Hi Curtis,

we simply remove our galaxy server from the SGE submit host list. So we can't submit anything to the Queue. That works quite well.

Cheers, Bjoern

Am 01.05.2014 21:17, schrieb Curtis Hendrickson (Campus):

Peter,

Thanks for the quick reply. Saved me a lot of google/grep time.

Here's what I did to work-around (added it to the card as a comment). Anyone got a more thorough/elegant solution?

Regards, Curtis

Workaround:

1. Backup migrated_tools_conf.xml, shed_tool_conf.xml and tool_conf.xml
2. edit each to remove all sections
3. edit static/welcome.html to notify users with a big orange announcement

That gives you an empty tool pane, though it doesn't prevent anyone from hitting "re-run" on an existing dataset, or doing something through the API.

-----Original Message----- From: Peter Cock [mailto:p.j.a.cock@googlemail.com] Sent: Thursday, May 01, 2014 1:03 PM To: Curtis Hendrickson (Campus) Cc: galaxy-dev PSU list-serv (galaxy-dev@lists.bx.psu.edu) Subject: Re: [galaxy-dev] Cleanest way to disable job-runs during a long cluster maintenace window?

Hi Curtis,

This used to be easy - there was a setting on the admin pages to lock new job submission, which I would use before a planned shutdown.

See https://trello.com/c/BTDaHy9m/1269-re-institute-job-lock-feature (and vote on it to help prioritize the issue).

Right now I'm not sure if there is an easy alternative :(

Peter

On Thu, May 1, 2014 at 6:45 PM, Curtis Hendrickson (Campus) curtish@uab.edu wrote:

...

Folks

The cluster under our galaxy server will be down for a WEEK, so it can be relocated.

The fileserver that hosts the galaxy datasets, the galaxy server and the galaxy database will stay up.

What’s the most elegant way to disable people’s ability to run new jobs, without blocking them from browsing existing histories and downloading/viewing their data?

Our configuration:

             External Auth

                             universe_wsgi.ini:use_remote_user =

True

             Job runner: SGE

                             universe_wsgi.ini:start_job_runners =

drmaa

             and the new_file_path and job_working_directories are

on a fileserver that will be down.

grep -H /scratch/share/galaxy universe_wsgi.ini

universe_wsgi.ini:file_path = /scratch/share/galaxy/staging # available

universe_wsgi.ini:new_file_path = /scratch/share/galaxy/temp # down

universe_wsgi.ini:job_working_directory = /scratch/share/galaxy/job_working_directory # down

We’d rather something nicer than just letting the jobs go to the queue and never get run.

Regards,

Curtis

---

Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/

---

Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/