Re: Receiving objects: 44% (121381/272167), 124.05 MiB | 78 KiB/s

[Jeff Schroeder <jeffschroed gmail com>]

On Tue, May 5, 2009 at 10:49 AM, Pierre <pierre pwi dyndns biz> wrote:
>> Other European users seeing a similar image?
>>
>> Simos
>
> From Bordeaux (France), I see :
>
> $ ping svn.gnome.org
> PING svn.gnome.org (91.189.93.3) 56(84) bytes of data.
> 64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=1 ttl=47 time=74.5 ms
> 64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=2 ttl=47 time=74.2 ms
> 64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=3 ttl=47 time=70.4 ms
> 64 bytes from cobalt.canonical.com (91.189.93.3): icmp_seq=4 ttl=47 time=71.0 ms
>
> $ ping git.gnome.org
> PING git.gnome.org (209.132.176.202) 56(84) bytes of data.
> 64 bytes from git.gnome.org (209.132.176.202): icmp_seq=1 ttl=46 time=262 ms
> 64 bytes from git.gnome.org (209.132.176.202): icmp_seq=2 ttl=46 time=262 ms
> 64 bytes from git.gnome.org (209.132.176.202): icmp_seq=3 ttl=46 time=258 ms
> 64 bytes from git.gnome.org (209.132.176.202): icmp_seq=4 ttl=46 time=260 ms
>
> # mtr -rwc5 git.gnome.org
> HOST: kermit.pwi.net                             Loss%   Snt   Last   Avg  Best  Wrst StDev
>  1. moustik.pwi.net                             0.0%     5    0.3   0.3   0.2   0.3   0.0
>  2. 88.182.20.254                               0.0%     5   53.5  53.2  52.3  53.9   0.6
>  3. 213.228.9.190                              60.0%     5   54.7  54.8  54.7  54.9   0.1
>  4. bzn-crs16-1-be1100.intf.routers.proxad.net  0.0%     5   63.4  62.6  61.7  63.4   0.8
>  5. th2-crs16-1-be2000.intf.routers.proxad.net  0.0%     5   65.0  63.5  62.5  65.0   1.0
>  6. xe-11-1-0.edge3.Paris1.Level3.net           0.0%     5   63.4  63.3  62.7  63.9   0.5
>  7. ae-32-56.ebr2.Paris1.Level3.net             0.0%     5   68.0  66.1  63.6  68.8   2.4
>  8. ae-1-100.ebr1.Paris1.Level3.net             0.0%     5   73.1  71.2  66.9  76.2   3.5
>  9. ae-43.ebr1.London2.Level3.net               0.0%     5   73.2  74.0  71.3  82.3   4.7
>  10. ae-47-107.ebr2.London2.Level3.net           0.0%     5   71.4  78.1  71.4  82.8   5.0
>  11. ae-42-42.ebr1.London1.Level3.net            0.0%     5   71.5  71.3  69.8  72.6   1.1
>  12. ae-100-100.ebr2.London1.Level3.net          0.0%     5   70.9  72.1  69.7  77.0   2.9
>  13. ae-41-41.ebr1.NewYork1.Level3.net           0.0%     5  136.8 137.3 136.6 138.5   0.8
>  14. ae-71-71.csw2.NewYork1.Level3.net           0.0%     5  137.2 140.7 137.2 148.1   4.5
>  15. ae-74-74.ebr4.NewYork1.Level3.net           0.0%     5  143.7 140.4 136.8 143.7   3.0
>  16. ae-2.ebr4.SanJose1.Level3.net               0.0%     5  278.4 278.8 276.9 280.3   1.3
>  17. ae-74-74.csw2.SanJose1.Level3.net           0.0%     5  228.5 232.0 227.0 237.4   4.6
>  18. ae-23-79.car3.SanJose1.Level3.net          20.0%     5  275.2 277.8 275.2 280.4   2.2
>  19. te-4-1-73.rp0-1.snjsca13.Level3.net        60.0%     5  228.6 229.5 228.6 230.5   1.3

60% packet loss? Yeah that is certainly a problem. Strange but there
is 0% loss to this device from Beverly Hills, CA.

>  20. g2-2-0.a1.hywr.broadwing.net               20.0%     5  231.2 230.3 227.9 232.4   1.9
>  21. 216.140.0.65                               20.0%     5  230.5 229.8 226.1 231.3   2.5
>  22. p2-0-0.a0.phnx.broadwing.net               20.0%     5  339.5 339.1 335.2 340.9   2.7
>  23. 216.142.38.22                              20.0%     5  304.7 309.5 304.7 313.0   3.8
>  24. git.gnome.org                              20.0%     5  303.0 308.2 303.0 313.0   4.8

Since you only sent 5 packets in that mtr output, can you run it for
about a minute and send us the output again?

Subversion supports a transparent write-through proxy[1] that would
get around some of the problems we are seeing. Oh git gods *cough
owen* does git have any support for something like this? Google
doesn't make it look so promising. Perhaps cloning from a
geographically local server and pushing to a remote at git.gnome.org?
I'm throwing out random ideas but this does seem like an issue.

[1] http://subversion.tigris.org/svn_1.5_releasenotes.html#webdav-proxy

-- 
Jeff Schroeder

Don't drink and derive, alcohol and analysis don't mix.
http://www.digitalprognosis.com