I am using sidekiq gem to run API calls in background. I ran sidekiq in Daemon process like:
bundle exec sidekiq -d
Now I made some changes in my method, so I want to restart sidekiq. I tried to kill the sidekiq by using the below command:
kill -9 process_id
but it's not working. I want to know the command to restart sidekiq process. If you have any idea please share with me.
I tried the below command also:
sidekiqctl stop /path/to/pid file/pids/sidekiq.pid
So after you find you proces_id, use the below command, that will stop the workers from getting new jobs, and they will finish existing jobs.
kill -USR1 [PROCESS_ID]
After that you can kill them
kill -TERM [PROCESS_ID]
Also there is a page on sidekiq/wiki about this, called Signals.
[edit]
Here is the signal page.
[edit]
Check video
Start:
$ bundle exec sidekiq -d -P tmp/sidekiq.pid -L log/sidekiq.log
where -d demonize, -P pid file, -L log file.
Stop:
$ bundle exec sidekiqctl stop tmp/sidekiq.pid 0
Sidekiq shut down gracefully.
where 0 is number of seconds to wait until Sidekiq exits.
To keep the daemon running you should definitely have some good error handling in the HardWorker classes, but you can also use the command below to restart the sidekiq runners if they are not found in the system processes.
x=`ps aux | grep sidekiq | grep -v grep | awk '{print $2}'`; [ "$x" == "" ] && cd /path/to/deploy && bundle exec sidekiq -d -L /path/to/deploy/log/sidekiq.log -C /path/to/deploy/config/sidekiq.yml -e production
This basically looks for the PID using ps aux | grep sidekiq | grep -v grep | awk '{print $2}' and then stores it in variable x. Then, if it's empty, it will run a daemonized sidekiq process.
You can stick this guy into a cron job or something. But if your jobs are failing continually, you'll definitely want to figure out why.
EDIT: Added path to deploy from cron.
