Unable to login to piccolo_admin in production

[arunk]

Hi, I managed to setup and get everything running perfectly in my development environment. But when I moved the code to production, I can see the site running, there are no errors in the log files. I have also made changes in production such as adding the allowed_hosts kwarg to the create_admin method, and specified the domain that the site is on. I have also passed production: True to create_admin, I have also tried to turn debug: True just to see if will give some additional information somewhere. But whatever I've tried, I've been unable to login to the site. I'm able to check the users from running piccolo user ls and other commands, I've even changed password just to make sure it wasn't a fat finger issue. The user I'm trying to logging as has admin, superadmin and is active.
The only error I see is a little orange bar at the bottom of the screen that says Problem logging in .
Also attaching a screenshot of the error:

[arunk]

Ok found something in the browser console:

Request failed index.eebd4d8d.js:79:53332
{
  "data": "Client host can't be found.",
  "status": 400,
  "statusText": "Bad Request",
  "headers": {
    "connection": "keep-alive",
    "content-length": "27",
    "date": "Sat, 05 Oct 2024 08:24:05 GMT",
    "server": "nginx/1.18.0 (Ubuntu)"
  },
  "config": {
    "transitional": {
      "silentJSONParsing": true,
      "forcedJSONParsing": true,
      "clarifyTimeoutError": false
    },
    "adapter": [
      "xhr",
      "http",
      "fetch"
    ],
    "transformRequest": [
      null
    ],
    "transformResponse": [
      null
    ],
    "timeout": 0,
    "xsrfCookieName": "XSRF-TOKEN",
    "xsrfHeaderName": "X-XSRF-TOKEN",
    "maxContentLength": -1,
    "maxBodyLength": -1,
    "env": {},
    "headers": {
      "Accept": "application/json, text/plain, */*",
      "Content-Type": "application/json",
      "X-CSRFToken": "xxx-xxx-xxx-xxx"
    },
    "method": "post",
    "url": "./public/login/",
    "data": "{\"username\":\"arun\",\"password\":\"xxxxx\"}"
  },
  "request": {}
}
Logging in

I have crossed out my password, and the tokens, but the information there is all correct. And I'm logging in from a Firefox browser, in case that makes any difference.

[arunk]

Also tried Chromium and got the same error.

[arunk]

Didn't notice one more console log, it was higher up in the console. AxiosError:

{
    "message": "Request failed with status code 401",
    "name": "AxiosError",
    "stack": "AxiosError: Request failed with status code 401\n    at yl (https://z69.in/admin/assets/index.eebd4d8d.js:3:1055)\n    at XMLHttpRequest.u (https://z69.in/admin/assets/index.eebd4d8d.js:3:6125)\n    at dr.request (https://z69.in/admin/assets/index.eebd4d8d.js:5:1965)\n    at async te.fetchUser (https://z69.in/admin/assets/index.eebd4d8d.js:103:30723)",
    "config": {
        "transitional": {
            "silentJSONParsing": true,
            "forcedJSONParsing": true,
            "clarifyTimeoutError": false
        },
        "adapter": [
            "xhr",
            "http",
            "fetch"
        ],
        "transformRequest": [
            null
        ],
        "transformResponse": [
            null
        ],
        "timeout": 0,
        "xsrfCookieName": "XSRF-TOKEN",
        "xsrfHeaderName": "X-XSRF-TOKEN",
        "maxContentLength": -1,
        "maxBodyLength": -1,
        "env": {},
        "headers": {
            "Accept": "application/json, text/plain, */*"
        },
        "method": "get",
        "url": "./api/user/"
    },
    "code": "ERR_BAD_REQUEST",
    "status": 401
}

[arunk]

So the follow up is I tried to setup ProxyHeaderMiddleware inside my blacksheep app, but I don't think I'm doing it right. It isn't working. I have asked for help in the Blacksheep matrix channel. Lets see if they can help. Also from this - https://www.uvicorn.org/deployment/#unix-domain-sockets-uds it sounds like there is an issue with Unix Domain Sockets especially when it comes to uvicorn. So maybe I'm best off running it on a TCP/IP port.

[dantownsend]

Yes, I would try running it on a port. That has always worked well for me.

[arunk]

So just an update on this, running it on a localhost port is working fine. It would be nice to debug this issue further and figure out why it is not working. My investigations lead me to Python's standard module - asyncio and its Transport class. Now its one thing to fork some libraries but obviously it won't be possible to do that with Python itself. It will be a longer process to submit changes. But again I don't know if it is a problem with the asyncio module and the way it handles sockets, or the way that uvicorn is looking for information from the remote connection using the library. Anyway, I think I'll just be happy with running it on a IP address and port and worry about running it on a unix socket some other time. Thanks for your help and time.

[dantownsend]

OK, glad you got to the bottom of it. I haven't tried running it using sockets before.

[arunk]

From what I understand running servers like this on a domain socket, versus a network socket there is a big difference in performance. The domain socket is able to handle a lot more data in and out. At some point I'll revisit this issue and see if I can figure out what is going on.