[sage-devel] Re: DSage server: start clean?
On 10/23/07, John Voight [EMAIL PROTECTED] wrote: I think I figured out a brutal way: Just delete the $HOME/.dsage/dsage directory! There must be a better way... I had precisely the same complaint very recently, and Yi added a new command to deal with exactly this problem: sage: d = DSage()# connect to the server you want to clear sage: d.kill_all() # remove all jobs from the job queue -- William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage under Windows/VMware
Thanks William, John: I just tested it and it works fine with execfile. Cheers, Yi -- http://www.yiqiang.org On 10/12/07, William Stein [EMAIL PROTECTED] wrote: On 10/12/07, John Voight [EMAIL PROTECTED] wrote: Thanks Yi, It's definitely a start. I absolutely need Cython compatibility--the whole reason I'm using SAGE is the ease at which I can write optimized code. Very interesting :-) Should we make this a trac ticket? Why can't the dsage worker just execute a load command? Because the workers don't run Ipython, etc. They are just running pure python after all preparsing, etc. However, don't worry, you can easily just embed/inline cython code in a .sage file as illustrated below, so the fact that loading .sage files works is enough. E.g., [EMAIL PROTECTED]:~/tmp$ more foo.sage cython( def foo(x,y): return x*y ) print foo(2,3) [EMAIL PROTECTED]:~/tmp$ [EMAIL PROTECTED]:~/tmp$ sage -- | SAGE Version 2.8.6, Release Date: 2007-10-06 | | Type notebook() for the GUI, and license() for information.| -- sage: load foo.sage 6 sage: Exiting SAGE (CPU time 0m0.04s, Wall time 0m14.25s). --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage under Windows/VMware
On 10/12/07, John Voight [EMAIL PROTECTED] wrote: Thanks Yi, It's definitely a start. I absolutely need Cython compatibility--the whole reason I'm using SAGE is the ease at which I can write optimized code. Very interesting :-) Should we make this a trac ticket? Why can't the dsage worker just execute a load command? Because the workers don't run Ipython, etc. They are just running pure python after all preparsing, etc. However, don't worry, you can easily just embed/inline cython code in a .sage file as illustrated below, so the fact that loading .sage files works is enough. E.g., [EMAIL PROTECTED]:~/tmp$ more foo.sage cython( def foo(x,y): return x*y ) print foo(2,3) [EMAIL PROTECTED]:~/tmp$ [EMAIL PROTECTED]:~/tmp$ sage -- | SAGE Version 2.8.6, Release Date: 2007-10-06 | | Type notebook() for the GUI, and license() for information.| -- sage: load foo.sage 6 sage: Exiting SAGE (CPU time 0m0.04s, Wall time 0m14.25s). --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage Zombie issue on sage.math
Hi Michael, I think there is a problem with the way the individual workers are restarted which leaves them as zombie processes. I'll look into this. Cheers, Yi -- http://www.yiqiang.org On 10/11/07, mabshoff [EMAIL PROTECTED] wrote: Hello, zombies have taken over sage.math: top - 09:48:46 up 36 days, 18:09, 17 users, load average: 2.24, 1.79, 1.44 Tasks: 864 total, 3 running, 550 sleeping, 0 stopped, 311 zombie Cpu(s): 8.1%us, 1.0%sy, 13.1%ni, 76.7%id, 1.2%wa, 0.0%hi, 0.0%si, 0.0%st They all appear to be dsage jobs started by was yesterday. Any ideas what happened? Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage Zombie issue on sage.math
On 10/11/07, Yi Qiang [EMAIL PROTECTED] wrote: Hi Michael, I think there is a problem with the way the individual workers are restarted which leaves them as zombie processes. I'll look into this. Cheers, Yi Yep. I killed them all, since my entire huge job finished (yes!). But yeah, it would be good to make sure zombies don't get created in the future. William -- http://www.yiqiang.org On 10/11/07, mabshoff [EMAIL PROTECTED] wrote: Hello, zombies have taken over sage.math: top - 09:48:46 up 36 days, 18:09, 17 users, load average: 2.24, 1.79, 1.44 Tasks: 864 total, 3 running, 550 sleeping, 0 stopped, 311 zombie Cpu(s): 8.1%us, 1.0%sy, 13.1%ni, 76.7%id, 1.2%wa, 0.0%hi, 0.0%si, 0.0%st They all appear to be dsage jobs started by was yesterday. Any ideas what happened? Cheers, Michael -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage Zombie issue on sage.math
On Oct 11, 2007, at 10:19 AM, William Stein wrote: On 10/11/07, Yi Qiang [EMAIL PROTECTED] wrote: Hi Michael, I think there is a problem with the way the individual workers are restarted which leaves them as zombie processes. I'll look into this. Cheers, Yi Yep. I killed them all, since my entire huge job finished (yes!). But yeah, it would be good to make sure zombies don't get created in the future. FWIW, zombies are fairly light-weight processes :-} They are really skeletons of what the process was: essentially enough state information so that the parent can reap them and collect the status information (I think it amounts to a process descriptor in the kernel), so even 300 of them shouldn't be a drain (despite Hollywood's unflattering depiction of them :-}). It does pay to get rid of them, of course. Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for the Absorption of Federal Funds Men are from Earth. Women are from Earth. Deal with it. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage Zombie issue on sage.math
Yup, you are right. The processes are dead but have an entry in the process table. I fixed this issue btw. Cheers, Yi -- http://www.yiqiang.org On 10/11/07, Justin C. Walker [EMAIL PROTECTED] wrote: On Oct 11, 2007, at 10:19 AM, William Stein wrote: On 10/11/07, Yi Qiang [EMAIL PROTECTED] wrote: Hi Michael, I think there is a problem with the way the individual workers are restarted which leaves them as zombie processes. I'll look into this. Cheers, Yi Yep. I killed them all, since my entire huge job finished (yes!). But yeah, it would be good to make sure zombies don't get created in the future. FWIW, zombies are fairly light-weight processes :-} They are really skeletons of what the process was: essentially enough state information so that the parent can reap them and collect the status information (I think it amounts to a process descriptor in the kernel), so even 300 of them shouldn't be a drain (despite Hollywood's unflattering depiction of them :-}). It does pay to get rid of them, of course. Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for the Absorption of Federal Funds Men are from Earth. Women are from Earth. Deal with it. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage
What is screen? Unfortunately, it is not an easy thing to Google! JV --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage
2007/10/10, John Voight [EMAIL PROTECTED]: What is screen? Unfortunately, it is not an easy thing to Google! Here is a link to screen : http://www.gnu.org/software/screen/ There's a quick tutorial here: http://www.kuro5hin.org/story/2004/3/9/16838/14935 didier --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage
On 10/10/07, didier deshommes [EMAIL PROTECTED] wrote: 2007/10/10, John Voight [EMAIL PROTECTED]: What is screen? Unfortunately, it is not an easy thing to Google! Here is a link to screen : http://www.gnu.org/software/screen/ There's a quick tutorial here: http://www.kuro5hin.org/story/2004/3/9/16838/14935 Thanks for posting those. Screen is a really amazing program that anybody run calculations on a couple of machines could find very very useful. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage under Windows/VMware
On 10/10/07, John Voight [EMAIL PROTECTED] wrote: Anyone get DSage to work under Windows/VMware? I'm able to get workers (under Windows) to connect to sage.math when I run the client/server there, so that works. But repeating the same client/server steps on a Windows machine has yielded nothing. Just wondered if anyone has tried this--or maybe you all live in a happy world where you're not surrounded by Windows machines... I don't even understand what you're asking. Do want the dsage *server* to run under windows, in addition to the workers? This would probably require changing the VMware machine to use shared networking instead of NAT. You'll also have to possibly do all kinds of things to get around Windows firewalling, anti-virus programs, etc. But why do this? What's wrong with just running the server under UNIX and the clients under Windows? It's probably more secure, etc. By the way, I think dsage has tremendous potential, but is ill-documented and rough around the edges, so you should report all issues you have, ask questions, etc. Here is something Yi just wrote that might be helpful for your question: http://sage.math.washington.edu/tmp/dsage.html ---William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage
On 10/10/07, William Stein [EMAIL PROTECTED] wrote: On 10/10/07, didier deshommes [EMAIL PROTECTED] wrote: 2007/10/10, John Voight [EMAIL PROTECTED]: What is screen? Unfortunately, it is not an easy thing to Google! Here is a link to screen : http://www.gnu.org/software/screen/ There's a quick tutorial here: http://www.kuro5hin.org/story/2004/3/9/16838/14935 Thanks for posting those. Screen is a really amazing program that anybody run calculations on a couple of machines could find very very useful. Yup. In case anyone finds it useful, here's my .screenrc file. In particular it has tricks to make it visually easy to track multiple logins into many machines (via the statusbar), and it remaps the screen escape to Ctrl-J, so C-a can be used for beginning-of-line, like god intended. planck[~] cat .screenrc vbell off # Set the escape to be Ctrl-j instead of C-a, which I want for line editing escape ^Jj # define a bigger scrollback, default is 100 lines defscrollback 1024 # fix backspace so it works inside jed and other curses programs bindkey -k kb stuff \177 # An alternative hardstatus to display a bar at the bottom listing the # windownames and highlighting the current windowname in blue. (This is only # enabled if there is no hardstatus setting for your terminal) hardstatus on hardstatus alwayslastline hardstatus string %{.bW}%-w%{.rW}%n %t%{-}%+w %=%{..G} %H %{..Y} %m/%d %C%a #- # STARTUP SCREENS #- # Example of automatically running some programs in windows on screen startup. # Use the form: # screen [-flag] [window name] [window number] [program to run in window] # The -t flag means that the window in which that program is running # closes when you exit that program #screen -t littlewood 0 tcsh #screen -t littlewood 1 tcsh #screen -t planck 2 ssh planck #screen -t szego 3 ssh szego # I cribbed most of this from /etc/screenrc # Check it out on your own to learn more about keybindings and whatnot. ## EOF Cheers, f --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage under Windows/VMware
Hi John, I have only tested DSage with workers running in VMWare and not servers. Everyone else does live in the 'happy' place :) I will try to get both the server and workers running in Windows if I can this weekend and document the process. Cheers, Yi -- http://www.yiqiang.org On 10/10/07, John Voight [EMAIL PROTECTED] wrote: Anyone get DSage to work under Windows/VMware? I'm able to get workers (under Windows) to connect to sage.math when I run the client/server there, so that works. But repeating the same client/server steps on a Windows machine has yielded nothing. Just wondered if anyone has tried this--or maybe you all live in a happy world where you're not surrounded by Windows machines... JV --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage under Windows/VMware
John, I also wanted to mention that you should apply this latest dsage bundle since it fixes a lot of the rough edges that William mentioned. http://sage.math.washington.edu/home/yqiang/dsage_latest.hg There is documentation on how to do this here: http://www.sagemath.org/doc/html/ref/module-sage.misc.hg.html Let me know if you need any help doing this. Cheers, Yi -- http://www.yiqiang.org On 10/10/07, John Voight [EMAIL PROTECTED] wrote: Thanks! Yes, if you can actually believe it, I was asking (quite gently) about running the DSage client/server under Windows. Other than my dual- boot machine at home, I don't have access to a Linux machine here at UVM--in particular the 31-machine lab where I'd like to run projects from is all Windows machines. I've gotten it to work with sage.math running as the client/server-- does that especially bother you, Will? Yours, John Voight Assistant Professor of Mathematics University of Vermont [EMAIL PROTECTED] [EMAIL PROTECTED] http://www.cems.uvm.edu/~voight/ --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSage under Windows/VMware
Thanks! Yes, if you can actually believe it, I was asking (quite gently) about running the DSage client/server under Windows. Other than my dual- boot machine at home, I don't have access to a Linux machine here at UVM--in particular the 31-machine lab where I'd like to run projects from is all Windows machines. I've gotten it to work with sage.math running as the client/server-- does that especially bother you, Will? Yours, John Voight Assistant Professor of Mathematics University of Vermont [EMAIL PROTECTED] [EMAIL PROTECTED] http://www.cems.uvm.edu/~voight/ --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
The web user should also be able to change the format of time stamps. Creation time should not have underscore and same with monitor_id. For completed job there should be a time stamp for how long the job took. On 4/25/07, Timothy Clemans [EMAIL PROTECTED] wrote: I like it a lot. A lot of job ids are repeated. It says 449 jobs but I don't see nearly that many. You need an HTML title. I don't have access to sage.math so I can't start jobs. There should be say a WebSAGE server that this app would be apart of and admin(s) chose whether or not to enable it, the notebook, and wiki. On 4/25/07, alex clemesha [EMAIL PROTECTED] wrote: Hi all, I've been working on making a web interface to dsage, check it out: http://sage.math.washington.edu:/ made with Twisted, MochiKit and sqlite. Couple of notes: This is an Ajax app, I've only viewed it in Firefox, but I will support FF, IE6 and IE7, Safari, Opera (via the excellent js library MochiKit, http://mochikit.com ) 'SELECT STATUS OF TYPE' is not turned on yet, would that be a good addition? there could be efficiency issues with very large dynamically changing stats in browsers, right now the view is basically: jobs_you_are_seeing = [random.choice(query_result) for _ in range(20)] One feature I will be putting in is a 'user console' to manage one's personal jobs, (ie, view stats, starting, stopping, uploading new jobs). I have code that will enable this feature to be *SSL and cookie-based-session enabled* Other thoughts? Other stat views? Feature requests, styling issues? This code is alpha-ish, so be nice ;) Now start some jobs on sage.math and view their status! -Alex p.s. This server was started running this command in the dir /home/agc/dsage_stats on sage.math: $ sage td.py -oy dsage_stats_server.tac to kill the server started in this way type (in the same dir): $ kill -9 `cat twistd.pid` --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
On Apr 25, 2007, at 13:28 , alex clemesha wrote: Hi all, I've been working on making a web interface to dsage, check it out: http://sage.math.washington.edu:/ made with Twisted, MochiKit and sqlite. Really nice! I like the fact that you can see the code that's behind the job. Having said that, might it be a good idea to allow 'locking' of this feature (perhaps with password protection)? I can foresee situations in which this would be useful. Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for the Absorption of Federal Funds Men are from Earth. Women are from Earth. Deal with it. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
A couple more comments: 1) Could you make the job_id column entries links to the specific job? It is a lot easier to click on it than to copy paste it into the box. 2) Could you make the default refresh time to be like 10 seconds or so? 5 seconds feels to short and was disconcerting since it jumped around so much. On Apr 25, 2007, at 1:28 PM, alex clemesha wrote: Hi all, I've been working on making a web interface to dsage, check it out: http://sage.math.washington.edu:/ made with Twisted, MochiKit and sqlite. Couple of notes: This is an Ajax app, I've only viewed it in Firefox, but I will support FF, IE6 and IE7, Safari, Opera (via the excellent js library MochiKit, http://mochikit.com) 'SELECT STATUS OF TYPE' is not turned on yet, would that be a good addition? there could be efficiency issues with very large dynamically changing stats in browsers, right now the view is basically: jobs_you_are_seeing = [random.choice(query_result) for _ in range(20)] One feature I will be putting in is a 'user console' to manage one's personal jobs, (ie, view stats, starting, stopping, uploading new jobs). I have code that will enable this feature to be *SSL and cookie-based-session enabled* Other thoughts? Other stat views? Feature requests, styling issues? This code is alpha-ish, so be nice ;) Now start some jobs on sage.math and view their status! -Alex p.s. This server was started running this command in the dir /home/agc/dsage_stats on sage.math: $ sage td.py -oy dsage_stats_server.tac to kill the server started in this way type (in the same dir): $ kill -9 `cat twistd.pid` Cheers, Yi -- http://www.yiqiang.net --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
On Apr 25, 2007, at 18:18 , Yi Qiang wrote: On Apr 25, 2007, at 4:49 PM, Justin C. Walker wrote: On Apr 25, 2007, at 13:28 , alex clemesha wrote: Hi all, I've been working on making a web interface to dsage, check it out: http://sage.math.washington.edu:/ made with Twisted, MochiKit and sqlite. Really nice! I like the fact that you can see the code that's behind the job. Having said that, might it be a good idea to allow 'locking' of this feature (perhaps with password protection)? I can foresee situations in which this would be useful. I think a easy solution would be to be able to have 'private' jobs. I.E. the job_id column would just show 'PRIVATE'. If you knew the job id you can just put it into the the search box alex already made. That was the reason job ids were random to begin with anyways. What do you think of this solution? At first glance, it seems like a good idea. I really don't know how important it is to have 'private' jobs, though. If it is *really* important, then having *really* random job IDs would be equally important :-}. Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for the Enhancement of the Director's Income Experience is what you get when you don't get what you want. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
On Apr 25, 2007, at 6:18 PM, Yi Qiang wrote: On Apr 25, 2007, at 4:49 PM, Justin C. Walker wrote: On Apr 25, 2007, at 13:28 , alex clemesha wrote: Having said that, might it be a good idea to allow 'locking' of this feature (perhaps with password protection)? I can foresee situations in which this would be useful. I think a easy solution would be to be able to have 'private' jobs. I.E. the job_id column would just show 'PRIVATE'. If you knew the job id you can just put it into the the search box alex already made. That was the reason job ids were random to begin with anyways. What do you think of this solution? First, let me say that this dsage job page is pretty cool. About the random job ids, sounds like security through obscurity to me. There's already the concept of job ownership, and levels of trust controlling who can submit/process a job, so I think (long term at least) that infrastructure could be used to control viewing (or even showing up in the list). --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
On Wed, 25 Apr 2007, Robert Bradshaw wrote: On Apr 25, 2007, at 6:18 PM, Yi Qiang wrote: On Apr 25, 2007, at 4:49 PM, Justin C. Walker wrote: On Apr 25, 2007, at 13:28 , alex clemesha wrote: Having said that, might it be a good idea to allow 'locking' of this feature (perhaps with password protection)? I can foresee situations in which this would be useful. I think a easy solution would be to be able to have 'private' jobs. I.E. the job_id column would just show 'PRIVATE'. If you knew the job id you can just put it into the the search box alex already made. That was the reason job ids were random to begin with anyways. What do you think of this solution? First, let me say that this dsage job page is pretty cool. About the random job ids, sounds like security through obscurity to me. There's already the concept of job ownership, and levels of trust controlling who can submit/process a job, so I think (long term at least) that infrastructure could be used to control viewing (or even showing up in the list). Gonna have to second that. If you have a web-based interface that takes as input an id #, a BASH one-liner can get all your data in a very short amount of time. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage stats
On Apr 25, 2007, at 19:06 , Robert Bradshaw wrote: On Apr 25, 2007, at 6:18 PM, Yi Qiang wrote: On Apr 25, 2007, at 4:49 PM, Justin C. Walker wrote: On Apr 25, 2007, at 13:28 , alex clemesha wrote: Having said that, might it be a good idea to allow 'locking' of this feature (perhaps with password protection)? I can foresee situations in which this would be useful. [snip] About the random job ids, sounds like security through obscurity to me. There's already the concept of job ownership, and levels of trust controlling who can submit/process a job, so I think (long term at least) that infrastructure could be used to control viewing (or even showing up in the list). Agreed. If that can work simply and intuitively, it's a big improvement over obscurity... Justin -- Justin C. Walker, Curmudgeon-At-Large, Director Institute for the Enhancement of the Director's Income The path of least resistance: it's not just for electricity any more. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSAGE Questions
On 3/18/07, Yi Qiang [EMAIL PROTECTED] wrote: Are you trying to run DSAGE between different versions of SAGE? The way that jobs are handled changed quite a bit, which is why it complains about some attributes not being there. Try it again using sage-2.3 throughout. I suspect that is what is happening too. By the way, could you make it so the .sage/dsage config files have some sort of version, so instead of getting weird random errors, the users gets a message that they have to re-run dsage.[whatever] to generate new config files? This would save a lot of confusion. Likewise, it would be very useful if a job is run on a worker that is an old version, there is a warning, or error, or something. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSAGE Questions
That really shouldn't be happening. It looks like Worker 1 and Worker 14 are getting the same job which is definitely not right. Your job database might have gotten screwed up. I think the easiest thing to try right now is to remove the $HOME/.sage/dsage/db and $HOME/.sage/dsage/tmp_worker_files directory and try again. This happens even after wiping .sage/dsage and running dsage.setup(). Are you trying to run DSAGE between different versions of SAGE? The way that jobs are handled changed quite a bit, which is why it complains about some attributes not being there. Try it again using sage-2.3 throughout. I've been tesitng with two separate DSAGE instances so that there is no talking between different versions of DSAGE. It works fine on my local machine with 2.1.4, but gives the Job attribute when I try just running it the 2.3 install. This is the job method that I'm using: def t32_job(self, i): job = Job(file= %s h_basis = load_da_basis(%s) k_basis = load_da_basis(%s) pairs = %s result = map(lambda x: theorem32(h_basis[x[0]], k_basis[x[1]], %s), pairs) save(result, 'result') DSAGE_RESULT = 'result.sobj' % (self.file_text, str(self.h), str(self.k), str(self.p_pairs[i]), str(self.rep) ) , name='t32') job.i = i return job Is there something obvious that I'm missing with it? --Mike --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: DSAGE Questions
On Mar 18, 2007, at 2:21 PM, Mike Hansen wrote: That really shouldn't be happening. It looks like Worker 1 and Worker 14 are getting the same job which is definitely not right. Your job database might have gotten screwed up. I think the easiest thing to try right now is to remove the $HOME/.sage/dsage/db and $HOME/.sage/dsage/tmp_worker_files directory and try again. This happens even after wiping .sage/dsage and running dsage.setup(). Are you trying to run DSAGE between different versions of SAGE? The way that jobs are handled changed quite a bit, which is why it complains about some attributes not being there. Try it again using sage-2.3 throughout. I've been tesitng with two separate DSAGE instances so that there is no talking between different versions of DSAGE. It works fine on my local machine with 2.1.4, but gives the Job attribute when I try just running it the 2.3 install. This is the job method that I'm using: def t32_job(self, i): job = Job(file= %s h_basis = load_da_basis(%s) k_basis = load_da_basis(%s) pairs = %s result = map(lambda x: theorem32(h_basis[x[0]], k_basis[x[1]], %s), pairs) save(result, 'result') DSAGE_RESULT = 'result.sobj' % (self.file_text, str(self.h), str(self.k), str(self.p_pairs[i]), str(self.rep) ) , name='t32') job.i = i return job Is there something obvious that I'm missing with it? Mike and I figured this one out over google talk. It appears that the problem was with setting the attribute 'job.i'. For some reason, setting that attribute caused unpickling the Job object on the otherside to fail. Changing that one line to 'job.n = i' fixed the problem. Cheers, Yi --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: dsage docstring
On Wed, 31 Jan 2007, Yi Qiang wrote: Hey what docstrings were not available? When I type dsage? I get a docstring, as well as dsage.server? and dsage.worker? Yi, My opinion was that Release early, release often is great but one should keep an key on documentation and testing. In particular, my concerns were as follows (and not as stated above): 1. the nodoctests phrase appearing at the beginning of the dsage/dsage.py file 2. tab completion does not bring up DistributedFactor 3. the lack of examples in the docstrings It became apparent during your talk and the discussion which followed that #2, #3 will be addressed in the immediate future. Regards, Ifti --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---