Hello,
I would like to add a new tool to my local Galaxy install but I fail. I have put the script and the xml description file in the tools/ directory and modified the tool_conf.xml file. I type sh run.sh and the end of the message on the screen is :
socket.error: (98, 'Address already in use') galaxy.jobs INFO 2010-05-07 11:19:20,397 sending stop signal to worker thread galaxy.jobs INFO 2010-05-07 11:19:20,397 job queue stopped galaxy.jobs.runners.local INFO 2010-05-07 11:19:20,397 sending stop signal to worker threads galaxy.jobs.runners.local INFO 2010-05-07 11:19:20,398 local job runner stopped galaxy.jobs INFO 2010-05-07 11:19:20,398 sending stop signal to worker thread galaxy.jobs INFO 2010-05-07 11:19:20,398 job stopper stopped
and the new tool was not available on the refreshed page of the browser.
The script being already running , I thought that could be problem (Am I right ?) and tried "run.sh --stop-daemon", which returned me "No PID file exists in paster.pid".
What is the right thing to do ?
Thanks for your help,
Loraine
Hi Loraine,
It looks like you have a copy of Galaxy already running on that port. If you are just using the command 'sh run.sh' to run Galaxy, it is not being run as a daemon, so that '--stop-daemon' will not stop the already running server. You can use 'ctrl-c' to stop the server that you have started within a console. If you don't have access to the console view of this (e.g. you did a 'sh run.sh &'), you can do:
$ ps x | grep "python ./scripts/paster.py serve universe_wsgi.ini" 13294 s001 S+ 14:25.14 python ./scripts/paster.py serve universe_wsgi.ini 20592 s003 S+ 0:00.00 grep python ./scripts/paster.py serve universe_wsgi.ini
and then kill the process directly (e.g. 'kill 13294')
To run Galaxy as a daemon start it like 'sh run.sh --daemon', then you can use 'sh run.sh --stop-daemon' to end it later.
Thanks for using Galaxy,
Dan
On May 7, 2010, at 5:31 AM, Loraine Brillet wrote:
Hello,
I would like to add a new tool to my local Galaxy install but I fail. I have put the script and the xml description file in the tools/ directory and modified the tool_conf.xml file. I type sh run.sh and the end of the message on the screen is : socket.error: (98, 'Address already in use') galaxy.jobs INFO 2010-05-07 11:19:20,397 sending stop signal to worker thread galaxy.jobs INFO 2010-05-07 11:19:20,397 job queue stopped galaxy.jobs.runners.local INFO 2010-05-07 11:19:20,397 sending stop signal to worker threads galaxy.jobs.runners.local INFO 2010-05-07 11:19:20,398 local job runner stopped galaxy.jobs INFO 2010-05-07 11:19:20,398 sending stop signal to worker thread galaxy.jobs INFO 2010-05-07 11:19:20,398 job stopper stopped
and the new tool was not available on the refreshed page of the browser.
The script being already running , I thought that could be problem (Am I right ?) and tried "run.sh --stop-daemon", which returned me "No PID file exists in paster.pid".
What is the right thing to do ?
Thanks for your help,
Loraine
-- Loraine Brillet
Service Informatique et Genomique Plateforme ABiMS (Analyses and Bioinformatics for Marine Sciences) Station Biologique de Roscoff FR 2424 CNRS UPMC Place Georges Teissier 29682 Roscoff Cedex tel: ++33 (0)2 98 29 25 50 fax: ++33 (0)2 98 29 23 24 email: loraine.brillet@sb-roscoff.fr http://www.sb-roscoff.fr http://www.sb-roscoff.fr/abims/
galaxy-dev mailing list galaxy-dev@lists.bx.psu.edu http://lists.bx.psu.edu/listinfo/galaxy-dev
On Fri, May 07, 2010 at 11:31:07AM +0200, Loraine Brillet wrote:
The script being already running , I thought that could be problem (Am I right ?) and tried "run.sh --stop-daemon", which returned me "No PID file exists in paster.pid".
What is the right thing to do ?
Your assessment that another galaxy instance is already running seems correct. Off the top of my head here are a few possible reasons why --stop-daemon might not be working:
1) If the running instance were started in another disk location/copy-of-galaxy you need to be there when stopping the daemon
2) If it was run not as a daemon, (eg "./run.sh") then it's not a daemon and can't be stopped thusly. If that's the case go find the window where it was started and press Ctrl-C in it.
3) If the daemon were started by a user other than you it's possible you're unable to stop their daemon instance and read their .pid file
Some options you have now inlcude:
1) find the python galaxy process and kill it directly:
$ ps auxwww | grep python galaxy 3743 0.0 2.3 332716 72532 ? Sl May05 2:30 python -E ./scripts/paster.py serve --log-file /var/log/galaxy/galaxy.log universe_wsgi.ini --pid-file=/var/lib/galaxy/galaxy.pid --daemon $ kill 3743
or
2) Find out what user+location combo ran the daemon and stop it
or
3) Start the new instance on a different by changing the port= value in universe.wsgi to something ohter than 8080, and update the URL in your browser's address bar thusly.
Sorry if that's too much info, but the cause and fix are likely somewhere in that braindump.
galaxy-dev@lists.galaxyproject.org
-
Daniel Blankenberg -
Loraine Brillet -
Ry4an Brase