Many upstream related errors


  • | 1556 points

    I've been looking around for clues on what could be causing this, but I haven't found anything much conclusive. I'm working on this simple L5 app that will be consumed by no more than 30 some clients. A couple of such clients will be hitting the server frequently (kiosk apps) while all the others will be making random human-initiated requests (tablets). Right now the aggregate request load is quite small, but nevertheless Laragon/nginx/php seems to be having some trouble handling this. This is what I'm getting at nginx's error log:

    2018/08/27 21:13:52 [error] 11168#9512: *5431 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.150, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/lanche HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:14:28 [error] 11168#9512: *5482 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:15:07 [error] 11168#9512: *5535 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:15:46 [error] 11168#9512: *5588 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:16:25 [error] 11168#9512: *5639 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.150, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/lanche HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:17:04 [error] 11168#9512: *5694 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:17:43 [error] 11168#9512: *5747 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:18:22 [error] 11168#9512: *5800 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:19:01 [error] 11168#9512: *5853 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:19:43 [error] 11168#9512: *5910 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:20:25 [error] 11168#9512: *5965 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.150, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/lanche HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:21:04 [error] 11168#9512: *6000 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:21:43 [error] 11168#9512: *5998 WSARecv() failed (10054: FormatMessage() error:(15105)) while reading response header from upstream, client: 10.100.98.150, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/lanche HTTP/1.1", upstream: "fastcgi://127.0.0.1:9012", host: "10.100.98.141:8080"
    2018/08/27 21:21:46 [error] 11168#9512: *5994 upstream timed out (10060: FormatMessage() error:(15105)) while connecting to upstream, client: 10.100.98.150, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/lanche HTTP/1.1", upstream: "fastcgi://127.0.0.1:9019", host: "10.100.98.141:8080"
    2018/08/27 21:21:46 [error] 11168#9512: *5996 upstream timed out (10060: FormatMessage() error:(15105)) while connecting to upstream, client: 10.100.98.149, server: bein_energisa-museu-ms-intranet.local, request: "GET /api/setup/folder/graaf HTTP/1.1", upstream: "fastcgi://127.0.0.1:9020", host: "10.100.98.141:8080"

    A whole bunch of WSARecv errors, all complaining of not being able to properly read the response header along with a few upstream timeouts. Just a while ago the upstream timeout errors where in much greater number, but since I've increased nginx's Upstream value to 25 (usr\laragon.ini), they appear less frequently.

    I'm really out of my wits here trying to find the real issue. Perhaps some other tweaking needs to be done (besides just increasing the Upstream value), but I'm not sure what and why. Any help will be greatly appreciated.


  • administrators
    | 89600 points

    @mpenna : The default nginx.conf setting is for development. How about adding these configuration in Menu > Nginx > nginx.conf, under http:

    http {
        fastcgi_buffers 16 16k;
        fastcgi_buffer_size 32k;
        proxy_buffer_size   128k;
        proxy_buffers   4 256k;
        proxy_busy_buffers_size   256k;
    ...
    ...
    ...

  • | 1556 points

    Thanks. I have applied the suggested configuration and it does seem that the number and frequency of events in the error log have diminished, but nonetheless they are still occurring and being logged.


Log in to reply
 

Looks like your connection to Laragon was lost, please wait while we try to reconnect.