-- Main.simonk - 22 Dec 2005

Wish lists

General wish list

Condor wish list

List improvements to the way Condor works locally that you would like to see.

  • (Simon King) Wake on LAN - for vacations, when machines are powered down. This is probably technically possible with Condor, but does DICE support WOL? We're currently looking at this as part of a power management project. -- IainRae - 9 Dec 2008

GridEngine wish list

  • (Partha Lal) Could we view ganglia from outside Informatics?
    • This is possible by tunneling an ssh connection - see the main GridEngine wiki page

  • (Partha Lal) It'd be helpful to be able to mount more drives. Some directories, e.g. /group/project/ami2/ are visible to Grid Engine jobs on townhill but others e.g. /group/cstr/projects/phd/s0565860/ aren't. -- Main.s0565860 - 04 Apr 2006 TBH I'm not sure why these directories are visible at all, they were set up for a week or so as a courtesy to one user when I was testing townhill during commissioning and the exports were removed from the server (hippocampus) a while ago. Generally the only network filesystems we make available to the nodes are /home and /group/beowulf mounted on har which is connected to the same switch as most of the nodes. /group/beowulf is also available on any DICE machine. The filesystem mentioned is mounted from hippocampus which is at BP so any traffic is going to travel across EDLAN. If there were jobs running on all 64 the lion nodes reading a 10Gb file then you'd be generating 640Gb of traffic across EDLAN and impacting on other users. You'd also take a performance hit, it'd be faster to pull the files off of har, and if you're doing multiple reads or writes to large files it'll be much fasted to copy them into /disk/scratch on the nodes. Finally if hippocampus were to go down or any reason for an extended time any nfs mounts which were written to would go stale and it's usually only possible to clear stale mounts by rebooting the client machines. Having just one fileserver minimises the risk of this happening. -- IainRae - 07 Apr 2006

  • Most of the above is in the FAQ, under the "What happened to my home directory?" section. I'll make it more explicit on Monday. -- IainRae - 07 Apr 2006


Below are two sections for me (Simon King) to gather your feedback on experiences with Condor and GridEngine. Please comment on what aspects of the configuration are working well, as well as things that could be improved. For Condor, I am especially interested in the impact Condor has on your desktop machine, and what configuration tweaks we need to make to improve this, so that we can roll out Condor on the desktop machine of every unsuspecting person in Informatics smile

2008 feedback on Condor

Please enter your comments in this section - and place your name against each comment.

2008 feedback on GridEngine

Please enter your comments in this section - and place your name against each comment.

Topic revision: r21 - 09 Dec 2008 - 12:06:13 - IainRae
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies