• speedtest ok from site, fails on command line

    From Roger Price@21:1/5 to All on Fri May 9 18:00:01 2025
    If I visit speedtest.net and click on "Go", I get 356Mbps download, and 19Mbps upload with multiconnection via Starlink and a server in Paris. But when I load
    Debian 12 package speedtest-cli I get

    rprice@maria ~ speedtest
    Retrieving speedtest.net configuration...
    Testing from Starlink (216.128.23.xxx)...
    Retrieving speedtest.net server list...
    Selecting best server based on ping...
    ERROR: Unable to connect to servers to test latency.

    I get this same error result all day and from Debian 11 and 12. Has anyone got speedtest-cli to work?

    Roger

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From The Wanderer@21:1/5 to Roger Price on Fri May 9 18:10:01 2025
    This is an OpenPGP/MIME signed message (RFC 4880 and 3156)
    On 2025-05-09 at 11:57, Roger Price wrote:

    If I visit speedtest.net and click on "Go", I get 356Mbps download, and 19Mbps
    upload with multiconnection via Starlink and a server in Paris. But when I load
    Debian 12 package speedtest-cli I get

    rprice@maria ~ speedtest
    Retrieving speedtest.net configuration...
    Testing from Starlink (216.128.23.xxx)...
    Retrieving speedtest.net server list...
    Selecting best server based on ping...
    ERROR: Unable to connect to servers to test latency.

    I get this same error result all day and from Debian 11 and 12. Has anyone got
    speedtest-cli to work?

    This looks like bug #1024830.

    --
    The Wanderer

    The reasonable man adapts himself to the world; the unreasonable one
    persists in trying to adapt the world to himself. Therefore all
    progress depends on the unreasonable man. -- George Bernard Shaw


    -----BEGIN PGP SIGNATURE-----

    iQIzBAEBCgAdFiEEJCOqsZEc2qVC44pUBKk1jTQoMmsFAmgeJ10ACgkQBKk1jTQo Mmvorg//UnksOAmvxA1nN6Cb3i+LrzNhmGkjfc33K1XOJAgpaftlRkT9ok9BQVfM PNR6XjAN9ewUe5TtuM7lZXGNmnWWCsCAT+vymmFHXgNX9rb30mpMP17+w9Z5ZwmC 2yCmuAEzYf+l2JmEfzpvB1484ftnkSGGohQZDzI4yMSiX38fnic6NANzieQpNENS BZKW5viqW+i/gc1K+ZD/32ofeSsLuYwgfZUL9g+iK4nYecbtv9tG7/QDid4QHp7n mUU0F0tscbAa/Wt2MIXhJ/m5g+BKQSpHplGed1f96QGU3XG/JzeCLXhD2SBPzuSh t5HXeNWZzuPSalS7q824IJR0BFiAhI21ya4LLvZpBAhXj/e/Y94fQFma17HtFFRV vlNuAHFYPCiHfcUnekbiaKZntW0uSE1LOdaa6It5+6aow9fmZbQvakeLaYbmGQ+r dDOITsbGrO+ud20FRWr4rrVcASZ8y+GekymP6RQGE3SEDmc8g4kpoFDPHO1u2w5z InDKkH3RoaB3pwV6tZjfy0t9bH4vUT/G7IwYpeMd4tS8wW9jzFRHfhwaeKDLXI08 ANO7tH+vDAVN9B2H6Lwc/ruGSdPbwDSRdOfxHZsEIIUswJx7XOTpoxKdSbDgFo+k VhdW3Hf72ybbNOobOYaytMglfype
  • From Roger Price@21:1/5 to The Wanderer on Fri May 9 19:40:01 2025
    On Fri, 9 May 2025, The Wanderer wrote:

    This looks like bug #1024830.

    On Fri, 9 May 2025, Kent West wrote:

    westk@westkent:~$ speedtest

    Download: 496.67 Mbit/s
    VERSION_CODENAME=trixie

    So it looks as if this will be fixed in Debian 13. Roger

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From The Wanderer@21:1/5 to Roger Price on Fri May 9 21:50:01 2025
    This is an OpenPGP/MIME signed message (RFC 4880 and 3156)
    On 2025-05-09 at 13:32, Roger Price wrote:

    On Fri, 9 May 2025, The Wanderer wrote:

    This looks like bug #1024830.

    On Fri, 9 May 2025, Kent West wrote:

    westk@westkent:~$ speedtest

    Download: 496.67 Mbit/s
    VERSION_CODENAME=trixie

    So it looks as if this will be fixed in Debian 13. Roger

    What leads you to that conclusion?

    My reading of the bug report I referenced is that the issue is that the
    server is rejecting connections because it's receiving too many
    connections too fast from the same IP address.

    There is a link to a GitHub pull request (from 2023, I think) which
    looks like it's supposed to do something about the problem - but that
    pull request doesn't seem to be merged as far as I could see, the bug
    report appears to still be open, and I don't see any sign there that the problem has been fixed for trixie.

    Based on that, I would interpret the fact that it works for some people
    and not others as a coincidence relating to what the server thinks it
    knows (and/or is currently seeing) about the specific place it sees the
    request as coming from.

    --
    The Wanderer

    The reasonable man adapts himself to the world; the unreasonable one
    persists in trying to adapt the world to himself. Therefore all
    progress depends on the unreasonable man. -- George Bernard Shaw


    -----BEGIN PGP SIGNATURE-----

    iQIzBAEBCgAdFiEEJCOqsZEc2qVC44pUBKk1jTQoMmsFAmgeWzcACgkQBKk1jTQo Mmv/rg//YU9UMk23yP72BRMqxwBZw4PprFNDZddP7v6kKBz+nnRWAjuL+AZgRrtd 6wdglIoTnu+ut2+0nGG9h8JYH2bN1UN6R0fL7PPBlYcJQkrHKXB/jDwVuQK1QVaD 2fpquEEr4mG90+G6C3gUpT8urE/JJKPCju/iNo5FYhEeFMo4WPP7u6Mm5AxNatiU J8d3ArLJRel6sYMdzicQEKi40HJ4hqnJuTJpuqFuQYyjPS3QpHgtvqfkXCKXgJof grbGicnKAVg+LHVVzF5a52Yk0vCqrEBXDg8SAjirKZx+sPcMV/V/U5b4obph/T2z qcwWZj13+SbF7NLO3h+vkaEW9iwxyJac6RTCrc4uUhysOGtr4pcJH8zCoOk63Y8t +lET4t7wGcsx5ExTC/yvWsk+gAQfIPzrh62/1v9HPJRGkt6mcYMXymPS0WrvyPkG 4PDIvjU+6vL+oqAPwFF+btIOyZWwTfLgV0f0PjisiCP94hAKZVH6rIn88/oh6ISR Cv5vOmWe1SZmum+vSFAz/5QF5sDO6+8yEMro+pT7LXS0n0kKWYBNSGCePeaayuZp rofnk7FuB8X9B8yviPfdGLnmKNpSShUlMWMAdG9jZcs0+SvxAqe650JiItUffL3m I4Z53EctLNTxZ0pNFTiEb/gH7cgc
  • From Roger Price@21:1/5 to The Wanderer on Fri May 9 22:30:01 2025
    On Fri, 9 May 2025, The Wanderer wrote:

    So it looks as if this will be fixed in Debian 13. Roger

    What leads you to that conclusion?

    My reading of the bug report I referenced is that the issue is that the server is rejecting connections because it's receiving too many
    connections too fast from the same IP address.

    Then why would speedtest work correctly from a browser connected to the speedtest.net site, but not from speedtest-cli coming from the same IP ? My Starlink IP does vary, possibly every 4 minutes, but that doesn't explain the persistent difference in behaviour between the site and the command line. Roger

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From The Wanderer@21:1/5 to Roger Price on Fri May 9 22:50:01 2025
    This is an OpenPGP/MIME signed message (RFC 4880 and 3156)
    On 2025-05-09 at 16:26, Roger Price wrote:

    On Fri, 9 May 2025, The Wanderer wrote:

    So it looks as if this will be fixed in Debian 13. Roger

    What leads you to that conclusion?

    My reading of the bug report I referenced is that the issue is that
    the server is rejecting connections because it's receiving too
    many connections too fast from the same IP address.

    Then why would speedtest work correctly from a browser connected to
    the speedtest.net site, but not from speedtest-cli coming from the
    same IP ? My Starlink IP does vary, possibly every 4 minutes, but
    that doesn't explain the persistent difference in behaviour between
    the site and the command line. Roger

    I don't have the full story, but that's what the error code described in
    the bug report apparently means.

    Looking at the patch in the pull request, what it apparently does in an
    attempt to prevent that error code from occurring is to pull the list of servers from a different URL (and, I'm guessing, a different API). If
    that does in fact fix it, then the error code is misleading, and the
    issue is that the method the CLI tool is using to figure out what
    server(s) to contact is no longer correct.

    In that case, the reason why it works from the Website but not from the
    CLI tool would probably be because the Website backend is looking at the correct server list, and the CLI tool is looking at the wrong one. (And, potentially, some of the servers that can be contacted based on the
    wrong list give this error, and some don't.)

    I do not even use this service, much less this tool. I am going entirely
    based on what I found via reportbug for the package with the tool in it,
    and then the information I found (and/or found linked to from) there. My analysis could be correct, or it could be entirely off-base. I just
    didn't want to leave the conclusion that "it's fixed in trixie"
    unaddressed, when that did not seem to me to be the case.

    --
    The Wanderer

    The reasonable man adapts himself to the world; the unreasonable one
    persists in trying to adapt the world to himself. Therefore all
    progress depends on the unreasonable man. -- George Bernard Shaw


    -----BEGIN PGP SIGNATURE-----

    iQIzBAEBCgAdFiEEJCOqsZEc2qVC44pUBKk1jTQoMmsFAmgeaOQACgkQBKk1jTQo MmuSFw/+O9k84R6rgQuOfvjGtVPTITmO++ZIWRDgXQxALAm22XXLCN2cObwbbVdT KA56VfNK+y5skD6cJARoM7hHPEf0Xw7knXlNiG6MDYjjoLkayU/5sMmIEjXqD3t9 NbHBz+LACwxgFrcHgACRO+5OCkP8Gv6HIzf7KISpmlcnbV6uHun10v0hGABWdrcZ Hatzrq8PWyd93pXBlMEJUIbvxu9T6qifl5K2IGLLkkPBy2PGcq6klg2vZRAOOhJR lOMIQWu8YYT9BOgOCDyVa9nYcapnxkOScIZGpG9ddVhRRmFxbFs8p0TIo+lJsVBA 5XcVihpDwUCTBO13MkWVyaKU0GE7J5mkEvbjlcLkIvvX6S91j88sPs0gaQ90xS0w RrAcl1cOfH7ceX3QQZuycx1fcOiUzKCPUBNo4CYd+6Sphqr/AQ/BGtF51+dTIU0e dRtmUaGPUJI+fIkYfUIylJ+8MW5rmB56AYHThqZDjcPrgsXdB/qz+rvy5iaDg+pV ER6P4LQfcLx3ewfT7QZs8NzaGuKRvMmXNcwA7gnjJmJriq9THpU220Vn7rXp/ww3 NXhAago8RJ7fQEyeyU1XGfGilAdbx4rFiRSR/uvKSGNySEZFVS6bouNrHnQcA24C Xjgt94BczegAEzwV+UIOQ2aS42CO
  • From Michael Stone@21:1/5 to Roger Price on Sat May 10 16:10:01 2025
    On Fri, May 09, 2025 at 10:26:49PM +0200, Roger Price wrote:
    Then why would speedtest work correctly from a browser connected to the >speedtest.net site, but not from speedtest-cli coming from the same IP ? My >Starlink IP does vary, possibly every 4 minutes, but that doesn't explain the >persistent difference in behaviour between the site and the command line. Roger

    Did you try the --secure option?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Roger Price@21:1/5 to Michael Stone on Sat May 10 16:50:01 2025
    This message is in MIME format. The first part should be readable text,
    while the remaining parts are likely unreadable without MIME-aware tools.

    On Sat, 10 May 2025, Michael Stone wrote:

    On Fri, May 09, 2025 at 10:26:49PM +0200, Roger Price wrote:
    Then why would speedtest work correctly from a browser connected to the speedtest.net site, but not from speedtest-cli coming from the same IP ?

    Did you try the --secure option?

    Yes, it made no difference. However by looking at what the speedtest-cli python3
    does, and imitating this with wget I got a clearer diagnostic:

    rprice@maria ~ wget http://c.speedtest.net/speedtest-servers-static.php
    --2025-05-10 10:30:13-- http://c.speedtest.net/speedtest-servers-static.php
    Resolving c.speedtest.net (c.speedtest.net)... 151.101.2.219, 151.101.66.219,
    151.101.194.219, ...
    Connecting to c.speedtest.net (c.speedtest.net)|151.101.2.219|:80... connected.
    HTTP request sent, awaiting response... 429 Too Many Requests
    2025-05-10 10:30:13 ERROR 429: Too Many Requests.

    My previous guess that the bug would be fixed in Debian 13 looks to be wrong. The real problem seems to be that Ookla doesn┤t like Starlink, at least those Starlink users who land in Paris.

    Roger

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Michael Stone@21:1/5 to Roger Price on Sun May 11 21:50:02 2025
    On Sat, May 10, 2025 at 04:46:45PM +0200, Roger Price wrote:
    2025-05-10 10:30:13 ERROR 429: Too Many Requests.

    Perhaps a wishlist bug for speedtest to display that error so people
    can understand what's going on.

    My previous guess that the bug would be fixed in Debian 13 looks to be wrong. >The real problem seems to be that Ookla doesn┤t like Starlink, at least those >Starlink users who land in Paris.

    Sounds like the real problem is too many people trying to test the speed
    of their starlink connection.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)