-
Notifications
You must be signed in to change notification settings - Fork 3.5k
Description
When I use redis as input, it fails to stop using service logstash stop (which sends kill -TERM $PID).
When debugging, here is what I get when sending SIGTERM to the process (not started with service):
SIGTERM received. Shutting down the pipeline. {:level=>:warn, :file=>"logstash/agent.rb", :line=>"127", :method=>"execute"}
Sending shutdown signal to input thread {:thread=>#<Thread:0x3c8c03ae run>, :level=>:info, :file=>"logstash/pipeline.rb", :line=>"249", :method=>"shutdown"}
And it waits forever, doing nothing. I can send more SIGTERM and more of those appear on the screen.
Additionally, I noticed the following a few lines above, when it was logging and I don't know if it's related or not:
Starting stale files cleanup cycle {:files=>{"/var/log/logstash/output.log"=>#<IOWriter:0x6fcc6d1e @active=true, @io=#<File:/var/log/logstash/output.log>>}, :level=>:info, :file=>"logstash/outputs/file.rb", :line=>"196", :method=>"close_stale_files"}
Here is the first config file:
input {
}
filter {
mutate {
add_field => {
"received_at" => "%{@timestamp}"
"received_from" => "%{host}"
}
}
}
output {
elasticsearch { host => localhost }
file {
path => "/var/log/logstash/output.log"
}
}
I know it could be simplified and some stuff in there do not make much sense but I was testing Logstash (and that's the config I had when the bug happened) and you can reproduce the bug every single time with this config.
Here is the second config file:
input {
redis {
host => "192.168.110.143"
data_type => 'list'
key => 'logstash:redis'
type => 'redis-input'
password => 'howaboutthisverylongpassword?'
}
}
Another side note is that it seems that the redis plugin only works for the first few minutes. I just changed 2 lines in the configuration file for redis (from the package install on Ubuntu 14.04.2 LTS 64 bit). Let me know if you need the redis config file.