We have seen a issue in log
FATAL: lock file "/tmp/.s.PGSQL.5432.lock" already exists
HINT: Is another postmaster (PID 4696) using socket file "/tmp/.s.PGSQL.5432"?
LOG: could not bind IPv4 socket: Address already in use
We have deleted /tmp/.s.PGSQL.5432.lock
We restarted the postgres service the service restarted
But the psql -h localhost -d database -U newuser
shows error role 'newuser' does not exists
but if we connect without host
psql -d database -U newuser
It worked, How can We work with host options ?
Is there we ,missing somthing after deleting tmp file?
Same as the last one, you have two instances of PostgreSQL running on the machine.
You delete the socket file rather than shutting down the instance. This allows your second instance to start, but it's only listening on the unix socket. Presumably listen_addresses is unset/empty so it doesn't attempt to bind to a TCP/IP socket, since it'd fail if it tried and port 5432 was already in use.
The other instance is still running and still listening on port 5432 for TCP/IP connections. When you specify -h localhost you get a TCP/IP connection instead of the default unix socket connection, so you are connecting to the other instance of PostgreSQL on your system.
All your problems are caused by going around deleting and killing things rather than identifying the underlying problem - the second PostgreSQL instance - and addressing that.
Related
When I'm trying to connect to a docker MySQL that's running and forwarding to my local TCP:3306 I get the following answer
mysql -u root -pPASSWORD
mysql: [Warning] Using a password on the command line interface can be insecure.
ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/tmp/mysql.sock' (2)
but when I do mysql -u root -pPASSWORD -h127.0.0.1 it connects wonderfully.
Any clue?
[UPDATE]
Considering the comments and this post I create ~/.my.cnf
with this content
[Mysql]
user=root
password=yourpass
host=127.0.0.1
port=3306
Giving these changes I could connect the localhost with the 127.0.0.1 address.
If you don't specify a host with -h (or a host directive in your .my.cnf), then MySQL defaults to connect to localhost. Connections to localhost use the UNIX domain socket interface, not TCP/IP. So it's not connecting to a TCP/IP port, and therefore does not forward to your docker container.
This distinction between localhost and 127.0.0.1 is a historical oddity of MySQL. Normally localhost and 127.0.0.1 are assumed to be equivalent. But MySQL treats the hostname "localhost" as special, using it to invoke the UNIX domain socket interface. This is a bit faster than using TCP/IP, but of course only works if the connection is on the local computer.
I'm trying to restore a backup to my localhost machine to review the data - so that I can actually make sure the backup is working. I'm using Heroku, postgres, and rails.
I clean installed the app, so there is no data in the localhost database. I ran db:reset, and db:migrate.
Here is the string I'm trying to use for pg_restore:
pg_restore --verbose --clean --no-acl --no-owner -h localhost -p 3000 -U ian -d backup_production latest.dump
The error I receive says:
"connection to database "backup_production" failed: server closed the connection unexpectedly"
I've never had to do this before, so any help would be appreciated. I assume that development is what localhost uses (instead of test or production)?
I'm not sure what I'm missing, but I'll answer any question you may have.
Thanks a lot.
edit,. Here is a question, for this database string I assumed that I need to to use all the localhost information, because I created a backup using heroku capture/download. Am I wrong in that assumption? Should I instead be using the database name and port, etc, from the heroku credentials?
Turns out, I shouldn't have added the port. Removing -p 3000 , made this command work.
Try this
Edit the postgre configuration file typing in terminal
sudo gedit /etc/postgresql/POSTGRE_VERSION/main/pg_hba.conf
Then change your configuration to this
# Database administrative login by Unix domain socket
local all all trust
# 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
Restart your postgre server
sudo /etc/init.d/postgresql restart
This should be the solution to access without password
I'm trying to run my rails projects but it keeps falling because it depends on postgres to be running. When I try to start postgres it keep saying its already running (or the port is taken). I checked activity monitor and there are ~6 postgres processes running, when I try to terminate/force-quit the processes, they just appear again...
not sure how to get around this issue.
actual error when running postgres -D /usr/local/var/postgres:
LOG: could not bind IPv6 socket: Address already in use
HINT: Is another postmaster already running on port 5432? If not, wait a few seconds and retry.
LOG: could not bind IPv4 socket: Address already in use
HINT: Is another postmaster already running on port 5432? If not, wait a few seconds and retry.
WARNING: could not create listen socket for "localhost"
FATAL: could not create any TCP/IP sockets
and when I try to stop postgres with pg_ctl -D /usr/local/var/postgres stop -s -m fast, I get this error:
pg_ctl: PID file "/usr/local/var/postgres/postmaster.pid" does not exist
Is server running?
I ended up starting from scratch:
uninstall postgres through homebrew (brew uninstall postgresql)
uninstall postgres though uninstaller (/Library/Postgresql/version/uninstall)
delete postgres through file system (delete the folder: /Library/Postgresql/)
then just do a homebrew installation:
brew install postgresql
and start postgres manually.
1:
pg_ctl -D /usr/local/var/postgres stop -s -m fast
2:
brew services stop postgresql
We have a rails 4 application with postgres
Sometimes our server hanged and if we reboot it .
The server gets started but it can not connect to postgres.
status shows postmaster dead but pid file exists
log shows
FATAL: lock file "/tmp/.s.PGSQL.5432.lock" already exists
HINT: Is another postmaster (PID 4696) using socket file "/tmp/.s.PGSQL.5432"?
LOG: could not bind IPv4 socket: Address already in use
and the postgresql could not restarted.
--- After deleting the /tmp/.s.PGSQL.5432 file and restarting the postgres service the service restarted but the postgres can not connect with host we added.
What can we do to fix this issue ?
Thanks in advance
I have changed the port of my PostgreSQL server, hence the socket it creates is no longer named ".s.PGSQL.5432".
Is there a way to update/force Rails/the pg gem to change the socket name it looks for? I have tried uninstalling/reinstalling the pg gem in line with recommendations but it still tries to look for a socket named .s.PGSQL.5432, not the .s.PGSQL.50000 that it should and as per the database.yml comments, the 'socket' config line does not apply for PostgreSQL.
I can obviously create a symlink, but I was hoping there was a more elegant way around this.
Change the port in your database.yml
PostgreSQL supports two kinds of connections as you are aware: tcp/ip and unix domain sockets. Both can use host and port but use them in different ways.
When working with TCP/IP, host and port have the standard meanings with regard to networs running TCP/IP.
When working with Unix domain sockets, host provides the directory, and port provides the numeric portion of the socket name. Paths should be absolute. The same should hold true for any libpq-based connection (including through the pg gem), so consider this:
psql -h /tmp -p 5000
This gives an error message of:
$ psql -h /tmp -p 5000
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.5000"?
Similarly:
$ psql -h /foo -p 50000
psql: could not connect to server: No such file or directory
Is the server running locally and accepting
connections on Unix domain socket "/foo/.s.PGSQL.50000"?