Issues when testing Windows Update #37
Comments
|
this might be the issue for ms updates http://mailman.nginx.org/pipermail/nginx/2007-May/000961.html |
|
|
Change this line in nginx.conf: To this: |
|
ok will give a try this week and let you know |
|
still getting this when doing updates , note it looks like a windows 10 updates are the one having the issues , I do not get any errors when doing updates with windows 7 017/04/03 18:31:35 [crit] 688#0: *3450 bind(10.0.4.21) failed (22: Invalid argument) while connecting to upstream, client: 10.0.4.158, server: microsoft, request: "HEAD /d/msdownload/update/software/crup/2017/03/windows10.0-kb4013418-x64_e35782f290df71c9f00a52b0203a772b876d6e97.psf HTTP/1.1", upstream: "http://[2a01:111:2003::50]:80/d/msdownload/update/software/crup/2017/03/windows10.0-kb4013418-x64_e35782f290df71c9f00a52b0203a772b876d6e97.psf", host: "au.download.windowsupdate.com" |
|
Microsoft Windows Updates °|-lc-host-vint:7so I added this to the unbound.conf and it --- still testing !!! |
|
this is the ip that all the updates are coming from 13.107.4.50 it looks like a cdn for MS updates .. still trying to find out what the DNS name or cname is for this ip |
|
more on this IP Address |
|
more on this IP 13.107.4.50 https://www.robtex.com/dns-lookup/7.tlu.dl.delivery.mp.microsoft.com |
|
Names pointing to same IP address as this host name (45 shown) |
|
Let me know how it goes; I still haven't had the chance to set up my lancache as I'm debugging an error on my internet wich is causing random disconnects on bf1 :-S |
|
So what I see on the client pc running Windows 10 is this wireshark flags
reset set when that happens it ask to redownload and it looks in nginx in
the data tmp folder.
…On Apr 4, 2017 12:59 AM, "bntjah" ***@***.***> wrote:
Let me know how it goes; I still haven't had the chance to set up my
lancache as I'm debugging an error on my internet wich is causing random
disconnects on bf1 :-S
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#37 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/APwRgeh0Ofl8qQh0rXelYyFjqbkK2LNSks5rsdwlgaJpZM4MvJqX>
.
|
|
… On Apr 4, 2017 7:27 AM, "Travus Elm" ***@***.***> wrote:
So what I see on the client pc running Windows 10 is this wireshark flags
reset set when that happens it ask to redownload and it looks in nginx in
the data tmp folder.
On Apr 4, 2017 12:59 AM, "bntjah" ***@***.***> wrote:
> Let me know how it goes; I still haven't had the chance to set up my
> lancache as I'm debugging an error on my internet wich is causing random
> disconnects on bf1 :-S
>
> —
> You are receiving this because you authored the thread.
> Reply to this email directly, view it on GitHub
> <#37 (comment)>,
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/APwRgeh0Ofl8qQh0rXelYyFjqbkK2LNSks5rsdwlgaJpZM4MvJqX>
> .
>
|
|
… On Apr 4, 2017 8:13 AM, "Travus Elm" ***@***.***> wrote:
http://serverfault.com/questions/679705/why-is-my-web-server-dropping-
connections-with-a-tcp-reset-at-high-load
On Apr 4, 2017 7:27 AM, "Travus Elm" ***@***.***> wrote:
> So what I see on the client pc running Windows 10 is this wireshark flags
> reset set when that happens it ask to redownload and it looks in nginx in
> the data tmp folder.
>
> On Apr 4, 2017 12:59 AM, "bntjah" ***@***.***> wrote:
>
>> Let me know how it goes; I still haven't had the chance to set up my
>> lancache as I'm debugging an error on my internet wich is causing random
>> disconnects on bf1 :-S
>>
>> —
>> You are receiving this because you authored the thread.
>> Reply to this email directly, view it on GitHub
>> <#37 (comment)>,
>> or mute the thread
>> <https://github.com/notifications/unsubscribe-auth/APwRgeh0Ofl8qQh0rXelYyFjqbkK2LNSks5rsdwlgaJpZM4MvJqX>
>> .
>>
>
|
|
so at this point disabled proxybind in nginx config for mswindows and this is the error I am getting now 2017/04/04 14:14:31 [error] 859#0: *1 connect() to [2001:1900:2302:6f01::ff]:80 failed (101: Network is unreachable) while connecting to upstream, client: 192.168.30.211, server: microsoft, request: "GET /c/msdownload/update/sof tware/secu/2017/03/windows6.1-kb4012215-x64_31cc0bb569162c3ec6c1f992476f1ad9f58e0f16.psf HTTP/1.1", upstream: "http://[2001:1900:2302:6f01::ff]:80/c/msdownload/update/software/secu/2017/03/windows6.1-kb4012215-x64_31cc0bb569162c 3ec6c1f992476f1ad9f58e0f16.psf", host: "download.windowsupdate.com" |
|
ok did this and testing now Re: disable ipV6 solves lookup slowness for Debian but not for Ubuntu |
|
I missed where the resolver was set in a second place. Change this: To this: |
|
I did find the other location for ipv6=off in lancache resolver and did the change but it was still looking up ipv6 the errors only stopped when I did the command |
|
@nexusofdoom did this actually solve the issue? If so we'll add it to the
readme.md & installer but just want to make sure that this resolved the
problem
2017-04-05 13:19 GMT+02:00 nexusofdoom <notifications@github.com>:
… I did find the other location for ipv6=off in lancache resolver and did
the change but it was still looking up ipv6
the errors only stopped when I did the command
sudo sysctl net.ipv6.conf.all.disable_ipv6=1
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#37 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AGE_0fdCGP8R0r0_rM-W9_zDqdbcUH3oks5rs3i_gaJpZM4MvJqX>
.
|
|
still doing some tweaks for windows 10 and testing and testing :-) this is what i have so for for nginx settings #Windows Updates } once I am all done I will attache a tar with all the changes |
|
trying to get this folder to not cache set $no_cache ""; something that windows 10 is doing with updates i think its something like a update list it pushes and it pulls it down over and over there might also be a way to block it so still checking in ref url below |
|
more fun :-) Hello, |
|
so don't know if i need to remove this uri from caching? any one know how I could do this ? REQUEST="GET /filestreamingservice/files/3787205b-0105-4bfb-a64e-985d61a84e3b?P1=1491419171&P2=301&P3=2&P4=7L1llVny%2fZSfMprTfHkBkhZ0hiELzvopzaXFgiJ%2frIU%3d HTTP/1.1" - REQUEST_URI="/filestreamingservice/files/3787205b-0105-4bfb-a64e-985d61a84e3b?P1=1491419171&P2=301&P3=2&P4=7L1llVny%2fZSfMprTfHkBkhZ0hiELzvopzaXFgiJ%2frIU%3d" - URI="/filestreamingservice/files/3787205b-0105-4bfb-a64e-985d61a84e3b" |
|
After setting ipv6=off for both resolver definitions and restarting nginx I was able to update a fresh install of Windows 10 without seeing those errors about Invalid Argument. |
|
current setup #Windows Updates }Microsoft Windows Updates °|-lc-host-vint:7also did keep the ipv6=off in both configs -- and also have a "bind ipv6only=off" on the top of the lancahce-microsoft.conf |
|
wendelj if you are looking at nload do you still see it pulling down data..
|
|
if I use all the defaults along with all the ipv6 disabled not every update for windows comes over the lancache it routes over the internet direct |
|
going to attache wireshark logs |
|
dns wireshark file csv on pc that was doing windows updates |
|
lancache config ms office #Windows Updates } |
|
when getting new updates I get this error 8a385923ff85ddc95ecd75ea41.psf HTTP/1.1", upstream: "http://8.254.220.190:80/c/msdownload/update/software/secu/2017/03/windows10.0-kb4013198-x64_1_b8835f35c094b78a385923ff85ddc95ecd75ea41.psf", host: "download.windowsupdate.com" |
|
trying this setup again #Windows Updates } |
|
got a message like this unexpected status code 400 in slice response while reading response header from upstream [error] 5801#0: *954 unexpected status code 400 in slice response while reading response header from upstream, client: 192.168.30.150, server: microsoft, request: "GET /c/msdownload/update/software/secu/2017/04/windows6.1-kb4015549-x64_bf8033c1dbede1fcd8b09e29328fddea5c44f8f4.psf HTTP/1.1", subrequest: "/c/msdownload/update/software/secu/2017/04/windows6.1-kb4015549-x64_bf8033c1dbede1fcd8b09e29328fddea5c44f8f4.psf", upstream: "http://13.107.4.50:80/c/msdownload/update/software/secu/2017/04/windows6.1-kb4015549-x64_bf8033c1dbede1fcd8b09e29328fddea5c44f8f4.psf", host: "au.download.windowsupdate.com" |
|
maybe try this ? that is, for users of ngx_http_proxy_module In addition to fastcgi, the proxy module also saves the request header in a temporary buffer. So you may need also to increase the proxy_buffer_size and the proxy_buffers, or disable it totally (Please read the nginx documentation). Example of proxy buffering configuration http { http { |
|
# sudo service nginx status Apr 11 19:45:37 lancache-server lancache[6088]: Starting : nginx: [warn] could not build optimal proxy_headers_hash, you should increase either proxy_headers_hash_max_size: 512 or pr..._bucket_size there is a "warn" about proxy header hash sizechange this in my ms-config "proxy_headers_hash_bucket_size 512;" #Windows Updates |
|
so i see it still does try to download over and over can we try this ? http://serverfault.com/questions/526441/nginx-cant-handle-multiple-downloads also should the file name be ?windows6.1-kb4015549-x64_bf8033c1dbede1fcd8b09e29328fddea5c44f8f4.psf" - KEY="microsoft/c/msdownload/update/software/secu/2017/04/windows6.1-kb4015549-x64_bf8033c1dbede1fcd8b09e29328fddea5c44f8f4.psf when it downloads ? |
|
https://www.mail-archive.com/nginx-devel@nginx.org/msg06079.html ?? added this to the ms config *1230 etag mismatch in slice response while reading response header from upstream, client: 192.168.30.126, server: microsoft, request: "GET /msdownload/update/software/crup/2011/05/windows6.1-kb2533552-x64_fabe284ceaa1403debb09fad96cd5c0806f08648.psf HTTP/1.1", subrequest: "/msdownload/update/software/crup/2011/05/windows6.1-kb2533552-x64_fabe284ceaa1403debb09fad96cd5c0806f08648.psf", upstream: "http://8.254.250.190:80/msdownload/update/software/crup/2011/05/windows6.1-kb2533552-x64_fabe284ceaa1403debb09fad96cd5c0806f08648.psf", host: "download.windowsupdate.com" |
|
do we need to have proxy_pass ? https://www.liaohuqiu.net/posts/nginx-proxy-pass/ |
|
any issue whit this setup that any one can see ? #Windows Updates } |
|
can some one test this config I tested with windows 7 and windows 10 #Windows Updates } |
|
this might be a issue |
|
2017/04/15 07:49:19 [error] 1806#0: *555 upstream prematurely closed connection while reading upstream, client: 10.0.4.146, server: microsoft, request: "HEAD /msdownload/update/software/crup/2011/05/windows6.1-kb2533552-x64_fabe284ceaa1403debb09fad96cd5c0806f08648.psf HTTP/1.1", upstream: "http://8.254.250.78:80/msdownload/update/software/crup/2011/05/windows6.1-kb2533552-x64_fabe284ceaa1403debb09fad96cd5c0806f08648.psf", host: "download.windowsupdate.com" |
|
upstream prematurely closed connection while reading upstream |
|
so this looks like a issue. Too small listen queue on a backend resulting in occasional listen queue overflows (Linux is usually configured to just drop SYN packet in this case, thus making it indestinguishable from a packet loss; see netstat -s | grep listen to find out if it's the problem). 1 times the listen queue of a socket overflowed |
|
working on tweaking nginx https://www.linode.com/docs/web-servers/nginx/configure-nginx-for-optimized-performance/ |
|
so what I am running into is fresh install windows 7 sp1 has 194 updates |
|
so are we not doing this any more ? |
might want to add this also
|
|
another config working with map $upstream_http_content_length $flag_cache_empty { #Windows Updates proxy_set_header Host http://download.windowsupdate.com; access_log /srv/lancache/logs/Access/microsoft.log main buffer=128k flush=1m; location / { proxy_set_header X-Real-IP $remote_addr; } #################################################### 2017/04/15 17:13:30 [error] 5761#0: *434 upstream prematurely closed connection while reading upstream, client: 10.0.4.122, server: microsoft, request: "GET /d/msdownload/update/software/secu/2017/04/windows10.0-kb4015217-x64_3_075b7dc98dbdf6ef5bc42128cb7b23cdeb8e1291.psf HTTP/1.1", upstream: "http://8.253.185.121:80/d/msdownload/update/software/secu/2017/04/windows10.0-kb4015217-x64_3_075b7dc98dbdf6ef5bc42128cb7b23cdeb8e1291.psf", host: "download.windowsupdate.com" |
|
will it still cache the updates ? When buffering is disabled, a response is passed to a client synchronously, as soon as it is received from the back-end. nginx will not try to read the whole response from the proxied server. The maximum size of the data that nginx can receive from the server at a time is set by the proxy_buffer_size directive. proxy_buffering off; https://tweaked.io/guide/nginx-proxying/ will it still cache the files ? |
|
--------------------- the working config that I am using at my office ------------- map $upstream_http_content_length $flag_cache_empty { #Windows Updates listen lancache-microsoft deferred default; proxy_ignore_client_abort on; proxy_http_version 1.1; proxy_hide_header Etag; proxy_set_header Host http://download.windowsupdate.com; proxy_cache_key "$scheme$proxy_host$request_uri $http_range"; #proxy_buffering off; access_log /srv/lancache/logs/Access/microsoft.log main buffer=128k flush=1m; location / { } |
|
Closing this as it works now and config is adapted as posted by Nexusofdoom |
[crit] 686#0: *271 bind(10.0.4.21) failed (22: Invalid argument) while connecting to upstream, client: 10.0.4.179, server: microsoft, request: "GET /d/msdownload/update/software/updt/2017/03/windows10.0-kb4015438-x64_3_aadae06fd0885db7996e702e453f784b4683d978.psf HTTP/1.1", upstream: "http://[2001:1900:2302:2f04::ff]:80/d/msdownload/update/software/updt/2017/03/windows10.0-kb4015438-x64_3_aadae06fd0885db7996e702e453f784b4683d978.psf", host: "download.windowsupdate.com"
2017/03/30 21:27:44 [crit] 692#0: *274 bind(10.0.4.21) failed (22: Invalid argument) while connecting to upstream, client: 10.0.4.179, server: microsoft, request: "GET /d/msdownload/update/software/updt/2017/03/windows10.0-kb4015438-x64_1_9881d994bf0f17a1120e12723450b6e54e499075.psf HTTP/1.1", upstream: "http://[2001:1900:2304:8f0a::ff]:80/d/msdownload/update/software/updt/2017/03/windows10.0-kb4015438-x64_1_9881d994bf0f17a1120e12723450b6e54e499075.psf", host: "download.windowsupdate.com"
2017/03/30 21:27:51 [crit] 692#0: *269 bind(10.0.4.21) failed (22: Invalid argument) while connecting to upstream, client: 10.0.4.179, server: microsoft, request: "HEAD /d/msdownload/update/software/updt/2017/03/windows10.0-kb4015438-x64_3_aadae06fd0885db7996e702e453f784b4683d978.psf HTTP/1.1", upstream: "http://[2001:1900:2304:af07::ff]:80/d/msdownload/update/software/updt/2017/03/windows10.0-kb4015438-x64_3_aadae06fd0885db7996e702e453f784b4683d978.psf", host: "download.windowsupdate.com"
The text was updated successfully, but these errors were encountered: