scrape_env.sh

[johnbent]

It would be really nice if there was a companion script to io500.sh called 'scrape_env.sh'.
It could take the path of the output dir and do the easy stuff like 'df' and 'mount' and 'statfs' on that dir. Then it would have a big if/else section for each particular file system. For example, if the mount identified it as 'lustre' then it would do a bunch of 'lfs_*' commands to get Lustre specific environmentals. Then slowly the different file systems would fill in their branches as well.

This would be a great way to collect info such as:

1. resilient or not (degree of parity protection)
2. asynchronous resilience or synchronous
3. transient namespace or permanent
4. shared namespace or private
5. stripe layout

[johnbent]

Here is something from LANL that could be used as a starting point:
https://github.com/SC-Tech-Program/Author-Kit/blob/master/collect_environment.sh
But it is generic to the compute and doesn't dig into the storage. It's pretty simple.