setuptools. To create a virtualenvironment, run
pip install virtualenv
virtualenv pythoncli
source pythoncli/bin/activate # This activates the virtual environment.
Installation Steps
To install from pypi using pip, execute the following command:
pip install solidfire-cli
-- or -- Before you begin, must have solidfire-sdk-python installed in your environment. To install from source using easy_install, navigate to solidfirecli*.tar.gz file (where the * stands for the version), and run
easy_install solidfirecli*.tar.gz
- Windows 7, 8, and 10
- Linux
- MacOS
The tool includes inline help that provides details about how to use the commands.
To see the top-level help, run:
sfcli --help
To see the help at the object level (for example, account), run:
sfcli account --help
To see command-specific help, run:
sfcli account getbyid --help
Autocomplete is only available on mac and linux. To enable it, enter the following into your .bashrc file or run it whenever you want autocomplete enabled:
eval "$(_SFCLI_COMPLETE=source sfcli)"
To run a command on a given connection without storing it away, use the mvip, username, and password options at the top level.
sfcli --mvip 10.117.60.15 --username username --password password account list
To connect to an individual node rather than the cluster admin node, add --port 442 to connection options.
To store a given connection, use the connection push command and supply the name option.
sfcli connection push --mvip 10.117.60.15 --username username --password password --name "Example"
There are three ways to access a stored connection: by name, by index, or by default. To access a connection by name, use -n or --name. To access a connection by index, use -c or --connectionIndex. To access a connection by default, leave the above two parameters off and the command will default to using the connection at index 0. The three possibilities are exemplified below:
sfcli -n Example account list # by name
sfcli -c 0 account list # by index
sfcli account list # use connection 0
To remove a given connection, use the connection remove command.
sfcli connection remove -n Example
sfcli connection remove -i -1 # Removes the newly pushed connection.
sfcli connection remove -i 0 # Removes the oldest pushed connection.
sfcli connection remove -i 1 # Removes the second oldest connection.
To list the stored connections, use the Connection List command.
sfcli connection list
To prune broken connections from the connections.csv file, use the Connection Prune command.
sfcli connection prune
Determine the needed parameters using the autocomplete functionality or the --help functionality and then append the required parameters at the end as shown below:
sfcli -c 0 account getbyid --accountid 3065
Example output:
account:
attributes:
target_secret:
secret: q7788;0a:Cd0xCE4
initiator_secret:
secret: r0Ju1}t:02"5p<L^
status: active
account_id: 3065
username: example
volumes:
A parameter might have optional subparameters that you need to use to specify an attribute of the the parameter. In the following example, the --volumes parameter has subparameters, such as volumeid, accessvolumes, name, newaccountid, newsize,and attributes:
sfcli volume clonemultiple --help
Usage: sfcli volume clonemultiple [OPTIONS]
CloneMultipleVolumes enables you to create a clone of a group of specified
volumes. You can assign a consistent set of characteristics to a group of
multiple volumes when they are cloned together. Before using groupSnapShotID
to clone the volumes in a group snapshot, you must create the group snapshot
by using the CreateGroupSnapShot API method or the Element OS Web UI. Using
groupSnapshotID is optional when cloning multiple volumes.
Note: Cloning multiple volumes is allowed if cluster fullness is at stage 2 or 3.
Clones are not created when cluster fullness is at stage 4 or 5.
Options:
--volumes Unique ID for each volume to include
in the clone. If optional parameters are not specified,
the values are inherited from the source
volumes. Has the following subparameters:
--volumeid --accessvolumes --name
--newaccountidvolumes --newsize --attributes
[required]
--volumeid INTEGER [subparameter] Required parameter for
"volumes" array: volumeID. [required]
--accessvolumes TEXT [subparameter] Access settings for the new
volume. readOnly: Only read operations are
allowed. readWrite: Reads and writes are
allowed. locked: No reads or writes are
allowed. replicationTarget: Identify a volume
as the target volume for a paired set of
volumes. If the volume is not paired, the
access status is locked. If unspecified, the
access settings of the clone will be the same
as the source.
--name TEXT [subparameter] New name for the clone.
--newaccountidvolumes INTEGER [subparameter] Account ID for the new volume.
--newsize INTEGER [subparameter] New size Total size of the
volume, in bytes. Size is rounded up to the
nearest 1MB size.
--attributes TEXT [subparameter] List of name-value pairs in
JSON object format.
--access TEXT New default access method for the new volumes
if not overridden by information passed in
the volume's array. readOnly: Only read
operations are allowed. readWrite: Reads and
writes are allowed. locked: No reads or
writes are allowed. replicationTarget:
Identify a volume as the target volume for a
paired set of volumes. If the volume is not
paired, the access status is locked. If
unspecified, the access settings of the clone
will be the same as the source.
--groupsnapshotid INTEGER ID of the group snapshot to use as a basis
for the clone.
--newaccountid INTEGER New account ID for the volumes if not
overridden by information passed in the
volumes array.
--help Show this message and exit.
To clone multipe volumes in the same group, run:
sfcli volume clonemultiple --volumes --volumeid 1 --accessvolumes readWrite --volumes --volumeid 1
This clones the volumes and applies readWrite access to volume 1 and uses the default for volume 2.
Occasionally, you will need to pass in an arbitrary dictionary of keys and values. In that case, you will need to provide a JSON string directly to the command line.
In the example below, if you want to bypass the logic and make a call directly to the API, you can do so by using SFApi Invoke.
$account = sfcli -c 0 SFApi Invoke --method GetAccountByID --parameters "{\"accountID\":94}"
The tree format enables you to grep the output and get a result without extra formatting. It uses whitespace to distinguish between different objects. To view this format, run a command without the -j or -k options.
See the following example:
accounts:
attributes:
storage_container_id:
int: <to see more details, increase depth>
account_id: 2404
status: active
initiator_secret:
secret: <to see more details, increase depth>
target_secret:
secret: <to see more details, increase depth>
volumes:
<to see more details, increase depth>
username: EXAMPLE
attributes:
storage_container_id:
int: <to see more details, increase depth>
account_id: 2405
status: active
initiator_secret:
secret: <to see more details, increase depth>
target_secret:
secret: <to see more details, increase depth>
volumes:
username: haxecliFV8QdeT6fn5DxtvFuYzjsFwWtc1YzXfT5-NQE5pHiQAQBelNqVskTsJY8
attributes:
storage_container_id:
int: <to see more details, increase depth>
account_id: 2406
status: active
initiator_secret:
secret: <to see more details, increase depth>
target_secret:
secret: <to see more details, increase depth>
volumes:
username: haxeclij0yA7YVfCDiq9jZXdkdiKfkSytK2flKk9Gi9NFq0677Fcg44QIDc9inqF
The output in JSON format resembles the output of the API. This can be useful if you are going to store the data for later use with Postman or if you want to import it via any JSON libraries. To get this output, use the -j option after "sfcli" in your command.
See the following example:
{
"accounts": [
{
"attributes": {},
"initiator_secret": {
"secret": "3,gG[sP02V'@911}"
},
"volumes": [
4588
],
"target_secret": {
"secret": "aAe6Bb&q]63zU0Ei"
},
"status": "active",
"account_id": 2404,
"username": "EXAMPLE",
"storage_container_id": {
"hex": "00000000000000000000000000000000"
}
},
{
"attributes": {},
"initiator_secret": {
"secret": "haxecliiSgmB"
},
"volumes": [],
"target_secret": {
"secret": "haxecliLUuvO9s"
},
"status": "active",
"account_id": 2405,
"username": "haxecliFV8QdeT6fn5DxtvFuYzjsFwWtc1YzXfT5-NQE5pHiQAQBelNqVskTsJY8",
"storage_container_id": {
"hex": "00000000000000000000000000000000"
}
},
{
"attributes": {},
"initiator_secret": {
"secret": "haxecli7tOAtk"
},
"volumes": [],
"target_secret": {
"secret": "haxecliXFty4F2"
},
"status": "active",
"account_id": 2406,
"username": "haxeclij0yA7YVfCDiq9jZXdkdiKfkSytK2flKk9Gi9NFq0677Fcg44QIDc9inqF",
"storage_container_id": {
"hex": "00000000000000000000000000000000"
}
}
]
}
The pickle format is the same as the JSON format except with an extra field for every object. This field names the object type. This is useful if you want to store the data and use it with a Python SDK later. If you unpickle the string using the Python SDK, you get a full object model instead of a nested dictionary. To get this output, use the -k option after "sfcli" in your command.
See the following example:
{
"py/object": "solidfire.models.ListAccountsResult",
"accounts": [
{
"account_id": 2404,
"attributes": {},
"status": "active",
"py/object": "solidfire.models.Account",
"target_secret": {
"secret": "aAe6Bb&q]63zU0Ei",
"py/object": "solidfire.custom.models.CHAPSecret"
},
"initiator_secret": {
"secret": "3,gG[sP02V'@911}",
"py/object": "solidfire.custom.models.CHAPSecret"
},
"username": "EXAMPLE",
"volumes": [
4588
],
"storage_container_id": {
"hex": "00000000000000000000000000000000",
"py/object": "uuid.UUID"
}
},
{
"account_id": 2405,
"attributes": {},
"status": "active",
"py/object": "solidfire.models.Account",
"target_secret": {
"secret": "haxecliLUuvO9s",
"py/object": "solidfire.custom.models.CHAPSecret"
},
"initiator_secret": {
"secret": "haxecliiSgmB",
"py/object": "solidfire.custom.models.CHAPSecret"
},
"username": "haxecliFV8QdeT6fn5DxtvFuYzjsFwWtc1YzXfT5-NQE5pHiQAQBelNqVskTsJY8",
"volumes": [],
"storage_container_id": {
"hex": "00000000000000000000000000000000",
"py/object": "uuid.UUID"
}
},
{
"account_id": 2406,
"attributes": {},
"status": "active",
"py/object": "solidfire.models.Account",
"target_secret": {
"secret": "haxecliXFty4F2",
"py/object": "solidfire.custom.models.CHAPSecret"
},
"initiator_secret": {
"secret": "haxecli7tOAtk",
"py/object": "solidfire.custom.models.CHAPSecret"
},
"username": "haxeclij0yA7YVfCDiq9jZXdkdiKfkSytK2flKk9Gi9NFq0677Fcg44QIDc9inqF",
"volumes": [],
"storage_container_id": {
"hex": "00000000000000000000000000000000",
"py/object": "uuid.UUID"
}
}
]
}
You can choose from four different levels of logging:
- 0: Critical logging. Only shows errors.
- 1: Warning logging. Shows errors and warnings.
- 2: Info logging. Shows errors, warnings, and information.
- 3: Debug logging. Shows errors, warning, information, and debug information.
To set the debug level, use the --debug option after "sfcli" in your command. The logger is a standard, Pythonic logger. In bash, the standard way to store information from different streams is to specify the stream you wish to redirect to a folder.
In the following example, when you run:
sfcli --debug 2 account list
You get the following information printed:
INFO in cmd_account.py@50: startaccountid = None;limit = None;
And you get the following data:
accounts:
If you want to store the response from the server, you can run the command as follows:
sfcli --debug 2 account list > data.txt
If you want to store the information from the server, you can run the command as follows:
sfcli --debug 2 account list 2> info.txt
In this case, the info.txt file only contains the text labeled "INFO" and the data.txt file only contains the text labeled "accounts:".
You can use the following options in the tool:
-m, --mvip TEXT
SolidFire MVIP.
-u, --username TEXT
SolidFire Cluster username.
-p, --password TEXT
SolidFire cluster password.
-v, --version TEXT
The version you would like to connect on.
-q, --port INTEGER
The port number on which you wish to connect.
-n, --name TEXT
The name of the connection you wish to use in connections.csv. You can use this if you have previously stored away a connection with ‘sfcli connection push’.
-c, --connectionIndex INTEGER
The index of the connection you wish to use in connections.csv. You can use this if you have previously stored away a connection with ‘sfcli connection push’.
-s, --verifyssl
Enable this to check ssl connection for errors especially when using a hostname. It is invalid to set this to true when using an IP address in the target.
-t, --timeout INTEGER
The request timeout in seconds.
-j, --json
To print the full output in json format, use this flag.
-k, --pickle
To print the full output in a pickled json format, use this flag.
--debug [0|1|2|3]
Set the debug level
--nocache
If you do not wish to cache the connection, supply this flag.
--help
Show this message and exit.
Here are the known issues. This list will be added to while they are open. Known issues will be removed from here if/when are fixed.
- Cannot pass blank strings for schedule command parameters (or any other commands, actually).
- If you need to blank out a value, it is currently impossible to pass blank strings via sfcli. Instead, you will need to either a) use a character that represents a blank (eg. _ or *) or b) use one of the other Host Integraion or User Interface tools.
- NEVER use
sudotopip install solidfire-cli! Here is a good explanation of why that is a bad idea.- Anyone on Linux should expect to use a virtual env.
- Windows and Mac users can get away without virtual envs.