Is this issue with indexfile, or search, or index?

[jason0]

Hello

I am perplexed: when I run firebrigade, and choose "detail | index detail" and having chosen a host, my index list is incomplete.

I see in the source where it gets input on line 20: "inputlookup fb_hostname_index_cache".

I see the contents of the fb_hostname_index_cache.csv file is the same incomplete list

I found the periodic search where it extracts the data to put into "fb_hostname_index_cache.csv", the search command being:

 index=summary search_name="DB inspection" | dedup orig_host, orig_index | sort orig_host, orig_index | table orig_host,orig_index | outputlookup fb_hostname_index_cache

When I run this search, I get "Error in 'outputlookup' command: The lookup table 'fb_hostname_index_cache' is invalid."

when I run the search without "| outputlookup fb_hostname_index_cache", I get an incomplete list of my indexes.

so a few things might be happening that I don't know how to determine.

1. splunk doesn't like something about the "fb_hostname_index_cache.csv" file
2. not all of the indexes are being returned from the query
3. something isn't right about the contents of "index=summary"

This issue appeared shortly after i upgraded from 8.0.3 to 8.1.6 to 9.0.0.1.

My current system has six dedicated indexers, and an independent search head.  there are also three heavy forwarders.

Can someone shed some light on this?  

Thank you!