Corrupt Index Exception Guidance


(Rohail Kidwai) #1

I'm looking for guidance on what my options are when I have a corrupt index.

I found this out when trying to snapshot and it failed due to a corrupt index.
For more Info:
https://discuss.elastic.co/t/understanding-index-shard-snapshot-failed-exception/159572/5

I don't know how to tell if it's a replica or primary so the only thing I know right now is the name of the corrupted index.

Also I am not seeing anything standout when I do
GET /INDEX/_stats

It seems a lot of information on corrupt indices seems to be on older ES versions. (I'm on 6.1) So before I follow one of those guides, I am just wondering what/where is the latest information and guidance to resolve this issue?


(Rohail Kidwai) #2

An update to this:

I figured out that the corrupted shard id is 0 just from looking closer at the snapshot error.
Then afterwards doing:

GET _cat/shards/winlog-2018.09.05

I get:

INDEX-2018.09.05 0 r STARTED 71460755 36.7gb IP.ADDRESS NODE-master-3
INDEX-2018.09.05 0 p STARTED 71460755 36.7gb IP.ADDRESS NODE-master-2

I'm assuming the r and p stand for replica and primary. I want to delete the primary in hopes that maybe only that is corrupt. Can someone point me to the correct api for this?