I'm running postgresql on a Vagrant centos 7.2 box on macOS. I have confirmed postgres is up and running on the Vagrant box on port 5432 by connecting to it using psql on the box itself. I am attempting to forward port 5432 on the Vagrant box to port 10001 on my host machine as follows:
config.vm.define "acc_db" do | acc_db |
acc_db.vm.box = "bento/centos-7.2"
acc_db.vm.hostname = "acc.db"
acc_db.vm.network :forwarded_port, guest: 5432, host: 10001
acc_db.vm.provision "shell",
inline: "yum upgrade -y -q --nogpgcheck"
acc_db.vm.provision "shell",
path: "install_postgres.sh"
I have altered my pg_hba.conf to bind to all ip addresses and allow password authentication, as below:
# TYPE DATABASE USER ADDRESS METHOD
# "local" is for Unix domain socket connections only
local all all peer
# IPv4 local connections:
host all all 0.0.0.0/0 md5
# IPv6 local connections:
host all all ::1/128 ident
I have turned off the firewall but I am still unable to connect to postgres on port 10001 on the host machine:
psql: could not connect to server: No such file or directory
Is the server running locally and accepting
connections on Unix domain socket "/tmp/.s.PGSQL.10001"?
How do I fix this so that port forwarding works?
I have seen https://gielberkers.com/fixing-vagrant-port-forwarding-osx-yosemite/ but haven't tried it as it involves touching files I'm unfamiliar with. Is it the correct approach? It looks like you have to explicitly allow every port you want to forward via Vagrant.
...connections on Unix domain socket "/tmp/.s.PGSQL.10001" means that you are not attempting a TCP/IP connection, so port forwarding cannot happen.
By default, psql use Unix domain sockets on unix-like OSes.
Use the -h option of psql to specify a host, presumably 127.0.0.1 if using IPv4.
Additionally, the PostgreSQL server must listen to the network interface to which the connection will be routed. By default, it listens only to localhost for security reasons.
Setting listen_addresses='*' in postgresql.conf makes it listen on all existing interfaces.
Related
I have installed PostgreSQL 13 on Windows 10 using the EDB installer and it seems to work fine. I can connect easily with psql in the included SQL Shell application. However, I can not connect from powershell:
~> psql -U postgres -h localhost
psql: error: could not connect to server: Connection refused (0x0000274D/10061)
Is the server running on host "localhost" (::1) and accepting
TCP/IP connections on port 5432?
could not connect to server: Connection refused (0x0000274D/10061)
Is the server running on host "localhost" (127.0.0.1) and accepting
TCP/IP connections on port 5432?
Any permutation of psql command line switches yields the same result. I checked pg_hba.conf and it looks sane - all local connections are allowed. All the search results for the error code above on the internet assume the server is not running but it is running just fine (see SQL Shell above). Within that shell, all the postgres tables are visible, etc. What do I need to do to connect psql to the server on localhost from powershell?
Default installed pg_hba.conf:
# "local" is for Unix domain socket connections only
local all all scram-sha-256
# IPv4 local connections:
host all all 127.0.0.1/32 scram-sha-256
# IPv6 local connections:
host all all ::1/128 scram-sha-256
The IPv4 line above should cover me in powershell? Yes? tcp connections are allowed in postgresql.conf:
listen_addresses = '*' # what IP address(es) to listen on;
Long time linux PostgreSQL user, first time in powershell on Windows, though...
This morning I found the problem - the EDB installer had setup the postgres service to run on port 5433. After a couple of decades or running it on 5432 I didn't even notice that the number was different in postgresql.conf. It was late...
Something else must be running on 5432. Will check later on.
I run on Windows 10
docker run -it --rm -e POSTGRES_PASSWORD=postgres -p5401:5432 postgres:12.3-alpine
it works fine. I can connect to PostgreSQL, create database etc, all is ok.
But when I run the same docker on my 2nd machine Windows 7 I receive
Could not connect to server: Connection refused (0x0000274D/10061) Is
the server running on host "localhost" (::1) and accepting TCP/IP
connections on port 5401?
What may be the reason? what to check?
That is pg_hba.conf
# TYPE DATABASE USER ADDRESS METHOD
# "local" is for Unix domain socket connections only
local all all trust
# IPv4 local connections:
host all all 127.0.0.1/32 trust
# IPv6 local connections:
host all all ::1/128 trust
# Allow replication connections from localhost, by a user with the
# replication privilege.
local replication all trust
host replication all 127.0.0.1/32 trust
host replication all ::1/128 trust
host all all all md5
I would say the possible issues are:
1) The server has not actually been started in the Docker server.
Verify server is running.
2) Port 5401 is not visible to the client that is trying to connect.
Ping the port. Could be a firewall issue.
I have two VirtualBox Vagrant machines running on my Mac:
Ubuntu 16.04.1, private network: 192.168.122.13, running as a webserver (PHP, Apache, etc.).
Ubuntu 16.04.1, private network: 192.168.122.14, running MySQL.
How do I connect to MySQL from the webserver?
I have enabled port forwarding on the MySQL box (3306 guest => 5629 host) and commented out the bind-address line in /etc/mysql/mysql.conf.d/mysqld.cnf. If I SSH onto the webserver and try to connect using mysql -host 192.168.122.14 -P 5629 -u xxx -p yyy the connection times out (ERROR 2003 (HY000): Can't connect to MySQL server on '192.168.122.14' (110)).
MySQL runs on the vagrant box with IP 192.168.122.14, and is listening on port 3306. When you forward ports, it means that the port 3306 from the virtual machine, is forwarded to your host OS on the port 5629.
From your web server (192.168.122.13) you could connect to mysql by
mysql -h 192.168.122.14 -P 3306
or
mysql -h 192.168.122.1 -P 5629
Here 192.168.122.1, refers to the IP of your host OS. The IP 192.168.122.1 is created by vagrant and assigned to your host OS
I'm running werkzeug (as part of a Tilestache setup) inside a Vagrant VM, running ubuntu 'precise.'
In my Vagrantfile, I have:
config.vm.network :forwarded_port, guest: 8080, host: 8080
When I start the server in the VM, I see:
* Running on http://127.0.0.1:8080/
If I curl that address from within the VM, I get the expected result. When I curl it from the host machine, I get:
curl: (52) Empty reply from server
And Chrome says "No data received."
Troubleshooting info:
The server responds to pings from the host machine
a port sniffer verifies that the port is open
running netstat -ntlp | grep 8080 in the vm shows that the server is listening on 8080
My local hostsfile doesn't have any weird conflicts
I'm also forwarding 22 => 2222, and I can ssh in with no trouble
I've disabled the firewall on the host, and i don't believe there's one on the guest (iptables and ufw are disabled, at least)
I've set auto_correct: true in case there are conflicts (there aren't)
I know I could set up a private network, but I'd like to understand why this isn't working and how to troubleshoot it.
Any other ideas?
When running a server from within a VM, start the server on 0.0.0.0 instead of 127.0.0.1.
127.0.0.1 is only accessible to the local machine, which for a VM means nothing outside of the VM can reach it! 0.0.0.0 is accessible from anywhere on the local network, which to a VM includes the host machine.
The answer came from here: Connection Reset when port forwarding with Vagrant
(Which apparently got its answer from here: https://stackoverflow.com/a/5999945/738675)
With help from: https://serverfault.com/questions/78048/whats-the-difference-between-ip-address-0-0-0-0-and-127-0-0-1
Google-bait:
Here are the errors you might receive if this is the problem:
Chrome: "No data received"
Firefox: "The connection was reset - The connection to the server was reset while the page was loading."
Safari: "Safari can’t open the page [URL] because the server unexpectedly dropped the connection"
curl: "Empty reply from server"
In the /etc/hosts inside the VM, change line
127.0.0.1 localhost -> 0.0.0.0 localhost
and then restart server
This can also be a problem with your firewall on the vagrant machine. If you can curl the address while on the vagrant box, then check your firewalld settings or turn it off:
on CENTOS:
sudo service firewalld stop
Then you should update your firewalld settings and restart it ;)
I'm working on a web services architecture. I've got some software that I need to run on the native host machine, not in Vagrant. But I'd like to run some client services on the guest.
Vagrant's config.vm.forwarded_port parameter will open a port on the host and send the data to the guest. But how can I open a port on the guest and send the data to the host? (It's still port forwarding, but in the reverse direction.)
When you run vagrant ssh, it's actually using this underlying command:
ssh -p 2222 -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -o LogLevel=ERROR -o IdentitiesOnly=yes -i ~/.vagrant.d/insecure_private_key vagrant#127.0.0.1
SSH supports forwarding ports in the direction you want with the -R guestport:host:hostport option. So, if you wanted to connect to port 12345 on the guest and have it forwarded to localhost:80, you would use this command:
ssh -p 2222 -R 12345:localhost:80 -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -o LogLevel=ERROR -o IdentitiesOnly=yes -i ~/.vagrant.d/insecure_private_key vagrant#127.0.0.1
As Eero correctly comments, you can also use the command vagrant ssh -- -R 12345:localhost:80, which has the same effect in a much more concise command.
In the book Vagrant: Up and Running (Pub. date: June 12, 2013), written by the creator of Vagrant, he mentioned that it is not possible for guest machine to access services running on the host machine.
Instead of using Forwarded Ports, you could set up a private network using Host-Only Networks.
Pros of using Host-Only Networks over Forwarded Ports
Guest machines may access the services running on host machine
This feature would solve your problem.
Guest machines may access the services running on other guest machine
This feature is very useful to separate services onto multiple machines to more accurately mimic a production environment.
Secure
Outside machines have no ways to access the services running on the guest machines
Less work
No need to configure every single Forwarded Port
How to configure Host-Only Networks
config.vm.network :"hostonly", "192.168.0.0" # Vagrant Version #1
config.vm.network :private_network, ip: "192.168.0.0" # Vagrant Version #2
Having this line in your Vagrantfile will instruct vagrant to create a private network that has a static IP address: 192.168.0.0
The IP address of the host is always the same IP address but with the final octet as a 1. In the preceding example, the host machine would have the IP address 192.168.0.1.
You can access ports on the host machine through the default gateway inside the guest OS. (Which typically has an IP of 10.0.2.2.)
For example, if you have a webserver running on port 8000 on your host machine...
echo 'Hello, guest!' > hello
python -m SimpleHTTPServer 8000
You can access it from inside the Vagrant VM at 10.0.2.2:8000 (provided 10.0.2.2 is the ip of the guest's default gateway):
vagrant ssh
curl http://10.0.2.2:8000/hello # Outputs: Hello, guest!
To find the IP of the default gateway inside the guest OS, run netstat -rn (or ipconfig on a Windows guest) and look for the row with a destination IP of 0.0.0.0 (or the field labeled "Default Gateway" on Windows):
$ netstat -rn
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt Iface
0.0.0.0 10.0.2.2 0.0.0.0 UG 0 0 0 eth0
10.0.2.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
192.168.33.0 0.0.0.0 255.255.255.0 U 0 0 0 eth1
You can extract this IP programmatically with netstat -rn | grep "^0.0.0.0 " | tr -s ' ' | cut -d " " -f2.
Sources: How to connect with host PostgreSQL from vagrant virtualbox machine; Connect to the host machine from a VirtualBox guest OS?
Add following to your ~/.ssh/config on the host machine:
Host 127.0.0.1
RemoteForward 52698 127.0.0.1:52698
It lets you access a service on host machine port 52698 from Vagrant, as long as you logged in via vagrant ssh.
You can confirm it works by running netstat -lt on vagrant VM and taking a note on the following lines:
tcp 0 0 localhost:52698 *:* LISTEN
tcp6 0 0 ip6-localhost:52698 [::]:* LISTEN
I can access services running on my host machine via its local IP address (not its loopback address). I tested by creating an http server on port 80 (and then on port 987) and curling 197.45.0.10:80 and 197.45.0.10:987 (actual ip address changed to protect the innocent). It worked both times, and I don't have any special vagrant configuration (no public_network, no forwarded_port) and while I do have some ports forwarded via PuTTY, I don't have ports 80 and 987 forwarded. So maybe try using the host machine's local or public IP address.
And if you want to access (ssh into) one guest vagrant instance from another, you can enable public_network as well as forwarding from port 22 in the Vagrantfile like this:
config.vm.network "public_network"
config.vm.network "forwarded_port", guest: 22, host: 2200
Then as long as that port is open (ie do some more port forwarding in your router config) you can access that machine from anywhere, even the outside world.