Hello I have following dockerfile on my macOS Sierr High (as example):
FROM richarvey/nginx-php-fpm:1.3.7
EXPOSE 80
And I'm able to build and run it on 192.168.8.101:8082 by
docker build -t myproject
docker run -d -p 192.168.8.101:8082:80 --name myproject myproject
and it works on http://192.168.8.101:8082 my local computer and devices (iPhone) connected to the same wifi network (I also have no problem to run container on localhost).
However when I try tu run this container on other IP like for example 192.168.8.102:8082 I get following error:
Mac-mini-Kamil:myproject Kamil$ docker run -d -p 192.168.8.102:8082:80 --name myproject myproject
f939d38243f420f812c859f5fe275faf49dc6e123d807583ec240fbdf0619a17
docker: Error response from daemon: driver failed programming external connectivity on endpoint myproject (0b546e63887e3ddeb4d2b21a8d6d15a94e33f1ff67c2765174a808bf6b13e120): Error starting userland proxy: listen tcp 192.168.8.102:8082: bind: cannot assign requested address.
I try also other addresses like 172.17.1.2 but with no success and I don't have idea why. Do someone cane give me advice (and/or instruction about what more information I should provide to this question to solve it) ? I would like to know: why I cannot use other IPs and what I can do to use other IPs?
UPDATE - additional informations:
After execute ifconfig I selected two interesting results :
en1: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
ether a8:8e:24:a3:d3:91
inet6 fe80::18fe:1146:6389:b630%en1 prefixlen 64 secured scopeid 0xa
inet 192.168.8.101 netmask 0xffffff00 broadcast 192.168.8.255
nd6 options=201<PERFORMNUD,DAD>
media: autoselect
status: active
bridge0: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
options=63<RXCSUM,TXCSUM,TSO4,TSO6>
ether 32:00:11:fc:70:40
Configuration:
id 0:0:0:0:0:0 priority 0 hellotime 0 fwddelay 0
maxage 0 holdcnt 0 proto stp maxaddr 100 timeout 1200
root id 0:0:0:0:0:0 priority 0 ifcost 0 port 0
ipfilter disabled flags 0x2
member: en2 flags=3<LEARNING,DISCOVER>
ifmaxaddr 0 port 9 priority 0 path cost 0
nd6 options=201<PERFORMNUD,DAD>
media: <unknown type>
status: inactive
The output of ifconfig (linux, OSX) or ipconfig (Windows) will probably tell you that your IP4 is 192.168.8.101.
That's reason why 192.168.8.101:8082:80 works. Docker can access the endpoint and can map port 80 of its container on port 8082 of 192.169.8.101.
When you use another IP this error will pop up:
Error response from daemon: driver failed programming external connectivity on endpoint
This means your Docker daemon can not access the endpoint 192.168.8.102 which is logical because it isn't the IPV4 address of your machine.
The bridge0 network is the default Docker network. If you don't specify a network your Docker container will be deployed inside this network.
You can do: docker inspect network bridge
The output of this command will return a subnet like "Subnet": "172.17.0.0/16". Every container will have an IP between this private range.
What you are actually doing is mapping 172.17.X.X:80 on 192.168.8.101:8082 which works when 192.168.8.101 is accessible from you machine.
i suspect that your previous container still using port 8082 try using another port. . i dont know how to check used port in mac but if its in linux you can check your open port with netstat -ntlp
Related
I know there are a lot of posts on the internet about this subject, but I have been debugging this over two days now and I don't seem to get any further.
The setup
I have a (pretty old) Vagrant box on my work laptop that's been setup mostly by a former collegae. Everything seems to be still working for me and my collegaes so we don't have had any reason to setup a completely new one.
The Vagrant box contains a Centos installation on which we develop websites.
The host machine is Windows 10.
The problem
When in the office, connected to the physical network (by cable), I can use XDebug without any problems. I enable XDebug from my Firefox browser plugin and XDebug on Centos then connects to PHPStorm on Windows, so I can step through code.
However, when at home, on WIFI (I don't have a cable) XDebug just won't work.
The XDebug log on the Vagrant machine currently states the following:
I: Checking remote connect back address.
I: Checking header 'HTTP_X_FORWARDED_FOR'.
I: Checking header 'REMOTE_ADDR'.
I: Remote address found, connecting to 10.10.10.1:9000.
E: Time-out connecting to client. :-(
Research
In many posts I read that the Vagrant host address should be something like 10.0.2.2. As far as my information goes in our case it has always been 10.10.10.1.
I also read that from the Vagrant box you should check the host IP by using netstat. The host IP would be the default Gateway.
At home (while on WIFI) I tested this and the output was:
netstat -rn
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt Iface
0.0.0.0 192.168.220.2 0.0.0.0 UG 0 0 0 ens32
10.10.10.0 0.0.0.0 255.255.255.0 U 0 0 0 ens33
169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 ens33
169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 ens34
192.168.56.0 0.0.0.0 255.255.255.0 U 0 0 0 ens34
192.168.220.0 0.0.0.0 255.255.255.0 U 0 0 0 ens32
Here the default Gateway is "192.168.220.2". So I tried to set that IP for xdebug, by manually setting XDebug's remote_host to 192.168.220.2 and disabling remote_connect_back
Now the log says:
I: Connecting to configured address/port: 192.168.220.2:9000.
W: Creating socket for '192.168.220.2:9000', poll success, but error: Operation now in progress (29).
E: Could not connect to client. :-(
Log closed at 2020-01-26 08:19:34
From other posts I understood that this is in fact worse and the IP-address is just wrong, because it should be an internal IP-adress like 10.10.10.1
Edit 1: I tested this today while in the office, connected to the physical network. XDebug works as expected, but the output of netstat is the same for the Gateway, so this probably has nothing to do with it:
Destination Gateway Genmask Flags MSS Window irtt Iface
0.0.0.0 192.168.220.2 0.0.0.0 UG 0 0 0 ens32
10.10.10.0 0.0.0.0 255.255.255.0 U 0 0 0 ens33
169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 ens33
169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 ens34
172.24.1.0 0.0.0.0 255.255.255.0 U 0 0 0 ens34
192.168.220.0 0.0.0.0 255.255.255.0 U 0 0 0 ens32
Edit 2: While in the office, I now also tested XDebug while on WIFI (by disconnecting the physical network cable) and then XDebug no longer works. So the problem does not seem to be specific to my home network, but rather to being on physical network VS WIFI.
PHPStorm says I'm fine
PHPStorm has this xdebug validate screen in the settings. Trying that passes everything with green checks. Appearently that validation doesn't validate enough?
Interesting to note
Simultaneously I was also trying to mount a Samba share from Windows into my Vagrant box. I wanted to experiment with that setup, but just like XDebug the mount command is unable to reach the host machine.
Edit: I tested this today while in the office on the physical network and this then doesn't work either. So we can probably forget about this for now as this seems to be a completely different issue.
Edit 27-01: I thought I fixed the problem, but this only worked when on WIFI at the office. At home, it still doesn't work.
The fix in the office was fiddling with IP routes in the Centos server.
Edit 29-01:
I still don't have this working, but after several tests I think I can state that the VM can reach the host:
- ping 10.10.10.1 works
- nmap -p 9000 10.10.10.1 seems to tell me that it can reach the port
Besides that today I found out that when PHPStorm is not listening on port 9000, the XDebug log file says:
"Creating socket for '192.168.220.2:9000', poll success, but error: Operation now in progress (29)."
But when PHPStorm is listening for a connection the XDebug log says:
"Remote address found, connecting to 10.10.10.1:9000"
The fact that the errormessage on the VM's side changes when the state of the host changes, makes me believe that the connection is coming through.
That would mean that - for some reason - PHPStorm simply is unable to handle the incoming connection. The only reason for that I could think of is that PHPStorm may not be able to talk back to the VM?? Does that sound plausible? And if so, how to further investigate this?
The following phrase seems to indicate that a connection is made.
W: Creating socket for '192.168.220.2:9000', poll success, but error: Operation now in progress (29).
Are you sure that it is PhpStorm that is listening on port 9000 locally? And that your Vagrant box can talk to it?
Let's find out the first one: In a shell on your host (which runs Windows in both cases, I presume), run:
C:\> netstat -a -b
And make sure that it is the PhpStorm process that is listening on port 9000.
If it is something else, change the port in the PhpStorm config to something else (say 9003), and make that same port the value for the xdebug.remote_port setting in php.ini. The turn off "listening for debugging connections" in PhpStorm, and turn it back on again.
Secondly, test whether the Vagrant box can talk to that port
I had a similar issue and for me it was an issue with the host machine network profile. It was not discoverable in the private network on allowing it for discover in the private network it did worked.
Below is the configuration in my Windows host machine after which it did worked.
Network Profile Configuration
I've got a app running on my computer in localhost:1235, and I'm trying to load test it.
I installed k6 container for docker to test it, but of course from the nature of docker, my container has a different localhost. I'm trying to understand what do.
I run the following command:
docker run -it --rm --net=host -v c:/users/k6:/k6 loadimpact/k6 run /k6/script
I read somewhere that --net=host doesn't work on windows, is that right? How would I find the host IP?
I've tried running by this tutorial:
http://blog.michaelhamrah.com/2014/06/accessing-the-docker-host-server-within-a-container/
The IP I find 172.17.0.1 doesn't work in my test.
I also tried adding -p 1235:1235 but it failed, I guess docker tries to bind this port and just forward to it.
Thanks in advance,
Chaim
Inside your k6 script use the url host.docker.internal to access something running on the host machine.
For example to access a service running on the host at http://localhost:8080
// script.js
import http from "k6/http";
import { sleep } from "k6";
export default function () {
http.get("http://host.docker.internal:8080");
sleep(1);
}
Then on windows or mac this can be run with:
$ docker run -i loadimpact/k6 run - <script.js
for linux you need an extra flag
$ docker run --add-host=host.docker.internal:host-gateway -i loadimpact/k6 run - <script.js
References:
Mac: https://docs.docker.com/docker-for-mac/networking/#use-cases-and-workarounds
Windows: https://docs.docker.com/docker-for-windows/networking/#known-limitations-use-cases-and-workarounds
Linux: https://stackoverflow.com/a/61424570/3757139
k6 inside the docker instance should be able to connect to the "public" IP on your host machine - the IP that is configured on your ethernet or Wifi interface. You can do a ipconfig /all to see all your interfaces and their IPs.
On my Mac I can do this:
$ python httpserv.py &
[1] 7824
serving at port 8000
$ ifconfig en1
en1: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
ether b8:09:8a:bb:f7:ed
inet6 fe80::148f:5671:5297:fc24%en1 prefixlen 64 secured scopeid 0x5
inet 192.168.0.107 netmask 0xffffff00 broadcast 192.168.0.255
nd6 options=201<PERFORMNUD,DAD>
media: autoselect
status: active
$ echo 'import http from "k6/http"; export default function() { let res = http.get("http://192.168.0.107:8000"); console.log(res.status); };' |docker run -i loadimpact/k6 run -
I.e. I start a simple HTTP server on port 8000 of the host machine, then executes the k6 docker image and tells it to access a URL based on the IP address of the physical, outward-facing en1 interface on the host machine. In your case, on Windows, you can use ipconfig to find out your external-facing IP.
I am running a windows docker container on a Windows Server 2016 host, running default configuration.
When running the docker container using the command:
docker run -it microsoft/windowsservercore powershell
When I run the command:
ping <hostIPAddress>
It just says that the request times out.
I have checked that I can ping 8.8.8.8 and google.com etc... and even other machines on the same subnet. The only one I cannot ping is the host.
I have added '--dns ' to the 'docker run' command but this only allows me to ping the host machine via hostname and not IP.
Has anyone else seen this problem and have a solution?
I found a workaround (I'm not willing to call it a solution):
Windows Container Network Drivers: create a 'transparent' network:
docker network create -d transparent trans
Attach container to this network
docker run --network=trans ...
Important: Please note, that with this network, your container needs to obtain an IP Adress from the Host Subnet and it is directly exposed to it.
maybe related (this is about access the containers from the host):
According to https://github.com/Microsoft/Virtualization-Documentation/issues/253#issuecomment-217975932 (JMesser81):
This is a known limitation in our Windows NAT implementation (WinNAT) that you cannot access the external port in a static port mapping directly from the container (NAT) host.
Hoping this might help somebody.
On Windows 10 when hosting a Linux container on 0.0.0.0:5057 I was able to ping my server from my Windows host (powershell) using the IP address of the vEthernet (Default Switch) NIC found in Control Panel>All Control Panel Items>Network Connections:
In my case I have a corporate managed McAfee firewall running on my Windows host. I could not add any additional rules on the firewall, but fortunately there was a rule that allowed access from 172.16.0.0/24.
I used "docker network create -d transparent trans" and it worked as described, but I was not happy with an IP from my host network assigned to the container.
I did the following:
docker network create --driver=nat --subnet=172.16.0.0/24 br0
Added --network=br0 to my docker run command
I am facing the same issue.
My workaround is to restart docker service, afterwards it works fine. I'm still looking for a permanent solution.
root#a6c40eb25cbf:/# ping xxx.xx.xx.xxx
PING xxx.xx.xx.xxx (xxx.xx.xx.xxx): 56 data bytes
64 bytes from xxx.xx.xx.xxx: icmp_seq=0 ttl=37 time=3.541 ms
64 bytes from xxx.xx.xx.xxx: icmp_seq=1 ttl=37 time=2.643 ms
64 bytes from xxx.xx.xx.xxx: icmp_seq=2 ttl=37 time=1.857 ms
^C--- xxx.xx.xx.xxx ping statistics ---
3 packets transmitted, 3 packets received, 0% packet loss
This works for me, I hope it works for you.
Currently, you must have WSL or WSL2 installed and have Virtualization enabled to run Docker on Windows.
The installation of WSL from PowerShell is with the following command.
wsl --install -d Ubuntu
Obviously you need to download and install Docker Desktop on Windows. It will be necessary to enable the WSL integration from the Docker desktop settings after installing it.
After configuring WSL and Docker Desktop, you can create/use your containers. Example:
docker run --name some-postgres -e POSTGRES_PASSWORD=mysecretpassword -d postgres
Finally you must identify the IP of WSL, you must open the Linux distribution that you installed in the first step, in our case Ubuntu, this will open your terminal and here we will execute:
ifconfig
and you will identify the ip of eth0. Example: 172.27.123.123
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 172.27.123.123 netmask 255.255.240.0 broadcast 172.27.127.255
inet6 fe80::215:5dff:fecf:b4 prefixlen 64 scopeid 0x20<link>
ether 00:15:5d:cf:00:b4 txqueuelen 1000 (Ethernet)
RX packets 4389 bytes 299784 (299.7 KB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 4343 bytes 315643 (315.6 KB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
Now you can ping from CMD in Windows to this IP.
NOTE: the connection will depend on the port of your container, example: 172.27.123.123:8080.
Running macOS and docker ubuntu on it, I am unable to ping my local network's IP addresses. Below is the network configuration of my docker container. I can ping my host machine's IP address but I am unable to access other IP addresses on the local network.
root# ping 172.17.101.192
PING 172.17.101.192 (172.17.101.192) 56(84) bytes of data.
From 172.17.0.3 icmp_seq=1 Destination Host Unreachable
From 172.17.0.3 icmp_seq=2 Destination Host Unreachable
root# ifconfig
eth0 Link encap:Ethernet HWaddr 02:42:ac:11:00:03
inet addr:172.17.0.3 Bcast:0.0.0.0 Mask:255.255.0.0
inet6 addr: fe80::42:acff:fe11:3/64 Scope:Link
Is there any network configuration which would help me to perform this operation?
I have tried net=host but it is not helping it.
I was using docker for mac, the issue was docker internal network conflicting with my local network addressing.
To resolve the this issue, i need to go to Preferences in the dokcer menu.
In Preferences menu Daemon>Advanced Menu, i can supply other bip.
{
"bip" : "12.12.0.1/24"
}
Click apply and restart. The next time docker will start with internal network as 12.12.0.1/24.
I think docker internal network could be conflicting with your local network addressing.
Try to run Docker changing the default internal network to something that doesn't conflict, as 172.31.0.0/24. Add to your Docker startup options --bip 172.31.0.1/24, probably located in /etc/default/docker:
# Use DOCKER_OPTS to modify the daemon startup options.
DOCKER_OPTS="--bip 172.31.0.1/24"
UPDATE: Although this is true in Linux, Mac implementation is different, see docs here, so it depends if you're using Docker for Mac or for Linux.
Just to test, try to attach your laptop to a different network, any in the range of 192.168.0.0/16 would suffice.
I'm guessing you're using docker-machine.
You should check out docker for mac as it has some networking improvements that may make this work for you. This is what I'm using and I'm able to reach other hosts on my internal network from docker.
I tried the following, and it works:
mac-os $> ifconfig
en0: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
ether 78:31:c1:bd:4b:84
inet6 fe80::1425:a90d:9c00:ef53%en0 prefixlen 64 secured scopeid 0x4
inet 10.191.41.1 netmask 0xfffff800 broadcast 10.191.47.255
nd6 options=201<PERFORMNUD,DAD>
media: autoselect
status: active
So my ip is 10.191.41.1. Then I do:
mac-os $> docker run -ti --rm ubuntu bash -c "apt-get update && apt-get install -y inetutils-ping && ping 10.191.41.1"
Which installs the "ping" command, and tries to ping my host. I get the right answers:
[...]
PING 10.191.41.1 (10.191.41.1): 56 data bytes
64 bytes from 10.191.41.1: icmp_seq=0 ttl=37 time=0.303 ms
64 bytes from 10.191.41.1: icmp_seq=1 ttl=37 time=0.536 ms
[...]
I wanted to throw this answer out as it happened to me on Windows, but could really on any environment. I was using a bridge network on my local machine for a service I was developing and ran into similar issues, what was weird was that pinging our production physical servers (which started with a 10.10... IP pattern) or the random sites on the internet I was able to receive a response. But the physical dev server which sat on a 172.16... pattern didn't work.
Finally I discovered through doing docker network inspect on the networks I had locally that one had created with the same subnet of 172.16... which overlapped my physical server IP address range. So when I was pinging the physical dev server, it was trying to find it on the virtual docker network, where it didn't exist.
After deleting that network (it was a dev one I didn't use), everything worked.
docker network rm {networkName}
If I had to keep the network, it is possible to change the subnet in the docker-compose file. If you need to change the conflicting subnet, the link to that is here How do configure docker compose to use a given subnet if a variable is set, or choose for itself if it isn't?
That would look something like this (if you wanted a 12.12 subnet):
networks:
default:
ipam:
config:
- subnet: 12.12.0.0/16
I am currently running a docker machine in local network that its IP & port is 192.168.99.100:8080
right now. I want to connect this machine from somewhere out of the local network and access it, so the best way I found is to NAT the sent request from my laptop computer to docker machine.
In other words, each request that has been sent to my laptop computer will be sent to the docker machine on its own IP and port!
I use pfctl to do this but when I check the port 8585 to find out which service is listening to this port, I can't find anything and it sounds like the requests do not arrive(get to) the destination!
I did as below:
1- csrutil disable
2- my pf role :
/etc/pf.anchors/mazafard:
rdr pass on en0 inet proto tcp from any to 192.168.20.203 port 8585 -> 192.168.99.100 port 8080
3- my pf conf
/etc/pf-mazafard.conf
rdr-anchor "forwarding"
load anchor "forwarding" from "/etc/pf.anchors/mazafard"
and
sudo pfctl -vnf /etc/pf.anchors/mazafard
pfctl: Use of -f option, could result in flushing of rules
present in the main ruleset added by the system at startup.
See /etc/pf.conf for further details.
rdr pass on en0 inet proto tcp from any to 192.168.20.203 port = 8585 -> 192.168.99.100 port 8080
and
sudo pfctl -evf /etc/pf.anchors/mazafard
fctl: Use of -f option, could result in flushing of rules
present in the main ruleset added by the system at startup.
See /etc/pf.conf for further details.
No ALTQ support in kernel
ALTQ related functions disabled
rdr pass on en0 inet proto tcp from any to 192.168.20.203 port = 8585 -> 192.168.99.100 port 8080
pfctl: pf already enabled
and
telnet 192.168.20.203 8585
Trying 192.168.20.203...
telnet: connect to address 192.168.20.203: Operation timed out
telnet: Unable to connect to remote host
My OSX v: 10.11.4
Docker version 1.9.1, build a34a1d5
Boot2Docker version 1.11.0-rc3, build