{"id":5566,"date":"2020-09-21T14:54:09","date_gmt":"2020-09-21T21:54:09","guid":{"rendered":"https:\/\/blogs.infoblox.com\/?p=5566"},"modified":"2020-09-21T14:54:09","modified_gmt":"2020-09-21T21:54:09","slug":"hacking-infoblox-discovery-for-fun-and-profit-its-just-another-csv","status":"publish","type":"post","link":"https:\/\/www.infoblox.com\/blog\/community\/hacking-infoblox-discovery-for-fun-and-profit-its-just-another-csv\/","title":{"rendered":"Hacking Infoblox Discovery for Fun and Profit, It’s Just Another CSV"},"content":{"rendered":"

An IPAM solution is not very useful unless it accurately reflects the state of your network. The discovery features in NIOS have always been one of the great benefits of having Infoblox DDI.\u00a0Network Insight<\/a>\u00a0and\u00a0NetMRI<\/a>\u00a0focus on discovery of network gear and what is connected to it. The\u00a0vDiscovery<\/a>\u00a0feature adds the ability to discover virtual infrastructure in both public and private cloud environments. All of these working together can give you a super accurate picture of what exists on your network\u2026if all your gear is supported.<\/p>\n

As much as it may seem like it, the discovery features are not magic. Infoblox engineers had to figure out how each\u00a0supported vendor\u2019s device<\/a>\u00a0works and how to query its details. New vendors, new models, and new firmware versions can all affect Infoblox\u2019s ability to properly discover your network gear. In many cases, customers are the ones driving Infoblox to make additions to the support list. But what do you do if you have an unsupported device and can\u2019t\u00a0(or don\u2019t want to)<\/em>\u00a0wait for the formal device support request process?<\/p>\n

You could just do it yourself. \ud83d\ude0f<\/p>\n

USING NETMRI\u2019S EXAMPLE<\/h3>\n

Because NetMRI is not part of the NIOS grid, it is a good model for figuring out how to upload our own discovery data from something external. The problem is we don\u2019t really know how it works under the hood. But we have hints from the Audit Log on the grid master.<\/p>\n

\"\"<\/p>\n

Every time NetMRI submits new discovery data, it calls a\u00a0DataUploadInit<\/code>\u00a0function followed by\u00a0SetDiscoveryCSV<\/code>. After searching the WAPI documentation, SetDiscoveryCSV<\/code>\u00a0appears to be a standard WAPI function tied to the\u00a0fileop<\/code> object type. This is great because anything that can use the WAPI via REST can theoretically submit new discovery data. You can view the WAPI doc on your own grid using the following URL and substitue your own grid master host\/IP, https:\/\/gm\/wapidoc\/objects\/fileop.html#upload-setdiscoverycsv<\/code>. Here is a screenshot from my lab GM\u2019s copy.<\/p>\n

\"\"<\/p>\n

Now we have to figure out how to format our CSV properly and we should be good to go.<\/p>\n

SUPPORT BUNDLE TO THE RESCUE<\/h3>\n

When I was first looking into this, I was afraid I\u2019d need to do something like\u00a0MitM<\/a> the connection between NetMRI and the grid master in order to capture the CSV file being uploaded while it was in transit. Thankfully, that was unnecessary due to Infoblox\u2019s support bundle<\/a>\u00a0functionality.<\/p>\n

If you have ever created an Infoblox support ticket, you have likely also been asked to generate and attach a support bundle to the ticket. It\u2019s a tar\/gzipped file containing a bunch of low-level diagnostic stuff from the appliance it was generated on (log files, config files, etc). It is also a great resource for better understanding Infoblox internals or even trying to solve your own support issues if you have the time and patience.<\/p>\n

While poking around in my lab GM\u2019s support bundle, I noticed a folder called\u00a0\/storage\/discovery\/last_csv_import<\/code>\u00a0which looked very promising. It contained another numbered folder likely tied to an internal ID of some sort and within that was a file called\u00a0import_file<\/code>.<\/p>\n

\"\"<\/p>\n

After extracting the file, it appeared to be a complete copy of the file NetMRI had submitted with all column headers intact. \ud83c\udf89 Hooray! \ud83c\udf89<\/p>\n

CSV FORMAT<\/h3>\n

Here is a list of all the headers included in the CSV file we pulled from the support bundle:<\/p>\n