Performance has been fairly reasonable lately, until just now... from PharoLauncher I downloaded Pharo 7 latest-32 and that zipped along okay, then immediately downloaded Pharo 7 latest-64 and it crawled. About 30 seconds or more per 1%. That would have been ~16:50TZ+8) I went to dinner. After dinner ~19:30TZ+8 I found it timed out and kicked off another that was a tiny bit better, but still crawling. Interestingly, from a terminal I then did... $ wget http://files.pharo.org/image/70/latest-64.zip and for the first 15% it crawled at 20kB/s. Then up to 50% it ramped steadily up to 400kB/s, then the last 50% it smoothly ramped up 1.5MB/s and by the end was zooming. Meanwhile, PharoLauncher continued plodding along and finished a few minutes later with a "CRCError: Wrong CRC-32 (expected 55B3346A got 2DD68EFA) (proceed to ignore)" , which I've never seen before. About five minutes later I tried wget with the same behaviour, and then PharoLauncher again and it was much better taking about 20 seconds to successfully download Pharo 7 latest-64.zip. Paying close attention the download speed of PharoLauncher seemed to increase *slightly* as it went along, but nothing like wget's progressive improvement. I'm not quite sure what to make of that. The overlapped downloads by PharoLauncher and wget performing differently seems significant. Does the server have some throttling mechanism that maybe adapts its policies due to overall load, which Pharo is interacting with poorly? Another observation, using Chromium I found pages [1] http://files.pharo.org/image/60/ and [2] http://files.pharo.org/image/70/ very slow, with the latter being worse. In one instance refreshing [2] I counted 40 seconds with no response (blank page). Then tried refreshing [2] again and at the the 6 second mark refreshed [1], then at a count of 20 both pages simultaneously appeared. No other internet services are slow. My current connection speedtest rates my link at 34Mbps (~4MB/s). Unfortunately I didn't get a traceroute while things were really slow. Here is one after it got back to normal (so the poor performance lasted about 50 minutes). $ traceroute files.pharo.org traceroute to files.pharo.org (128.93.162.72), 30 hops max, 60 byte packets 1 192.168.43.1 (192.168.43.1) 3.122 ms 3.180 ms 4.437 ms 2 * * * 3 * * * 4 Bundle-Ether14.wel-edge901.perth.telstra.net (139.130.16.49) 70.991 ms 71.000 ms 70.984 ms 5 bundle-ether6.wel-core3.perth.telstra.net (203.50.6.214) 70.964 ms 70.944 ms 72.089 ms 6 bundle-ether7.fli-core1.adelaide.telstra.net (203.50.11.18) 85.918 ms 73.619 ms 82.271 ms 7 bundle-ether9.win-core10.melbourne.telstra.net (203.50.11.91) 98.565 ms 98.568 ms 98.560 ms 8 bundle-ether12.ken-core10.sydney.telstra.net (203.50.11.122) 109.279 ms 109.298 ms 109.288 ms 9 bundle-ether1.pad-gw11.sydney.telstra.net (203.50.6.61) 109.232 ms 98.702 ms 101.279 ms 10 bundle-ether1.sydp-core04.sydney.reach.com (203.50.13.90) 92.535 ms 104.351 ms 113.158 ms 11 i-52.tlot-core02.bx.telstraglobal.net (202.84.137.101) 294.263 ms 283.547 ms 281.576 ms 12 i-0-7-0-13.tlot-core01.bi.telstraglobal.net (202.40.149.165) 279.967 ms i-0-2-0-11.tlot-core01.bi.telstraglobal.net (202.40.149.185) 279.948 ms i-0-7-0-11.tlot-core01.bi.telstraglobal.net (202.84.251.233) 277.390 ms 13 gtt-peer.tlot02.pr.telstraglobal.net (134.159.63.182) 280.217 ms 288.853 ms 275.764 ms 14 xe-1-2-0.cr0-par7.ip4.gtt.net (89.149.185.53) 461.558 ms 445.153 ms 445.097 ms 15 renater-gw-ix1.gtt.net (77.67.123.206) 445.083 ms 443.652 ms 441.766 ms 16 * * * 17 inria-rocquencourt-te1-4-inria-rtr-021.noc.renater.fr (193.51.184.177) 421.283 ms 480.646 ms 424.906 ms cheers -ben On 29 September 2017 at 03:11, Stephane Ducasse <[hidden email]> wrote: We opened a new ticket. |
We are experimenting with a mirror that seems to work good.
I mailed you some info privately. I am not sure if we can make this public already (Marcus ?) > On 17 Nov 2017, at 13:05, Ben Coman <[hidden email]> wrote: > > Performance has been fairly reasonable lately, until just now... > from PharoLauncher I downloaded Pharo 7 latest-32 and that zipped along okay, > then immediately downloaded Pharo 7 latest-64 and it crawled. About 30 seconds or more per 1%. That would have been ~16:50TZ+8) > I went to dinner. After dinner ~19:30TZ+8 I found it timed out and kicked off another that was a tiny bit better, but still crawling. > Interestingly, from a terminal I then did... $ wget http://files.pharo.org/image/70/latest-64.zip > and for the first 15% it crawled at 20kB/s. Then up to 50% it ramped steadily up to 400kB/s, then the last 50% it smoothly ramped up 1.5MB/s and by the end was zooming. Meanwhile, PharoLauncher continued plodding along and finished a few minutes later with a "CRCError: Wrong CRC-32 (expected 55B3346A got 2DD68EFA) (proceed to ignore)" , which I've never seen before. > > About five minutes later I tried wget with the same behaviour, and then PharoLauncher again and it was much better taking about 20 seconds to successfully download Pharo 7 latest-64.zip. Paying close attention the download speed of PharoLauncher seemed to increase *slightly* as it went along, but nothing like wget's progressive improvement. > > I'm not quite sure what to make of that. The overlapped downloads by PharoLauncher and wget performing differently seems significant. Does the server have some throttling mechanism that maybe adapts its policies due to overall load, which Pharo is interacting with poorly? > > > Another observation, using Chromium I found pages [1] http://files.pharo.org/image/60/ and [2] http://files.pharo.org/image/70/ very slow, with the latter being worse. In one instance refreshing [2] I counted 40 seconds with no response (blank page). Then tried refreshing [2] again and at the the 6 second mark refreshed [1], then at a count of 20 both pages simultaneously appeared. > > No other internet services are slow. My current connection speedtest rates my link at 34Mbps (~4MB/s). > > Unfortunately I didn't get a traceroute while things were really slow. Here is one after it got back to normal (so the poor performance lasted about 50 minutes). > > $ traceroute files.pharo.org > traceroute to files.pharo.org (128.93.162.72), 30 hops max, 60 byte packets > 1 192.168.43.1 (192.168.43.1) 3.122 ms 3.180 ms 4.437 ms > 2 * * * > 3 * * * > 4 Bundle-Ether14.wel-edge901.perth.telstra.net (139.130.16.49) 70.991 ms 71.000 ms 70.984 ms > 5 bundle-ether6.wel-core3.perth.telstra.net (203.50.6.214) 70.964 ms 70.944 ms 72.089 ms > 6 bundle-ether7.fli-core1.adelaide.telstra.net (203.50.11.18) 85.918 ms 73.619 ms 82.271 ms > 7 bundle-ether9.win-core10.melbourne.telstra.net (203.50.11.91) 98.565 ms 98.568 ms 98.560 ms > 8 bundle-ether12.ken-core10.sydney.telstra.net (203.50.11.122) 109.279 ms 109.298 ms 109.288 ms > 9 bundle-ether1.pad-gw11.sydney.telstra.net (203.50.6.61) 109.232 ms 98.702 ms 101.279 ms > 10 bundle-ether1.sydp-core04.sydney.reach.com (203.50.13.90) 92.535 ms 104.351 ms 113.158 ms > 11 i-52.tlot-core02.bx.telstraglobal.net (202.84.137.101) 294.263 ms 283.547 ms 281.576 ms > 12 i-0-7-0-13.tlot-core01.bi.telstraglobal.net (202.40.149.165) 279.967 ms i-0-2-0-11.tlot-core01.bi.telstraglobal.net (202.40.149.185) 279.948 ms i-0-7-0-11.tlot-core01.bi.telstraglobal.net (202.84.251.233) 277.390 ms > 13 gtt-peer.tlot02.pr.telstraglobal.net (134.159.63.182) 280.217 ms 288.853 ms 275.764 ms > 14 xe-1-2-0.cr0-par7.ip4.gtt.net (89.149.185.53) 461.558 ms 445.153 ms 445.097 ms > 15 renater-gw-ix1.gtt.net (77.67.123.206) 445.083 ms 443.652 ms 441.766 ms > 16 * * * > 17 inria-rocquencourt-te1-4-inria-rtr-021.noc.renater.fr (193.51.184.177) 421.283 ms 480.646 ms 424.906 ms > > > > > > cheers -ben > > On 29 September 2017 at 03:11, Stephane Ducasse <[hidden email]> wrote: > We opened a new ticket. > > > On Tue, Sep 26, 2017 at 8:45 PM, Gabriel Cotelli <[hidden email]> wrote: > > I'm willing to do some other specific tests if needed. Just let me know. > > > > On Tue, Sep 26, 2017 at 3:36 PM, Stephane Ducasse <[hidden email]> > > wrote: > >> > >> Thanks Gabriel I forwarded this to christophe. > >> > >> On Tue, Sep 26, 2017 at 12:01 AM, Gabriel Cotelli <[hidden email]> > >> wrote: > >> > Still really slow from Argentina. > >> > > >> > In case it helps > >> > > >> > traceroute files.pharo.org > >> > traceroute to files.pharo.org (128.93.162.72), 30 hops max, 60 byte > >> > packets > >> > 1 192.168.1.1 (192.168.1.1) 0.574 ms 0.968 ms 1.538 ms > >> > 2 192.168.0.1 (192.168.0.1) 2.048 ms 2.499 ms 2.847 ms > >> > 3 10.33.32.1 (10.33.32.1) 12.264 ms 14.160 ms 14.336 ms > >> > 4 10.242.2.161 (10.242.2.161) 14.522 ms 14.643 ms 16.497 ms > >> > 5 telecentro.baires3.bai.seabone.net (195.22.220.33) 18.936 ms > >> > 19.575 ms > >> > 18.843 ms > >> > 6 ae0.baires3.bai.seabone.net (195.22.220.32) 19.554 ms 10.675 ms > >> > 13.028 ms > >> > 7 et10-1-0.miami15.mia.seabone.net (89.221.41.175) 246.284 ms > >> > et9-3-0.miami15.mia.seabone.net (195.22.199.179) 246.214 ms > >> > et4-3-0.miami15.mia.seabone.net (195.22.199.177) 246.137 ms > >> > 8 gtt.miami15.mia.seabone.net (89.221.41.197) 245.909 ms 245.924 ms > >> > 245.870 ms > >> > 9 xe-3-1-2.cr0-par7.ip4.gtt.net (141.136.109.18) 251.632 ms 274.447 > >> > ms > >> > 250.097 ms > >> > 10 renater-gw-ix1.gtt.net (77.67.123.206) 281.025 ms 259.509 ms > >> > 261.717 > >> > ms > >> > 11 * * * > >> > 12 inria-rocquencourt-te1-4-inria-rtr-021.noc.renater.fr > >> > (193.51.184.177) > >> > 271.263 ms 307.011 ms 306.876 ms > >> > 13 * * * > >> > 14 * * * > >> > 15 * * * > >> > 16 * * * > >> > 17 * * * > >> > 18 * * * > >> > 19 * * * > >> > 20 * * * > >> > 21 * * * > >> > 22 * * * > >> > 23 * * * > >> > 24 * * * > >> > 25 * * * > >> > 26 * * * > >> > 27 * * * > >> > 28 * * * > >> > 29 * * * > >> > 30 * * * > >> > > >> > httping -g http://files.pharo.org/get-files/61/pharo64.zip -c 20 > >> > PING files.pharo.org:80 (/get-files/61/pharo64.zip): > >> > connected to 128.93.162.72:80 (256 bytes), seq=0 time=625,90 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=1 time=842,02 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=2 time=638,83 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=3 time=637,63 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=4 time=637,85 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=5 time=580,27 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=6 time=1207,98 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=7 time=638,13 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=8 time=637,74 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=9 time=582,59 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=10 time=658,11 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=11 time=673,84 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=12 time=584,44 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=13 time=691,84 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=14 time=637,20 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=15 time=861,84 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=16 time=619,45 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=17 time=638,03 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=18 time=638,11 ms > >> > connected to 128.93.162.72:80 (256 bytes), seq=19 time=741,01 ms > >> > --- http://files.pharo.org/get-files/61/pharo64.zip ping statistics --- > >> > 20 connects, 20 ok, 0,00% failed, time 33781ms > >> > round-trip min/avg/max = 580,3/688,6/1208,0 ms > >> > > >> > > >> > On Thu, Sep 21, 2017 at 12:13 PM, Ben Coman <[hidden email]> wrote: > >> >> > >> >> Still seems a bit slow... > >> >> http://files.pharo.org/get-files/70/pharo64.zip > >> >> 16MB ==> 10 minutes (and actually it stalled after 3.4MB downloaded) > >> >> Local speed test: 40Mb/s down, 20Mb/s up > >> >> > >> >> but then I'm a long way across the internet, the cause could be > >> >> anywhere... > >> >> $ traceroute files.pharo.org > >> >> traceroute to files.pharo.org (128.93.162.72), 30 hops max, 60 byte > >> >> packets > >> >> 1 192.168.43.1 (192.168.43.1) 2.040 ms 2.403 ms 4.222 ms > >> >> 2 * * * > >> >> 3 * * * > >> >> 4 Bundle-Ether14.wel-edge901.perth.telstra.net (139.130.16.49) > >> >> 67.741 > >> >> ms 67.739 ms 67.730 ms > >> >> 5 bundle-ether6.wel-core3.perth.telstra.net (203.50.6.214) 67.696 ms > >> >> 67.697 ms 67.697 ms > >> >> 6 bundle-ether7.fli-core1.adelaide.telstra.net (203.50.11.18) 95.723 > >> >> ms > >> >> 91.753 ms 100.574 ms > >> >> 7 bundle-ether9.win-core10.melbourne.telstra.net (203.50.11.91) > >> >> 106.550 > >> >> ms 108.849 ms 108.851 ms > >> >> 8 bundle-ether12.ken-core10.sydney.telstra.net (203.50.11.122) > >> >> 117.770 > >> >> ms 117.767 ms 117.751 ms > >> >> 9 bundle-ether1.pad-gw11.sydney.telstra.net (203.50.6.61) 117.720 ms > >> >> 103.603 ms 102.171 ms > >> >> 10 bundle-ether1.sydp-core04.sydney.reach.com (203.50.13.90) 93.276 > >> >> ms > >> >> 95.822 ms 102.577 ms > >> >> 11 i-52.tlot-core02.bx.telstraglobal.net (202.84.137.101) 330.128 ms > >> >> 321.231 ms 322.522 ms > >> >> 12 i-0-4-0-10.tlot-core01.bi.telstraglobal.net (202.84.253.21) > >> >> 321.187 > >> >> ms i-0-2-0-10.tlot-core01.bi.telstraglobal.net (202.40.149.181) > >> >> 321.181 ms > >> >> i-0-2-0-9.tlot-core01.bi.telstraglobal.net (202.40.149.177) 308.786 ms > >> >> 13 gtt-peer.tlot02.pr.telstraglobal.net (134.159.63.182) 273.784 ms > >> >> 283.530 ms 281.079 ms > >> >> 14 xe-1-2-0.cr0-par7.ip4.gtt.net (89.149.185.53) 466.648 ms 479.910 > >> >> ms > >> >> 481.254 ms > >> >> 15 renater-gw-ix1.gtt.net (77.67.123.206) 479.900 ms 479.882 ms > >> >> 478.641 ms > >> >> 16 * * * > >> >> 17 inria-rocquencourt-te1-4-inria-rtr-021.noc.renater.fr > >> >> (193.51.184.177) > >> >> 436.776 ms 479.819 ms 478.907 ms > >> >> 18 * * * > >> >> 19 * * * > >> >> > >> >> $ httping -h files.pharo.org > >> >> PING files.pharo.org:80 (http://files.pharo.org:80/): > >> >> connected to 128.93.162.72:80 (169 bytes), seq=0 time=970.27 ms > >> >> connected to 128.93.162.72:80 (169 bytes), seq=1 time=928.26 ms > >> >> connected to 128.93.162.72:80 (169 bytes), seq=2 time=860.06 ms > >> >> connected to 128.93.162.72:80 (169 bytes), seq=3 time=976.04 ms > >> >> connected to 128.93.162.72:80 (169 bytes), seq=4 time=1238.72 ms > >> >> connected to 128.93.162.72:80 (169 bytes), seq=5 time=1239.65 ms > >> >> connected to 128.93.162.72:80 (169 bytes), seq=6 time=918.06 ms > >> >> > >> >> > >> >> cheers -ben > >> >> > >> >> On Thu, Sep 21, 2017 at 5:43 PM, Marcus Denker <[hidden email]> > >> >> wrote: > >> >>> > >> >>> > >> >>> > On 18 Sep 2017, at 08:37, Marcus Denker <[hidden email]> > >> >>> > wrote: > >> >>> > > >> >>> > > >> >>> >> On 16 Sep 2017, at 09:40, Stephane Ducasse > >> >>> >> <[hidden email]> > >> >>> >> wrote: > >> >>> >> > >> >>> >> Hi marcus > >> >>> >> > >> >>> >> I do not see why they could not find it. > >> >>> >> Can you tell me how you ask because we should reask? > >> >>> >> > >> >>> >> > >> >>> > > >> >>> > I spend a ticket and they wanted to close it last week. I have told > >> >>> > them to not close it > >> >>> > > >> >>> The problems seems to be fixed now. > >> >>> > >> >>> If someone sees any slowness -> send me a mail. > >> >>> > >> >>> Marcus > >> >>> > >> >>> > >> >> > >> > > >> > > > > |
Free forum by Nabble | Edit this page |