[+DOC] Shard Unassigned Reasons

[stefnestor]

Howdy! I like using CAT Shards's unassigned.reason as reference, but just ran into an undocumented one and realized the code includes a couple others as well.

Can you append info on these

• PRIMARY_FAILED
• FORCED_EMPTY_PRIMARY
• MANUAL_ALLOCATION
• INDEX_CLOSED
• NODE_RESTARTING

I'm not sure if you're prefer a literal code copy/paste, but am experiencing MANUAL_ALLOCATION if you have more information on it. Cheers!

[elasticmachine]

Pinging @elastic/es-docs (Team:Docs)

[elasticmachine]

Pinging @elastic/es-distributed (Team:Distributed)

[DaveCTurner]

FWIW I think we should expand the description of this field in the docs too. Today we say

Reason the shard is unassigned.

That's not really true, or at least it's a bit misleading to phrase it like this. The reason the shard is unassigned is often fairly complex and needs someone to use the allocation explain API to investigate. This field describes the event that caused the shard to become unassigned in the first place (edit:) most recently changed the state of this shard in the routing table, which is rarely very interesting.

MANUAL_ALLOCATION just means that someone has hit POST _cluster/reroute?retry_failed to reset the failed-allocation counter.