File not closed with JDBC input and CSV output

[romariol (romario)]

Hi,

I'm using a JDBC input to fetch data from a specific table. I want to put this data in a new file. For this I'm using the CSV output. What I'm seeing, is that the file remains open. I want this file to be closed, because after the creation of the CSV file, I want to place it into a different endpoint (like a blob storage). Is there a way to do this? My conf looks like this:

input {
  jdbc {
    jdbc_driver_library => "/driver/mssql-jdbc-7.2.2.jre8.jar"
    jdbc_driver_class => "com.microsoft.sqlserver.jdbc.SQLServerDriver"
    jdbc_connection_string => "jdbc:sqlserver://xxxxx\xxxx;database=xxxx;user=xxxxx;	password=xxxxxx"
jdbc_user => nil
schedule => "16 15 * * *"
statement => "SELECT * from [dbo].[yyyyyy]
    }
}

filter {

mutate {
	add_field => {"filename" => "%{+YYYY-MM-dd}" }
}

}

output {
csv {
    path => "zzz-yyyy-%{filename}.csv"
	fields => ["@timestamp","seq_num","delete_session_id","create_user"]
}
}

Thanks

[elasticforme (Sachin)]

This is something different requirement.
you can put output from mysql straight to csv file. why use logstash.

anyway how do you know that this file remains open?

[romariol (romario)]

I know, but we want to maintain this ingesting via Logstash and not via MSSQL. What I see is if I run a seperate script, the file is still locked.

[elasticforme (Sachin)]

I don't see any option for close in output-csv section.

https://www.elastic.co/guide/en/logstash/current/plugins-outputs-csv.html#plugins-outputs-csv-csv_options

[romariol (romario)]

No exactly. What I do see now is that the file closes based on the scheduler cron. So lets say you've set it to "* * * * *", the file will close after each next write actions. If you would run the cron every day, that would mean that the file remains open for an extremely long time.