On Sep 4, 2013, at 10:08 AM, Nikos Sidiropoulos wrote:
Hi Nate
Yes, it is a multiprocess setup. But is it because we have more than one web-servers, handlers or both?
Because if it's just the web-servers I could just scale it down to one since it doesn't seem necessary with our number of users. It's would be nice not to have to restart even for the tool-shed tools.
The problem in this case is that one of the web server processes loaded the shed tool but none of the handlers did. Scaling back to a single web process won't solve the problem since the handler(s) will still need to load the tool. Unfortunately you'll have to restart all processes (except, technically, the one web process which happened to install the tool). The only way to avoid the restart is to have a single process for all tasks (web serving and job handling), which comes at a performance penalty. --nate
Bests, Nikos
2013/9/4 Nate Coraor <nate@bx.psu.edu> On Sep 4, 2013, at 9:17 AM, Nikos Sidiropoulos wrote:
Hi Bjørn
that does not look like a bismark error. Is it happen with other tools as well?
No, I have only experienced it with Bismark.
Not sure, sorry. But you should migrate to the new job configuration, better sooner than later.
I am starting to get that idea. There are jobs still running on the server. When they're done I'll migrate it and get back to you if the problem is fixed or not.
Hi Nikos,
The old-style configuration is still supported for now. I suspect this may be a problem of the tool not loading in the handler application. Are you running a multiprocess Galaxy setup? If so, you will need to restart all Galaxy processes after installing tools from the tool shed.
--nate
Bests, Nikos ___________________________________________________________ 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/