Paths, paths everywhere

So I realise I am way more rusty than I had thought when I set avout reinstalling everything from scratch, I guess this is a good time to ask some more silly questions, so here goes:
During the installation of the agent and micro services I had to specify a root path.
After that I had to specify more paths in the web UI for indexing.
Then I will have to set up more paths for the actual file channels.
This is where I am getting a bit confused and it feels like there is some redundancy in teh setup (or I am just not quite not getting it).

So during the agent installation I had to specify a root path, say:


In the web UI I have now set up an indexing path to:


Next I set up a file channel to:


Aren’t all subsequent paths sub paths of the rootpath set during the install? And thus they should be relative paths? If so, could their be a UI enhancements that makes all this a bit more streamlined vy suggesting valid paths for indexing and file channels?


1 Like

Hi Frank,

Aren’t all subsequent paths sub paths of the rootpath set during the install? And thus they should be relative paths?”

It is true all indexers are paths of the data root path. By using the full path, there is no ambiguity for what the user should be entering when creating new indexers or File Channels. You use the full path in all cases.

There are some user benefits to using the full path when using copy/paste of folders.

You are using the web app, where you type these things in manually. If the paths you want to use are available on the system you are using (and have same mount points as location where the files actually are), you can use the native file explorer to select a folder, copy (Ctrl-C) that folder, and then paste into the text box for indexer and File Channels. This will put the proper full path into the textbox in your browser.

Using the desktop app, you can use the native file selector mechanism (like Finder or Explorer or Nautilus) to graphically navigate your file tree and select the folder for indexers and File Channels. It will use the full path and you do not need to type anything.

There is another layer of detail beneath all of this. The mount location of a file on the system where the Yadle agent is running (let’s say /ohufx/lib) may not be the same as the mount point on a user system (perhaps /network/nfs/ohufx/lib). All indexers and File Channel paths must be specified by the full path as it is know by the agent system.

Your idea of using relative paths has merit. I need to think through all of the implications of this when I add the combinations of indexers, Yadle ‘VolumeIDs’, network mapping, etc. As always, you have provided more to think about!


p.s. VolumeIDs are a mechanism we have to allow files to be associated with a physical volume or device in a portable manner. A common scenario is an external drive that gets plugged into multiple systems. Using VolumeIDs, Yadle agents would not re-index the volumes as they more between systems.