DNS problem (docker-toolbox \ Win7x64 \ non-admin user)

[John-MEO-o]

Hi,
I've already complete setup & using devilbox tool on OS X
(it means that i understand how it works fine)

But on Windows7 x64 host i have a following issue.
Accordingly to documentation, we need to specify a "LOCAL_LISTEN_ADDR",
which is 127.0.0.1 by default. Docker tools software versions provided in attachment image.
In case of windows 7 we can use only docker-toolbox, and cannot use a docker-native client...
While it starts - it shows an IP 192.168.99.100
which must replace a default IP address (127.0.0.1:) from .env file.

Once I've set that - i'm launching the utility...
Starting procedure are the same as on OS X...
The tool sequentially launches a few web services and goes to idle state...

Afterwards my browser (Google chrome Version 60.0.3112.113 (Official Build) (64-bit))
can't get to the main web-panel, neither one of following addresses are recognized by DNS
127.0.0.1
192.168.99.100
172.16.238.11
localhost

I've tries to set all of them to "LOCAL_LISTEN_ADDR" value,
each time after any changes being made I've erase all present containers through following commands:
docker rm -f $(docker ps -a -q)
docker rmi $(docker images -q)
for now it is still doesn't see web-panel...

Who is sharp enough, please advise - what it can be? Is there any solution for that?
Thanks for paying attention, good luck

Software versions are

Working idle state of devilbox tool

[cytopia]

@jayson-lynnfield-zen unfortunately it is still doesn't support Docker Toolbox. If you have any change to use Windows 10, go for it, otherwise you will have to wait a bit, until I (or someone else) have found a way.

[John-MEO-o]

@cytopia thanks for reply

[cytopia]

@jayson-lynnfield-zen
Could you set LOCAL_LISTEN_ADDR=127.0.0.1: to LOCAL_LISTEN_ADDR= in .env and try to access it again?

By default the devilbox listens on 127.0.0.1, but as Docker Toolbox is an independent VirtualBox (another machine), you want to access it from the outside, so you must allow the devilbox to bind on all interfaces, hence leave the variable empty to accomplish it.

[John-MEO-o]

@cytopia you mean live it blank (empty)?
OK, I'll try, write about it soon

[John-MEO-o]

@cytopia, i've found one weird moment...
Once i've create internal service-dependent directories (www, mysql, pgsql and mongo)
and then get starting devilbox-tool, this folders are not filling by any of temporary working files... is that normal state?

-- My first attempt was a blank address,
such like: LOCAL_LISTEN_ADDR= web-interface inaccessible (output log in attachement)

-- Then i've try to add a colon mark,
like that: LOCAL_LISTEN_ADDR=: fails to start
-- Than i get rid of google's DNS (left only my own DNS Server), and tries to set the my own subnet address with colon-mark at the end,
such as: LOCAL_LISTEN_ADDR=192.168.0.0: web-interface inaccessible...
-- At the final attempt, i've set predefined ip (from router) and its still fails to start
Each time i've erase all containers, just like in documentation.

So maybe i'm getting wrong somewhere of the setup process,
because in OS X internal folders are filled by service-temporary files right after devilbox launches all containers.
My firewall is disabled, no any of AV software being started at that moment.
I'm logged as a standard user (non-admin)

[cytopia]

@jayson-lynnfield-zen can you paste the full log for httpd from your first try with an empty LOCAL_LISTEN_ADDR (no colon):

docker-compose up -d
docker-compose logs httpd

[John-MEO-o]

@cytopia yep, I'll send it, when get to the pc...
Through 4-5 hours or so

[John-MEO-o]

@cytopia i've execute that,results are below

[cytopia]

@jayson-lynnfield-zen I made a major rewrite to overcome some limitations. I hope I could also address your problems. Pull request is not merged yet (#130), however could you checkout the branch (REL-0.12) and give it a try:

1. just to make sure everything is reset

docker-compose down
docker-compose rm -f

2. Check out new branch

git fetch --all
git checkout REL-0.12

3. Compare new env-example with your local .env

vimdiff .env env-example

4. Start it up

docker-compose up -d

Hope it can fix some problems

[John-MEO-o]

Thanks man, I'll check it as fast as I can, write you later Best regards, Jason E. Отправлено из ProtonMail mobile

…

-------- Исходное сообщение -------- Вкл 28 сент. 2017 г., 2:26, cytopia написал(а):

***@***.***(https://github.com/jayson-lynnfield-zen) I made a major rewrite to overcome some limitations. I hope I could also address your problems. Pull request is not merged yet ([#130](#130)), however could you checkout the branch and give it a try: - just to make sure everything is reset docker-compose down docker-compose rm -f - Check out new branch git fetch --all git checkout REL-0.12 - Compare new env-example with your local .env vimdiff .env env-example - Start it up docker-compose up -d Hope it can fix some problems — You are receiving this because you were mentioned. Reply to this email directly, [view it on GitHub](#119 (comment)), or [mute the thread](https://github.com/notifications/unsubscribe-auth/AZjY4Gm8-WoL7mjm9gXUiqsv_9QEMbEzks5smr12gaJpZM4PGkxi).

[John-MEO-o]

@cytopia
I've downloaded latest version, it gives an error right from starting... (See attached file)
i've set dns server to my own, than to Google's, afterwards i change IP to null, and to docker-toolbox IP with no any results. Firewall was disabled.

[cytopia]

@jayson-lynnfield-zen

There seems to be two docker networks. One from your previous devilbox instance and the new one. However both do use the same IP.

So do the following (First shutdown the devilbox):

# show devilbox networks
docker network ls --filter 'name=devil*' 

# delete all networks starting by devilbox
docker network rm <network-name>

Alternatively use this command:

docker network rm $(docker network ls --filter 'name=devil*' -q | xargs)

That should fix your problems

[John-MEO-o]

@cytopia ok, I'll check it, thanks Best regards, Jason E. Отправлено из ProtonMail mobile

…

-------- Исходное сообщение -------- Вкл 28 сент. 2017 г., 20:33, cytopia написал(а):

There seems to be two docker networks. One from your previous devilbox instance and the new one. However both do use the same IP. So do the following (First shutdown the devilbox): # show devilbox networks docker network ls --filter ' name=devil* ' # delete all networks starting by devilbox docker network rm < network-name > Alternatively use this command: docker network rm $( docker network ls --filter ' name=devil* ' -q | xargs ) That should fix your problems — You are receiving this because you were mentioned. Reply to this email directly, [view it on GitHub](#119 (comment)), or [mute the thread](https://github.com/notifications/unsubscribe-auth/AZjY4IcC5rAnKm0yX1Y_nokXiB-AFkENks5sm7xTgaJpZM4PGkxi).

[John-MEO-o]

@cytopia Yep, it was not so simple as i want,
but anyway, i did removed those networks, forcefully...
Afterwards i perform a clean start of devilbox...
on default IP (127.0.0.1) and on docker-machine IP (192.168.99.100) web interface were still inaccessible.
on service IP (172.16.238.11) it gives an error (can't assign requested address)

[cytopia]

That means address or port is already in use on your side. Can you check what else is listening on the required ports on those IP addresses

[John-MEO-o]

@cytopia Can you advise, which command should i use for that checking?
basic netstat can show little bit more excess for security reasons...

also there's some outputs for executed commands...
Maybe this provide some help.

While IP is broadcasted (LOCAL_LISTEN_ADDR=)

• webUI not responding (sql & http logs are):
  

While IP is set to default (LOCAL_LISTEN_ADDR=127.0.0.1:)

• webUI not responding (sql & http logs are):
  
  and there's a bind log
  

While IP belongs to docker (LOCAL_LISTEN_ADDR=192.168.99.100:)

• webUI not responding (sql & http logs are):
  

While IP belongs to service itself (LOCAL_LISTEN_ADDR=172.16.238.11:)

• utility doesn't start at all (output as follows):
  

[cytopia]

@jayson-lynnfield-zen yes I can help :-)

You are using the master branch, however, all the fixes are currently not yet merged to master and are still in branch REL-0.12.

So as described above, checkout REL-0.12, compare the env-example with .env and start it up.

[cytopia]

Looks like the port assigned to bind in .env is already taken.

• Alter the bind port in .env
• Checkout above mentioned branch REL-0.12
• Start up again (try only with httpd, bind and php)

[John-MEO-o]

@cytopia I've finally got my windows-desktop, and apply REL-0.12 branch...
I've made a clean install, .env file being created right from your new env-example file...

first attempt with no any address specified (broadcasted):
localhost, 127.0.0.1, 172.16.238.11 are not responding
192.168.99.100 got step forward... it gave me a 404 error...

second attempt was with docker ip specified (192.168.99.100:)
404 on 192.168.99.100, others not responding

third time I've set 127.0.0.1: value (as in your default version)
all addresses are not responding

Then I've set a broadcast address again, change BIND port to 3000 (1053 was initially being set), and there's my logs-output

no responds from any address were found ((

[John-MEO-o]

@cytopia and there's output from logs httpd & bind services

5.39 is router

[cytopia]

@jayson-lynnfield-zen

I think, the only way to go from your debugging is:

1. LOCAL_LISTEN_ADDR= (empty, so it listens on all address)
2. Open URL: http://192.168.99.100

This leads to a 404, the furthest you got.

So can you paste the httpd error log from nginx here after you made a request to http://192.168.99.100

[John-MEO-o]

@cytopia correct me if i wrong...
httpd error log can be called by docker-compose logs httpd command right?

there's a log right after utility being started

and there's a log after i made a request... (seems all the same)

[cytopia]

@jayson-lynnfield-zen No. I mean the physical log file inside the ./log/nginx-stable/ or ./log/nginx-mainline/ folder.

[John-MEO-o]

@cytopia in my case log folder has only one, single file, called .keepme...
Maybe it switched off, or something?

[cytopia]

@jayson-lynnfield-zen there must be something as soon as the devilbox starts. Could you do the following and paste the folder structure here:

ls -lR log/

[John-MEO-o]

@cytopia

[cytopia]

@jayson-lynnfield-zen
You could also quickly join the gitter chat. I have a few more minutes before I have to leave: https://gitter.im/devilbox/Lobby?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge

[cytopia]

As discussed on gitter, I will close it and mark it as wont fix

[aozkan]

Hi, Was there a solution to this issue? It is identical to the problems I'm having using Windows 7.

[prototype2118]

Same issue as above. 404 nginx error on windows 10 home edition. I have to use Docker Toolbox and nothing is being created in the devilbox folders.

[cytopia]

I will re-open this issue as it seems to be active again. @aozkan @prototype2118 please provide detailed information about your system and how to reproduce this issue, so I can start debugging this.

[prototype2118]

Same issue as described above. I use Docker Toolbox on a Windows 10 Home Edition laptop and can compose up to create the container. Going to the IP address of the container brings me directly to the 404 nginx page. I cannot get to the devilbox admin page.

[cytopia]

@prototype2118

Going to the IP address of the container brings me directly to the 404 nginx page

How did you get the IP address of the container? Did you have a look here: https://devilbox.readthedocs.io/en/latest/howto/docker-toolbox/find-docker-toolbox-ip-address.html#howto-find-docker-toolbox-ip-address

[prototype2118]

@prototype2118

Going to the IP address of the container brings me directly to the 404 nginx page

How did you get the IP address of the container? Did you have a look here: https://devilbox.readthedocs.io/en/latest/howto/docker-toolbox/find-docker-toolbox-ip-address.html#howto-find-docker-toolbox-ip-address

The IP address of the container was shown upon startup. It's the same IP address in the link you provided.

[cytopia]

@prototype2118 can you attach your .env file.

[cytopia]

There's an upcoming PR which is going to mitigate DNS issues on Docker Toolbox (Windows and MacOS): #411

Please have a look and let me know how this works for your.

[cytopia]

@jayson-lynnfield-zen @aozkan @prototype2118

I have removed wontfix as development for this is already in progress.

This ticket is now a duplicate of: #248

I will therefore close this one and all further discussions should continue here: #248