laimaretto / taskAutom

simple task automation tool

README

This idea was born because of a need for a simple tool to automate the execution of simple configuration templates on Nokia SROS based routers. Data provided on a DATA file (CSV or Excel) and the configuration templates written in pure Python. Configuration scripts are the result of these templates being rendered with the data.

---

Setup

System Libraries

These libraries have been tested under Ubuntu 20.04 and Windows10 with Python3.8.

Ubuntu

pip3 install taskAutom

Windows

For Windows users, make sure you have Python and PIP installed.

py -m pip install taskAutom

Edit servers.yml

This file has configuration parameters for the Jump Host(s). Add as many as needed. If more than one jump-host is declared, the connections will be load balanced sequentially among them. You can comment out some servers, if needed.

srvr0:
    name: 'myServer'
    user: 'myUser'
    password: 'myPass'
    ip: '1.1.1.1'
    port: 22
srvr1:
    name: 'myServer'
    user: 'myUser'
    password: 'myPass'
    ip: '2.2.2.2'
    port: 22    

---

Bulk Script Execution on routers

The program needs two mandatory inputs: a) DATA file and b) a plugin, which is nothing but a configuration template.

DATA file

The DATA can be either a CSV file or an Excel file. In both cases you can define a header with column names, or not; it's optional.

• If no headers, the file must have in its first column (_1), the IP of the routers to which taskAutom will connect to.
  • Using a data file without headers is discouraged. It will be deprecated. It's only here for historical reasons.
• If using headers, there must be a column named ip with the IP addresses of the routers to which taskAutom will connect to.
  • You can chose a different column name by using the configuration option -gc/--dataGroupColumn myColName.

The first column _1 or the ip column (or eventually changed by -gc) allows taskAutom to group routers by that column when processing data. This is particularly useful if you have the same router along several rows in the DATA file.

If you want taskAutom not to group routers by the [ip|_1] column, you should use the -so/--strictOrder yes CLI parameter: this will process the routers' data in the order of the DATA file as is.

The next columns in the DATA file, are the variables that will be used in the configuration template.

Example: this is a CSV for two different routers, including the data to modify their interfaces. A header is being used in this case.

ip,name,port,interName,ipAddress
10.0.0.1,router1,1/1/1,inter1,192.168.0.1/30
10.0.0.2,router2,1/3/5,inter7,192.168.2.1/30

Plugin

The plugin is a Python code which is fed with each row of the DATA file at a time, in order to render a configuration script. It consists of a function called construir_cliLine() which accepts four arguments:

• m which is a counter, representing the row_id of the data.
• datos which is a Pandas object; the data itself.
• lenData which is the length of the Pandas dataFrame; i.e.: the amount of rows inside the grouped data.
• mop, a boolean.

m and lenData can be used to decide when some part of the code must be ran. mop is used when the configuration script needs to be verified before running; mop=True when the CLI parameter -j/--jobType is 0.

Example: use the previous data, to generate configuration scripts. The example is assuming no header has been defined in the DATA file, so column id is used to identify the proper variable.

def construir_cliLine(m, datos, lenData, mop=None):

    ipSystem   = datos.ip
    router     = datos.name
    port       = datos.port
    intName    = datos.interName
    address    = datos.ipAddress

    cfg        = ""

    if mop and m == 0:
        cfg = "\nHeading_2:Router: " + router + ", " + ipSystem + "\n"

    cfg = cfg + f'/configure router interface {intName} port {port}\n'
    cfg = cfg + f'/configure router interface {intName} address {address}\n'

    if m == lenData-1:
        cfg = cfg + f'/configure router interface {intName} no shutdown\n'

    return cfg

Notes on plugin

1. When writing plugins, do not to use abbreviated commands. This will potentially led to errors. For example: /configure rout int system add 10.0.0.1/32 is discouraged. Better off use /configure router interface system address 10.0.0.1/32.

2. Common practice: it is better to try to accommodate plugins so that they reflect they purpose. Then use the configuration parameter --pluginType=[show|config] to reflect the spirit of the plugin.

3. In general, use --cmdVerify=yes. Only disable cmdVerify if facing problems.

Inventory

By default, taskAutom connects to each and every router that exists inside the DATA data file. Optionally, an inventory file can be provided, with per router connection parameters. If so, the default connection values are overridden by those inside the inventory file.

ip	username	password	useSSHTunnel	readTimeOut	deviceType	jumpHost
10.0.0.1	user1	pass1	yes	15	nokia_sros	server1
10.0.0.2	user2	pass2	no	90	nokia_sros_telnet
10.0.0.3	user3	pass3	no	90	nokia_srl

If fieds in the inventory CSV file are left empty, default values are used.

MOP

When writing a plugin, is important to help taskAutom understand which string should be considered as a title. You do so be adding a prefix Heading_2 to the tiltle variable, under the if mop: statement. After this, a MOP is created with the intended information. There is also the possibility of using the prefix Heading_3.

Result

If taskAutom is invoked with option -j/--jobType 0, a text file with the rendered output, will be genereated.

$ taskAutom -d example/example.csv -py example/example.py -l test -j 0

Router: router1, 10.0.0.1
/configure router interface inter1 port 1/1/1
/configure router interface inter1 address 192.168.0.1

Router: router2, 10.0.0.2
/configure router interface inter7 port 1/3/5
/configure router interface inter7 address 192.168.2.1

Otherwise, if taskAutom is invoked with option -j/--jobType 2, it will connect to each and every router, and execute the commands. User and password must be provided in this case.

---

Bulk SCP/SFTP transfer to routers

There is also the possibility of bulk transfer of files to several routers. To do so you need to to the following:

• use jobtype=3
• prepare a dataFile with the following three column names: ip|ftpLocalFilename|ftpRemoteFilename.

Next, taskAutom will connect to each and every router, transfering the ftpLocalFilename and creating on the remote device a file with the name ftpRemoteFilename.