Reconnecting to Mongo after connection drop

[tokajuhani]

output plugin does not reconnecting to Mongo if connection is dropped.
Retry logic stuck in a loop. It will never retry to get connection to Mongo.

Howto reproduce:

1. Start logstash with a Mongo output plugin
2. Write something to logstash --> Data is stored to Mongo
3. Shutdown Mongo (I run Mongo as Service in Windows, so I stop Mongo service)
4. Write something to logstash
5. I get error: Failed to send event to MongoDB {:event=>2017-03-08T08:48:00.057Z _________ t4, :exception=>#<Mongo::Error::NoServerAvailable: No server is available matching preference: #<Mongo::ServerSelector::Primary:0x7354 @tag_sets=[], @options=:database=>"admin"},@server_selection_timeout=30>>,
6. Start Mongo service
7. Same error occurs again and again

I have no experience of Ruby.
but if this is called only once:

def register
   Mongo::Logger.logger = @logger
   conn = Mongo::Client.new(@uri)
   @db = conn.use(@database)
 end # def register

Maybe conn = Mongo::Client.new(@uri) and @db = conn.use(@database) should be called if connection is dropped in def receive(event) function.

[berlic]

There's a PR #29 to fix that. Why this patch is not applied? @jordansissel

[jordansissel]

Why this patch is not applied

If I had to guess, it's because there are two failures in that PR:

• ❌ CLA — Commit author has not signed the CLA and is not a member of Elasticsearch
• ❌ continuous-integration/travis-ci/pr — The Travis CI build failed

[colinsurprenant]

relates to #46

[colinsurprenant]

commented on proposed fix #29