View previous topic :: View next topic |
Author |
Message |
TheLexx Tux's lil' helper
Joined: 04 Dec 2005 Posts: 137 Location: Austin Tx
|
Posted: Thu Dec 28, 2017 12:37 am Post subject: initd scripts w/o initd |
|
|
I am trying to start services manually without using initd. It seams now a days, the only online documents I can find only talks about starting with initd or systemd. I did try running MySQL and Apache with initd, then simply using ps find the following CLI commands.
Code: | /usr/sbin/mysqld --defaults-file=/etc/mysql/my.cnf &
/usr/sbin/apache2 -D DEFAULT_VHOST -D INFO -D SSL -D SSL_DEFAULT_VHOST-D LANGUAGE -D PHP -d /usr/lib/apache2 -f /etc/apache2/httpd.conf -k start |
After starring and stopping with initd, I was able to start those services using just the CLI command shown above It works as long as a start and stop with initd first.
If on the other hand I reboot, then run those same CLI commands both Apache and MySQL will fail. This indicates to me, that the initd system preformed some operation prior to starting MySQL or Apache. I know that initd will automatically start prerequisite services, but looking at the output of "/etc/init.d/mysql start" no other services were started (same with Apache).
Reading /var/log/mysql/mysqld.err I see a prominent error that happens when I attempt to start MySQL without initd.
Code: |
2017-12-26 16:31:04 1899 [Note] Server hostname (bind-address): '127.0.0.1'; port: 3306
2017-12-26 16:31:04 1899 [Note] - '127.0.0.1' resolves to '127.0.0.1';
2017-12-26 16:31:04 1899 [Note] Server socket created on IP: '127.0.0.1'.
2017-12-26 16:31:04 1899 [ERROR] Can't start server : Bind on unix socket: Permission denied
2017-12-26 16:31:04 1899 [ERROR] Do you already have another mysqld server running on socket: /var/run/mysqld/mysqld.sock ?
2017-12-26 16:31:04 1899 [ERROR] Aborting
|
This error seems to indicate that the initd script does something that effects the permissions assigned to port 3306. Is there any way to figure out what the initd script is doing? I will report more info on the failure of Apache on a later post in this thread. Thanks in advance for any advice.
Version numbers:
sys-apps/sysvinit: 2.88-r9
sys-apps/openrc: 0.23.2
dev-db/mysql-init-scripts: 2.1-r1
dev-db/mysql: 5.6.35
www-servers/apache: 2.4.25
/etc/mysql/my.cnf https://pastebin.com/fwcdBxU2 |
|
Back to top |
|
|
Hu Administrator
Joined: 06 Mar 2007 Posts: 22742
|
Posted: Thu Dec 28, 2017 2:25 am Post subject: |
|
|
Why are you trying to avoid the init system?
You misunderstood the error message. It failed to bind the unix domain socket, not the TCP socket. The failed unix socket is on the next line: /var/run/mysqld/mysqld.sock. Most likely, your /run was made a tmpfs (by the init system ), so rebooting clears the startup work done by the initscript. If you don't use the initscript to start mysql, you need to do the job the initscript would have done. You could find these steps by reading the initscript in /etc/init.d. |
|
Back to top |
|
|
Jaglover Watchman
Joined: 29 May 2005 Posts: 8291 Location: Saint Amant, Acadiana
|
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54594 Location: 56N 3W
|
Posted: Thu Dec 28, 2017 9:45 am Post subject: |
|
|
TheLexx,
init runs things as root.
Code: | 2017-12-26 16:31:04 1899 [ERROR] Can't start server : Bind on unix socket: Permission denied | hinsts that you are not root. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
TheLexx Tux's lil' helper
Joined: 04 Dec 2005 Posts: 137 Location: Austin Tx
|
Posted: Wed Jan 03, 2018 7:59 pm Post subject: |
|
|
Hu wrote: | You misunderstood the error message. It failed to bind the unix domain socket, not the TCP socket. The failed unix socket is on the next line: /var/run/mysqld/mysqld.sock. Most likely, your /run was made a tmpfs (by the init system ), so rebooting clears the startup work done by the initscript. |
Thank you for your help, I did not realize that the key was the file mysqld.sock. Looking around I found that the directory /var/run/mysqld/ as not created until the initd script was run and the a reboot emoved the directory. I am able to get MySQL to run after I create he /var/run/mysqld/ owned by user mysql.
Hu wrote: | If you don't use the initscript to start mysql, you need to do the job the initscript would have done. You could find these steps by reading the initscript in /etc/init.d. |
That is basically that I was I am asking, how do I read the script and figure out what it is doing? The scripts are written in the /sbin/runscript language. Is there some sort of standard debugging that can be used to interpret the /sbin/runscript language? The /sbin/runscript language appears similar to train wreck syntax of the bash scripting language. Even with "post hoc analysis", I was not able to find the line in the script that created the needed directory.
Jaglover wrote: | No offense, but may it be we have an XY problem here? |
Hu wrote: | Why are you trying to avoid the init system? |
All I wish to do is decouple the starting of those two services from the initd system. Back in the 90's and 00's this was a perfectly legitimate way to run those services.
In a different thread, I explained why I wanted to do the decoupling, however I seamed to spend more and more time justifying myself instead of just doing the decoupling. The short answer is, if I do not have to rely on a functioning initd system (in the sandbox environment), it gives me more options for working in a sandbox environment.
I admit that the best way to start servives at boot is to use the initd system. However, right now, I am looking for a sandbox to experiment with. Because of this, manually starting services is not a problem. When I rent a VPS, I will be starting the services with initd. |
|
Back to top |
|
|
Hu Administrator
Joined: 06 Mar 2007 Posts: 22742
|
Posted: Wed Jan 03, 2018 11:56 pm Post subject: |
|
|
As I understand it, runscript, like ebuild, is fundamentally bash (and is executed by running GNU Bourne Shell), but with various standard libraries sourced in automatically before the script begins evaluation. |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Thu Jan 04, 2018 10:00 am Post subject: |
|
|
Hu wrote: | As I understand it, runscript, like ebuild, is fundamentally bash (and is executed by running GNU Bourne Shell), but with various standard libraries sourced in automatically before the script begins evaluation. |
Hu ... actually, posix sh, so not like ebuilds (which are bash). All initscripts should not contain bashisms for this reason (though because /bin/sh is linked to /bin/bash by default they tend to pass).
TheLexx wrote: | That is basically that I was I am asking, how do I read the script and figure out what it is doing? The scripts are written in the /sbin/runscript language. Is there some sort of standard debugging that can be used to interpret the /sbin/runscript language? The /sbin/runscript language appears similar to train wreck syntax of the bash scripting language. |
The standard shell method, 'set -x ; <section of code to be debugged> ; set +x'.
TheLexx wrote: | Even with "post hoc analysis", I was not able to find the line in the script that created the needed directory. |
It's this:
/etc/init.d/mysql: | checkpath -d --owner mysql:mysql --mode 0755 "$piddir" |
HTH & best ... khay |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|