MySQL Bugs: #37833: Setup problem

Bug #37833	Setup problem
Submitted:	3 Jul 2008 9:39	Modified:	28 Oct 2011 12:03
Reporter:	Nguyen Thanh Hai	Email Updates:
Status:	Can't repeat	Impact on me:	None
Category:	MySQL Server	Severity:	S5 (Performance)
Version:		OS:	Linux
Assigned to:		CPU Architecture:	Any

Description:
Error when I finish install:
I have copied mysql.server to /etc/init.d/mysql, then start MySQL with below command:
/etc/init.d/mysql start
So, I can't start mysql, it fails. And this is error notify:
"Manager of pid-file quit without updating file"
How can I fix this problems?
Thanks very much!

How to repeat:
Error when I finish install:
I have copied mysql.server to /etc/init.d/mysql, then start MySQL with below command:
/etc/init.d/mysql start
So, I can't start mysql, it fails. And this is error notify:
"Manager of pid-file quit without updating file"
How can I fix this problems?
Thanks very much!

Thank you for a problem report. What exact version of MySQL server, x.y.z, do you use?

I have the same error using 5.1.25. I'm running this on a virtual instance and I think it is possible that because the image was saved with MySQL active and running, the pid file is in a weird state that confuses MySQL when a new instance is launched at a later time and MySQL next tries to run and finds a pid file with a value in it.  In any case, I haven't found a work around for this yet.

For me, this was a permission issue.  The problem was resolved when I ran chown -R <your mysql user>:<your mysql group> /<path-to-mysql>.

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

I have seen this problem regularly when trying out different configurations of my.cnf files.

I think it is a mysql bug - it is very fragile to this PID file problem.

I forgot to mention my platform as OS X 10.5, Intel.

Mike,

Please, explain what exactly you do on Mac OS X after installation and copy-paste all the error messages you get. What MySQL server version do you use?

I'm using server 5.1.30.

It turns out that there is an option in my.cnf the server does not like. I have not yet determined which one.

I had a database configured and operating fine, and I have changed some options-file settings (a lot of settings) and now my server is either not starting, or when it does start, it is not recognising InnoDB as an engine (registration fails).

I should drop all of my databases and rebuild... but it is worrying as I only intended to tweak performance parameters.

I suspect I have caused InnoDB to modify things so much that now when it starts it reports that it cannot register InnoDB as a storage engine.

Perhaps this is not so much a bug as user error... however if mysql added a more useful error report in the error file, that would help... :)

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

I am having this same problem.  I have OSX 10.5.7.  According this thread this problem hasn't been resolved.  I assume after all this time somebody figured it out.  Maybe I have to uninstall everything and try again.

I have the same problem.  Because no one answers in a month does not mean that this is still not a problem.  We need a solution.  Thanks.

Same problem here.
Just installed mysql-5.1.41-osx10.5-x86_64.pkg on SnowLeopard 10.6.2
Use the workbench to start the server, then created a database.
Changed the config, so I shutdown the server, and now it won't start up.

At least make this message more informative, I have no clue where this pid file is, what I have to do with it, etc...

The error message about pid is misleading, it's just a consequence of a startup failure. The real reason of the crash is that the new my.cnf sets a different innodb_log_file_size (default value is way too small by modern standards). You can see this in the error log:

InnoDB: Error: log file ./ib_logfile0 is of different size 0 5242880 bytes
InnoDB: than specified in the .cnf file 0 268435456 bytes!
100426 13:12:03 [ERROR] Plugin 'InnoDB' init function returned error.
100426 13:12:03 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
100426 13:12:03 [ERROR] Unknown/unsupported table type: InnoDB
100426 13:12:03 [ERROR] Aborting

Solution: You must erase the existing ib_logfile0 and ib_logfile1 in /var/lib/mysql (don't worry that won't kill any data), then restart MySQL and it will create new log files with the size specified in my.cnf.

Hops this helps,

-- JFG

I had the same problem; worked initially, but not on 2nd start (installed via MacPorts). I had to

chown -R _mysql /opt/local/var/db/mysql

as it was owned by root. Perhaps that'll work for some of you folks having problems as well. I would say this is a bug in the MacPorts MySQL 5 package, not really MySQL itself.

I have the same problem. I downloaded mysql 5.1.1, installed it (in OS X 10.6.4), started it (as root?), created tables, loaded data. Then I restarted my mac, and tried to restart mysql with "sudo /Library/StartupItems/MySQLCOM/MySQLCOM start" from the prompt, and also using MySQL Workbench. Same negative result.

I had not changed any settings, or any file ownerships.

I have now tried 
sudo chown -R _mysql /usr/local/mysql/bin/mysql
and
sudo chown -R mysql /usr/local/mysql/bin/mysql
(gives same result?)
but that does not change anything, i.e. I still get the same error message when I try to start the database.

All reporters,

Please, check if the problem is still repeatable with current versions of MySQL server, 5.1.54 or 5.5.8, MySQL binaries. If it is, please, specify the exact OS version and URL used to download the binaries.

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

For me, this was the error:
[27587/47483191445840] [11/02/15 23:10:55] [ERROR] Manager: password file (/etc/mysqlmanager.passwd) does not exist.
Segmentation fault

And the solution:
touch /etc/mysqlmanager.passwd

Dear Balazs Vamos,

What exact version of MySQL server, and on what OS, you had used?

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

This happens on CentOS when SELinux is in use and you try to have your data files somewhere else than /var/lib/mysql.

The problem is that you don't get any clue to what went wrong:

110614 18:39:07 mysqld_safe Starting mysqld daemon with databases from /data/mysql/var
110614 18:39:07 mysqld_safe mysqld from pid file /var/run/mysql/mysqld.pid ended

(Usually I expect to see some helpful error message between those two lines.)

If you do "/usr/sbin/setenforce 0" and then start MySQL, it works.

Suggested fix:
Provide helpful error message when SELinux is preventing MySQL from writing to data files.)

On a variant of 5.1.52 I got this error by adding to my.cnf:
skip-name-resolve=1

The correct line is
skip-name-resolve
...which makes the error go away and mysqld starts normally again. AFAICS there are no other error messages in any logs, apart from this misleading message. (But it could be the server is using non-standard init script, in particular mysqld_multi.)

All reporters,

Please, concentrate on Linux OS only here. If you still have repeatable test case with recent 5.1.60+ or 5.5.17+ versions of MySQL servers, please, re-open this bug. For all other OSes, namely Mac OS X, please, open new bug reports.

In many cases this is related to file permissions on Linux or SELinux settings. I see no point to mix all other cases on top of that in one report.

Posted by developer:
 
No feedback on how to reproduce