Running Galaxy on PBS/Torque cluster
I have had Galaxy running on a local BAS, and after success with that we're ready to create an instance on our local cluster computing service. The local cluster uses PBS/Torque. I'm familiar with the basics of configuration, but our particular scheduler rewards those who specify accurate walltimes. I would like to be able to define specific walltimes for each tool. I'm mapping out a strategy for how best to specify accurate walltimes. For instance, if someone is running tophat, it will take longer if they're mapping against a complex reference genome like human, as opposed to nematode. I'd like to use different walltimes to account for that. Is there anyone who has tackled this and would be willing to share their approach with me? Thanks! Dan
On May 3, 2012, at 3:42 PM, Dorset, Daniel C wrote:
I have had Galaxy running on a local BAS, and after success with that we’re ready to create an instance on our local cluster computing service. The local cluster uses PBS/Torque.
I’m familiar with the basics of configuration, but our particular scheduler rewards those who specify accurate walltimes. I would like to be able to define specific walltimes for each tool. I’m mapping out a strategy for how best to specify accurate walltimes. For instance, if someone is running tophat, it will take longer if they’re mapping against a complex reference genome like human, as opposed to nematode. I’d like to use different walltimes to account for that.
Is there anyone who has tackled this and would be willing to share their approach with me?
Hi Dan, This doesn't currently exist in Galaxy, but John Chilton at UMN has written enhancements that do what you're looking for: https://bitbucket.org/galaxy/galaxy-central/pull-request/12/dynamic-job-runn... Unfortunately the patches no longer apply cleanly against galaxy-central, but I will work on getting them integrated soon. --nate
Thanks!
Dan ___________________________________________________________ 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:
On May 10, 2012, at 11:02 AM, Nate Coraor wrote:
On May 3, 2012, at 3:42 PM, Dorset, Daniel C wrote:
I have had Galaxy running on a local BAS, and after success with that we’re ready to create an instance on our local cluster computing service. The local cluster uses PBS/Torque.
I’m familiar with the basics of configuration, but our particular scheduler rewards those who specify accurate walltimes. I would like to be able to define specific walltimes for each tool. I’m mapping out a strategy for how best to specify accurate walltimes. For instance, if someone is running tophat, it will take longer if they’re mapping against a complex reference genome like human, as opposed to nematode. I’d like to use different walltimes to account for that.
Is there anyone who has tackled this and would be willing to share their approach with me?
Hi Dan,
This doesn't currently exist in Galaxy, but John Chilton at UMN has written enhancements that do what you're looking for:
https://bitbucket.org/galaxy/galaxy-central/pull-request/12/dynamic-job-runn...
Unfortunately the patches no longer apply cleanly against galaxy-central, but I will work on getting them integrated soon.
--nate
That would be really nice for our local users as well. I think the current hack around this was to have two versions of the same tools, configured differently, which is definitely suboptimal. chris
Thanks Nate! I'm looking forward to it! -----Original Message----- From: Nate Coraor [mailto:nate@bx.psu.edu] Sent: Thursday, May 10, 2012 11:02 AM To: Dorset, Daniel C Cc: galaxy-dev@lists.bx.psu.edu Subject: Re: [galaxy-dev] Running Galaxy on PBS/Torque cluster On May 3, 2012, at 3:42 PM, Dorset, Daniel C wrote:
I have had Galaxy running on a local BAS, and after success with that we're ready to create an instance on our local cluster computing service. The local cluster uses PBS/Torque.
I'm familiar with the basics of configuration, but our particular scheduler rewards those who specify accurate walltimes. I would like to be able to define specific walltimes for each tool. I'm mapping out a strategy for how best to specify accurate walltimes. For instance, if someone is running tophat, it will take longer if they're mapping against a complex reference genome like human, as opposed to nematode. I'd like to use different walltimes to account for that.
Is there anyone who has tackled this and would be willing to share their approach with me?
Hi Dan, This doesn't currently exist in Galaxy, but John Chilton at UMN has written enhancements that do what you're looking for: https://bitbucket.org/galaxy/galaxy-central/pull-request/12/dynamic-job-runn... Unfortunately the patches no longer apply cleanly against galaxy-central, but I will work on getting them integrated soon. --nate
Thanks!
Dan ___________________________________________________________ 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:
participants (3)
-
Dorset, Daniel C -
Fields, Christopher J -
Nate Coraor