[BlueOnyx:04642] Re: Help with recent Installs (/var Full)(Mysql quit)

[Jeff Folk]

On May 24, 2010, at 8:17 PM, publictv at cox.net wrote:

> Ok the mysql database, phpmyadmin, and The ftp service are all  
> working again!
> I ran the last 50 intries again then opened both of my websites that  
> use mysql and then conected to the server by Filezilla
> and everything is working! and after each step I reran the last 50  
> intries and still received the same responce which I have included  
> below...   I am pretty sure theat I should not be getting the  
> warnning messages and I was wondering not only how to fix it but  
> could it have anything to do with the problem in the 1st place or  
> was the problem cuasing the warning condition to happen?

Since there is now room to create the pid file, mysqld starts.

> The gui still says that I have 95% used in the /var is that because  
> the copy is still there?

You'll have to investigate what is taking so much space on /var (use  
ls -lah in all the directories to find large files). You don't want it  
that full. If you followed my instructions to the letter, your  
mysqld.log file should be very small and the copy should be over in / 
home/users/admin.

Try:

du -k /var/ | sort -nr | more

That should give you (page by page, space key for next page, return  
for next line) the listing of the directories in /var, and their size  
in kilobytes, largest first. That will help you track down the space  
hogs...

> Also the rm comand for to remove the copy of the mysql log file  
> could you give me that as I dont want to take a chance and guess  
> that I know what I am doing and mess it up again by syntaxing it  
> rong or incorect path and end up deleting the wrong file!

That's why I tried to keep you in /home/users/admin to copy the file,  
then you could perform the rm with just the file name...

[root at box1 admin]# rm -f mysqld.log

notice the   ^ (admin)? If it said 'log' instead, you surely would  
delete the wrong file, but mysql will recreate it anyway. I only had  
you cat the empty file into the existing file so we didn't have to  
bounce back and forth when tailing the log file.

To see if there are multiple copies, and where they are, type locate  
<filename> after updating the file database. If you followed my  
instructions, your session should look like this:

[root at box1 admin]# updatedb
[root at box1 admin]# locate mysqld.log
/home/.users/112/admin/mysqld.log (essentially the same as /home/users/ 
admin because of symlinks in system)
/var/log/mysqld.log

> 100524 19:45:14  mysqld restarted
> 100524 19:45:15 [Warning] option 'max_join_size': unsigned value  
> 18446744073709551615 adjusted to 4294967295
> 100524 19:45:15 [Warning] option 'max_join_size': unsigned value  
> 18446744073709551615 adjusted to 4294967295
> 100524 19:45:15  InnoDB: Started; log sequence number 0 43655
> 100524 19:45:15 [Note] /usr/libexec/mysqld: ready for connections.
> Version: '5.0.77'  socket: '/var/lib/mysql/mysql.sock'  port: 3306   
> Source distribution

If that is the current contents of the log file, it is the same as  
mine when mysql starts up. So...

How big is your /var partition? My server is using 365 MB (after 16  
months) of a 3.6 GB partition. I would think a 700 MB /var partition  
would be minimum, especially if you plan on processing mail. I don't  
recall the standard partition scheme in BX, but I think /var is  
usually north of 1 GB by default.

Jeff