Problem

Currently, when the application runs and is interrupted, there is no code to gracefully exit execution. Especially when scanning with subprocesses, as they have a tenancy to stick around after the interrupt.

Objective

Implement support for interrupts.

Ideas

• Using the signal module to get the subprocesses to ignore the KeyboardInterrupt, then terminating the workers normally

• Come up with a save format

• Find out how to retrieve the data too

Please let me know if you'd like to incorporate support for JUNOS devices in the future, I would be happy to help out.

Hello, I'm using python 3.6 with inside a virtualenv and I can't get information to display whatever the platform (cisco_nxos, cisco_ios or cisco_xe).

python run.py -sS -t 172.28.46.253 -p cisco_nxos
main.main : #3 Start new run
main.main : #3 ##### Starting Single Run #####
main.single_run : #3 Processing connection to 172.28.46.253
base_device.process_devices : #4 Processing device
cli.connect : ? SSH auth error to 172.28.46.253 using admin, In
cli.connect : ? SSH auth error to 172.28.46.253 using admin, In
cli.connect : #4 Successful ssh auth to 172.28.46.253 using administrator, In
base_device.process_devices : ? _enable - Error: 'CiscoNxosSSH' object has no attribute 'ip'
main.single_run : ? Connection to 172.28.46.253 failed: 'CiscoNxosSSH' object has no attribute 'ip'
Device processing failed
main.main : #3 ##### Single Run Complete #####

Could you please help to point what is wrong?

Objective

All parts of the program need more tests. Optimal test coverage should reach over 80%.

Concerns

A lot of the application depends on a populated database or a connection to a live device. How do we test without a live network to test on? Maybe through pre-downloaded samples of device output?

Problem

Netcrawl currently polls devices properly, but once the data is in the database there's no convenient way to interact with it outside of an SQL query

Objective

Implement a web-based front end to allow users to interact with the polled data. Obviously, there are a lot of ways to do this, but I'm not a web developer and I am open to ideas and help.

Objective

For testing purposes, we should have fully populated structures to work with. For this, we need helper functions which create on-demand, populated interfaces and devices, as well as filling the inventory and main databases with fake data.

When appropriate, these fake structures can be used alongside pre-made ones in situations where exactly known data is not needed.

Possible Implementation

Using the Faker package will expedite this

• Find out what format Solarwinds requires

• Output an importable file containing all detected devices

Please add new device types!

Each device will save it's mac address table per switchport. The result is that both the directly connected device and all devices in the same layer two path storing an end device's MAC address.

The MAC Audit module will take a csv (or any other input, as added in the future) and to compare it with the database of polled mac addresses. The purpose is for things like Rogue AP audits, where a rogue device may be detected, its MAC polled, and then compared against macs connected to switchports for a similar OUI.

For each csv mac address:

1. Get a list of all polled MAC addresses in the given subnet
2. Compare the csv MAC against the polled MACS
3. Generate a confidence rating that indicates how similar the MAC's are

The csv needs at minimum the mac column as well as either a network_ip column or an ip and subnet column.

Challenges:

• Currently, MAC addresses are stored by interface. The interface will not always have an associated subnet, due to being a layer 2 interface. The module should intelligently use the device's subnets when an interface specific one isn't available.

The setup scripts need to be updated.

Objective

When performing a scan with a pre-existing database, the application will identify pre-existing devices and update them rather than adding duplicate entries.

Possible Implementation

After polling a device, search the database for a duplicate unique_name.

If found:

• Full Replace (preferred): Cascade delete the old entry and add the new one.
• Update: Add the device_id of the match to the polled device. Compare each entry and update them.

Concerns

The application may be interested in tracking changes in device state over time. What would be the best way to do that?

I think there are extra SQL calls in there that can be very seriously trimmed down. Low priority.