Skip to content

TREE.md

Latest commit

 

History

History
67 lines (51 loc) · 4.09 KB

TREE.md

File metadata and controls

67 lines (51 loc) · 4.09 KB

image

Introduction

F5 configuration has a very complicate multiple-layer structure. The best way to think about that, it's an upside-down 30-foot big tree, with branches / sub-branches and leaves all over the place. Something like the picture below.

image

What's F5 Configuration Object Dependency Tree Resolving Algorithm

The F5 Configuration Object Dependency Tree Resolving algorithm, is a custom algorithm implemented by Sam Li. The goal is to identify and manage the complicate dependencies between a very large pool of F5 single configuration object efficiently.

The algorithm is based on Graph Theory, and the previous Python implementation from others, such as Ferry Boender.

Algorithm Improvements

In the data structure implementation, I keep the basic node structure. However, the tree structure are different in the following:

  • Build out the whole tree structure by using Python dictionary (vs. list). In theory, this may improve both computing speed - O(1) vs O(n), and the node retrieving flexibility of the algorithm.
  • Build the edges using node name instead of whole node. This simplify the resolving lookup process, and reduce the complexity of tree structure.
  • Simplify 'Circular reference check' by using existing global variable 'resolved' .

In the algorithm implementation, the basic ideas are still the same. However, there are many additions unique in our case:

  • Extending the F5Client class;
  • Build out the whole tree structure;
  • Automating the tree node loading process;
  • Additional sanity checks and better error handlings.

Overall, with an improvement of data structures and other implementation advantages above, there comes up with a more mature and efficient algorithm.

More Read

You can find better explanation of the concepts from Ferry Boender work.

Example 1 - Resolving LTM Configuration Dependency

Assuming you have this python package installed, and have the access to the F5 configuration file sample. Copy the following code block in a text editor. Then save it into a file name 'test3.py'. Or Download it as raw file.

# !/usr/bin/env python              # F5DepTree Test Example
import f5_admin                     # First we need to let Python know we're going to use the API
f5_node = "xx-dl4prd1"           # The F5 node we want to connect to.
with f5_admin.F5DepTree(None,None,None) as client:
    client.load(f5_node)
    conf_obj_name = "ltm virtual Forward_App_Servers"
    result = client.dep_resolve(conf_obj_name)
    print "Resolving configuration dependency on ", f5_node, 'for object:', conf_obj_name
    print result

Then you should be able to run the following code to resolve the configuration dependency for a LTM object 'Forward_App_Servers':

$ python src/bin/test3.py
Loading cache_config:  /Library/Python/2.7/site-packages/f5_admin/conf/xx-dl4prd1/xx-dl4prd1.txt
Loading complete
Resolving configuration dependency on  xx-dl4prd1 for object: ltm virtual Forward_App_Servers
['security log profile splunk-logging_local', 'security log profile splunk-logging', 'ltm profile fastl4 ftps_fastL4', 'net vlan DL4-prd-int-vlan752']

Example 2 - Resolving AFM configuration Dependency

TBD