Hi all
We are running thin on a production server. thin is the main server that responds to all requests coming into our production servers. So the problem is ever since an app went live. thin freezes everything on high load. I am lost as I am not a ruby person. so let me describe the situation.
1. Mobile devices hit our servers at socket 443, https
2. Ruby then redirects the request to glassfish servers. there is a 30 sec time out on thin redirect code.
3. Glassfish does processing and sends results back.
Now out thin script looks like this
thin -s 1 -threaded -C $dir/thin/config.yml -R $dir/thin/
config.ru $command --ssl-disable-verify --max-persistent-conns 1500 --no-epoll
thin -s 5 -threaded -C $dir/thin/config-ssl.yml -R $dir/thin/
config.ru $command --ssl-disable-verify --threadpool-size 1500 --no-epoll
the yml files look like
environment: production
chdir: /var/www
# address: 176.34.62.228
address: 0.0.0.0
user: root
group: root
port: 443
pid: /var/www/thin/thin.pid
log: /var/www/logs/thin.log
timeout: 30
servers: 10
threaded: true
no-epoll: true
daemonize: true
ssl: true
ssl-key-file: /var/www/thin/cert/server.key
ssl-cert-file: /var/www/thin/cert/server.crt
ssl-verify: false
Any help will be helpful. As our servers run fine when there is low to medium load. but on high load. thin seems to crash.
in log files I have noticed
/usr/local/rvm/gems/ruby-2.0.0-p247/gems/eventmachine-1.2.3/lib/eventmachine.rb:1076:in `block in spawn_threadpool'
Do let me know how to go around it.
regards
anindya