NodeBB ACP-access broken

[frissdiegurke]

I try to create a new NodeBB setup (branch master or v0.7.x).
Whenever I create a new admin via ./nodebb setup (since database was empty before) I CAN login but once I try to access /admin I automatically get logged out and my user is somehow broken so I cannot login anymore.
Happens with mongo and redis. No errors within development output.

[frissdiegurke]

@julian I just pulled the latest and tried again but this issue is still present.

Do you have any clue how to fix or may you provide redis-commands therefor I can give you more information what happens to the user-object.

Otherwise I sadly cannot setup local instance for testing (especially I wanted to try whether gh#1528 is still present).

And a command to drop whole redis-databases would be nice thought. Currently for each try I need use a new db-index...

psychobunny created this issue in NodeBB/NodeBB

closed [Settings] default values not being set for regular input types #1528

[julian]

@frissdiegurke flushdb in redis to drop just that database

What environment are you using?

Edit: I can't reproduce this -- the setup completes properly and I can log in and access the ACP normally...

[frissdiegurke]

I'm using redis 3.0.2, nodejs 0.12.7, npm 2.13.1, NodeBB commit 65ab5733ef8f6750ea710f490c8c92bea7721bbd and chromium 43.0.2357.134 on x86_64 Arch Linux 4.1.2-2.
Should all be latest within arch-repos.

After fresh git clone I ran

npm i
./nodebb setup
./nodebb dev # same problem when using ./nodebb start, etc.

When accessing NodeBB via browser now and logging in as the admin created within ./nodebb setup and accessing /admin afterwards the user gets somehow broken and I'm suddenly logged out.
I cannot login afterwards.

When using mongodb 3.0.4 (since I know it's cli) the objects-collection contains the following after trying to access /admin: (the user is called "frissdiegurke", for search ;)) http://pastebin.com/wvg7TKXk

maybe you see anything suspicious... I noticed that the user-status is still set to "online".

EDIT: here is the config.json:

{
    "url": "http://localhost:4567",
    "secret": "some uuid...",
    "database": "mongo",
    "mongo": {
        "host": "127.0.0.1",
        "port": "27017",
        "username": "nodebb",
        "password": "secret",
        "database": "nodebb"
    }
}

so everything quite default values...

[baris]

What happens when you try to login after you get logged out? 403? 500?

[frissdiegurke]

I've disabled the cache for this (request-header Cache-Control:no-cache) but without that the same result: 200.

General

Remote Address:[::1]:4567
Request URL:http://localhost:4567/login?local=1
Request Method:POST
Status Code:200 OK

Response Headers

Connection:keep-alive
Content-Length:6
Content-Type:text/html; charset=utf-8
Date:Sat, 18 Jul 2015 15:13:14 GMT
ETag:W/"6-EfazKFv3GoeBHD2tzLexZg"
set-cookie:express.sid=s%3AKZKFhtm5od-ur_JBgNkN1mZk-_MgogOF.r%2BamgtRRL1d%2Bun5NbhOxIuBW1W%2FJx7WAv9xp01UorXo; Path=/; Expires=Sat, 01 Aug 2015 15:13:14 GMT; HttpOnly
Vary:Accept-Encoding
X-Frame-Options:SAMEORIGIN
X-Powered-By:NodeBB

Request Headers

Accept:*/*
Accept-Encoding:gzip, deflate
Accept-Language:en-US,en;q=0.8,de;q=0.6
Cache-Control:no-cache
Connection:keep-alive
Content-Length:113
Content-Type:application/x-www-form-urlencoded; charset=UTF-8
Cookie:org.cups.sid=6592779d4f2a1725f4c410957b3ffef2; connect.sid=s%3AR0QEUPdPVSiS5tazLC8bkdhtLmjyYEnp.Q5Eucy7fsCVN22WViBWuZaWgmOzhFKrayNvnkdYuleY; express.sid=s%3AKZKFhtm5od-ur_JBgNkN1mZk-_MgogOF.r%2BamgtRRL1d%2Bun5NbhOxIuBW1W%2FJx7WAv9xp01UorXo
Host:localhost:4567
Origin:http://localhost:4567
Pragma:no-cache
Referer:http://localhost:4567/login?local=1
User-Agent:Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.134 Safari/537.36
x-csrf-token:CgqEiLEc-Cq3OXMXAyRURBWlCuo0Ep_tvsEU
X-Requested-With:XMLHttpRequest

Query String Parameters

local:1

Form Data

username:frissdiegurke
password:***
remember:on
returnTo:http://localhost:4567/login?local=1

I've also tried with other returnTo value (http://localhost:4567/tags) but same result 200 and redirect to /login?local=1

EDIT: After cleaning cookies (more specific express.sid) I can login again but as soon I try to access /admin it's the same all over again.

[frissdiegurke]

The request to /admin itself when logged in (and avatar is showed, etc.) results in a 302 as it seems:

General

Remote Address:[::1]:4567
Request URL:http://localhost:4567/admin
Request Method:GET
Status Code:302 Moved Temporarily

Response Headers

Connection:keep-alive
Content-Length:84
Content-Type:text/html; charset=utf-8
Date:Sat, 18 Jul 2015 15:41:47 GMT
Location:/login?local=1
set-cookie:express.sid=s%3ASnR2fDlqTEBDunA6DAyurgERQy9LvZwj.jmjluB9kDCE%2BSEhCtxTpzgcakpjJa%2BKOpuFYYSgtmIo; Path=/; Expires=Sat, 01 Aug 2015 15:41:47 GMT; HttpOnly
Vary:Accept, Accept-Encoding
X-Frame-Options:SAMEORIGIN
X-Powered-By:NodeBB

Request Headers

Accept:text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8
Accept-Encoding:gzip, deflate, sdch
Accept-Language:en-US,en;q=0.8,de;q=0.6
Cache-Control:no-cache
Connection:keep-alive
Cookie:express.sid=s%3ABq2PBi6iT5jaRQ6Qg3ojLy9PyWB0dRe6.VgpaSnS%2B%2BvdTxoJkxieZgw4ZpOWzUxJZOOH%2BWygCNqw; express.sid=s%3AYryeHEojKdlk3cc9SBmIegkusB9GeJ70.wJ5jXbD5oqvrygsWvUKMbq8mVJGIXrprt%2F%2F3OAH0jn0
Host:localhost:4567
Pragma:no-cache
Referer:http://localhost:4567/unread
User-Agent:Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.134 Safari/537.36

Maybe the two different express.sid request-cookies have sth to do with this o_O

[baris]

Can you try switching the session store to redis and see if that helps?

[frissdiegurke]

I've used redis in first place, just switched to mongodb for debugging purpose since I'm not familiar with redis commands.
So everything happens using redis the same way.

[julian]

Highly peculiar... and does this also happen on, say... a cloud install?

[baris]

@frissdiegurke Does this happen with node 0.10.x?

[frissdiegurke]

@baris said:

@frissdiegurke Does this happen with node 0.10.x?

Since there is no way to install node 0.10.x (neither iojs, wanted to try this ) without purging my current version I'm not willing to do this.

I've just tried to remove all global installed node_modules (rm -rf /usr/lib/node_modules), clean npm cache and reinstall npm and my global modules.
But it still didn't fix the problem (fixed some non-related errors though).

I'll try on my desktop-pc (also bleeding-edge arch-linux) later this day to see if it works there.

[julian]

You could try installing nvm, which would allow you to "install" Node from within your home directory...

Then you can switch between 0.10, 0.11, 0.12, and even iojs, with a single nvm use command.

Of course, you do have to npm rebuild each time, but that's not so bad.

[frissdiegurke]

@julian well, that's affordable didn't know nvm thanks!

Using nodejs v0.10.40 (npm 1.4.28) and npm rb within NodeBB root, flushing db and calling ./nodebb setup again... The issue still persists o_O

I've also tried the v0.6.x branch (using nodejs 0.12.x) where the same issue occurs, so it's definitely no fault within the NodeBB core (unless caused by lately bugfixes within that branch too)...

As I said I'll test on desktop-pc later on. Maybe I'll also try around with docker some time (never used it yet, definitely need to xD ).

[julian]

Eh, try LXC. I recommend it over Docker