Hi Harald On Wed, Jul 22, 2009 at 07:07:09AM -0700, Harald Schilly wrote: > Below is a link to a python script that tests the active and up-to- > date mirrors, selects one of the 3 fastest (to be "fair" and > distribute the load, there is much random noise anyways) and then > returns the link to the now mirrored standard packages. I think that's > a nice addition, but I'm not familiar with the "sage -update" script > and the whole process. So, someone should look into this please. Tell > me if the script works for you! > > http://sage.math.washington.edu/home/schilly/sage_mirror_upgrade.py
Well, the choosing from the top three was not so good; rerunning makes that mirror clearly the worst. There aren't many, and seem to get them all in a few runs. Perhaps the random noise is enough of randomness...? 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.switch.ch 419 [ms] mirror.yandex.ru 710 [ms] modular.math.jmu.edu 1168 [ms] modular.fas.harvard.edu 1251 [ms] www.cecm.sfu.ca 1311 [ms] sagemath.c3sl.ufpr.br 1911 [ms] selected mirror: http://mirror.switch.ch/mirror/sagemath/packages/standard/ You test this by opening index.html? Everything here is proxied, and in low bandwidth environments HAVE to be proxied. By policy and techically enforced. So the one I chose is the only one not proxied? :( 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] modular.math.jmu.edu 6 [ms] modular.fas.harvard.edu 7 [ms] www.cecm.sfu.ca 7 [ms] sagemath.c3sl.ufpr.br 8 [ms] mirror.switch.ch 363 [ms] selected mirror: http://modular.math.jmu.edu/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] www.cecm.sfu.ca 5 [ms] modular.math.jmu.edu 6 [ms] sagemath.c3sl.ufpr.br 6 [ms] modular.fas.harvard.edu 8 [ms] mirror.switch.ch 398 [ms] selected mirror: http://mirror.yandex.ru/mirrors/sage.math.washington.edu/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] modular.fas.harvard.edu 7 [ms] www.cecm.sfu.ca 7 [ms] modular.math.jmu.edu 8 [ms] sagemath.c3sl.ufpr.br 8 [ms] mirror.switch.ch 368 [ms] selected mirror: http://www.cecm.sfu.ca/sage/packages/standard/ Gar, enough of that. So I comment out random.shuffle(top) and try again, and get the same one all the time: 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] www.cecm.sfu.ca 5 [ms] modular.math.jmu.edu 6 [ms] modular.fas.harvard.edu 7 [ms] sagemath.c3sl.ufpr.br 7 [ms] mirror.switch.ch 363 [ms] selected mirror: http://mirror.yandex.ru/mirrors/sage.math.washington.edu/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] modular.fas.harvard.edu 4 [ms] modular.math.jmu.edu 5 [ms] sagemath.c3sl.ufpr.br 9 [ms] www.cecm.sfu.ca 10 [ms] mirror.switch.ch 363 [ms] selected mirror: http://mirror.yandex.ru/mirrors/sage.math.washington.edu/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 2 [ms] modular.fas.harvard.edu 6 [ms] www.cecm.sfu.ca 5 [ms] modular.math.jmu.edu 6 [ms] sagemath.c3sl.ufpr.br 6 [ms] mirror.switch.ch 372 [ms] selected mirror: http://mirror.yandex.ru/mirrors/sage.math.washington.edu/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] modular.math.jmu.edu 7 [ms] sagemath.c3sl.ufpr.br 8 [ms] modular.fas.harvard.edu 10 [ms] www.cecm.sfu.ca 9 [ms] mirror.switch.ch 453 [ms] selected mirror: http://mirror.yandex.ru/mirrors/sage.math.washington.edu/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py mirror.yandex.ru 3 [ms] modular.fas.harvard.edu 3 [ms] sagemath.c3sl.ufpr.br 6 [ms] modular.math.jmu.edu 7 [ms] www.cecm.sfu.ca 11 [ms] mirror.switch.ch 371 [ms] selected mirror: http://mirror.yandex.ru/mirrors/sage.math.washington.edu/packages/standard/ So I see my local mirror sun.ac.za is not there? Because the tar is not unpacked there? Also, I would print the actual command to run: sage -upgrade http://... A ping time would be better than opening. We have pings in the country of 1-100ms and through to Europe/US at 400-800. (Debian does this, I think, and mistakenly keeps giving me mirrors in Eastern Europe) Then again some servers don't allow pings... --- echidna.maths.usyd.edu.au ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 3998ms rtt min/avg/max/mdev = 504.555/531.527/548.337/16.717 ms --- mimas-nxge0.switch.ch ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4000ms rtt min/avg/max/mdev = 175.483/199.152/249.654/28.797 ms --- mirror.yandex.ru ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4332ms rtt min/avg/max/mdev = 206.620/230.290/241.420/12.302 ms --- modular.fas.harvard.edu ping statistics --- 5 packets transmitted, 0 received, 100% packet loss, time 4031ms --- modular.math.jmu.edu ping statistics --- 5 packets transmitted, 0 received, 100% packet loss, time 4030ms --- sagres.c3sl.ufpr.br ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4677ms rtt min/avg/max/mdev = 388.971/405.773/434.684/16.217 ms --- newweb.cecm.sfu.ca ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4889ms rtt min/avg/max/mdev = 310.205/339.340/384.164/27.543 ms --- ftp.sun.ac.za ping statistics --- 5 packets transmitted, 5 received, 0% packet loss, time 4005ms rtt min/avg/max/mdev = 52.209/53.966/55.909/1.346 ms A clear winner there for the local mirror (last, not on your list) So the proxy messes it all up, but I added mirrors.append('ftp://ftp.sun.ac.za/pub/mirrors/www.sagemath.org/') and tried again. 0 j...@muizenberg:~$python sage_mirror_upgrade.py echidna.maths.usyd.edu.au 4 [ms] www.cecm.sfu.ca 3 [ms] sagemath.c3sl.ufpr.br 8 [ms] modular.math.jmu.edu 8 [ms] modular.fas.harvard.edu 9 [ms] mirror.yandex.ru 10 [ms] mirror.switch.ch 372 [ms] ftp.sun.ac.za 743 [ms] selected mirror: http://www.cecm.sfu.ca/sage/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py echidna.maths.usyd.edu.au 2 [ms] sagemath.c3sl.ufpr.br 6 [ms] modular.fas.harvard.edu 6 [ms] mirror.yandex.ru 7 [ms] www.cecm.sfu.ca 6 [ms] modular.math.jmu.edu 7 [ms] mirror.switch.ch 389 [ms] ftp.sun.ac.za 656 [ms] selected mirror: http://echidna.maths.usyd.edu.au/sage/packages/standard/ hmm, ftp is not proxied. So I chanved to mirrors.append('http://ftp.sun.ac.za/ftp/.sdb1/www.sagemath.org2/') 0 j...@muizenberg:~$python sage_mirror_upgrade.py echidna.maths.usyd.edu.au 3 [ms] modular.fas.harvard.edu 6 [ms] modular.math.jmu.edu 6 [ms] www.cecm.sfu.ca 6 [ms] mirror.yandex.ru 8 [ms] sagemath.c3sl.ufpr.br 8 [ms] ftp.sun.ac.za 129 [ms] mirror.switch.ch 471 [ms] selected mirror: http://echidna.maths.usyd.edu.au/sage/packages/standard/ 0 j...@muizenberg:~$python sage_mirror_upgrade.py echidna.maths.usyd.edu.au 4 [ms] ftp.sun.ac.za 7 [ms] mirror.yandex.ru 9 [ms] sagemath.c3sl.ufpr.br 8 [ms] modular.fas.harvard.edu 11 [ms] www.cecm.sfu.ca 11 [ms] modular.math.jmu.edu 15 [ms] mirror.switch.ch 367 [ms] selected mirror: http://echidna.maths.usyd.edu.au/sage/packages/standard/ getting there?... So I log onto the proxy and unset http_proxy and try again: oops no function u.getcode. So I uncomment that for the moment and try again: This looks more realistic: 0 j...@musselcracker:~$python sage_mirror_upgrade.py ftp.sun.ac.za 190 [ms] mirror.switch.ch 485 [ms] mirror.yandex.ru 530 [ms] modular.math.jmu.edu 614 [ms] modular.fas.harvard.edu 661 [ms] www.cecm.sfu.ca 730 [ms] sagemath.c3sl.ufpr.br 929 [ms] echidna.maths.usyd.edu.au 1051 [ms] selected mirror: http://ftp.sun.ac.za/ftp/.sdb1/www.sagemath.org2/packages/standard/ 0 j...@musselcracker:~$python sage_mirror_upgrade.py ftp.sun.ac.za 126 [ms] mirror.switch.ch 375 [ms] modular.math.jmu.edu 518 [ms] mirror.yandex.ru 532 [ms] modular.fas.harvard.edu 567 [ms] www.cecm.sfu.ca 640 [ms] sagemath.c3sl.ufpr.br 796 [ms] echidna.maths.usyd.edu.au 954 [ms] selected mirror: http://ftp.sun.ac.za/ftp/.sdb1/www.sagemath.org2/packages/standard/ And really, as a sysadm, I found out which mirror to use. I was more thinking about the end user who wants it transparantly. But this is a good start! regards, Jan -- .~. /V\ Jan Groenewald /( )\ www.aims.ac.za ^^-^^ --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
