Linux – How to start Oracle Enterprise manager after restarting the Red Hat Server

If you ran make install and already had the mysql-server package installed, chances are you had a pre-existing mysql database schema in /var/lib/mysql. The version included with RHEL5 is 5.0.77, which could explain the mysql-upgrade error, as that script updates the schema between MySQL versions. The error could also be very similar if you did not create the initial schema as part of your installation.

If you're going to compile from source and not specify installation locations, you need to remove the system packages of mysql first. If you want to maintain both the system packages and the source installation, which there is sometimes argument for, you need to be careful not to stomp all over the system package. That typically means installing your source in a location such as /usr/local/mysql as opposed to the default of /usr/local. You can accomplish this with ./configure --prefix=/usr/local/mysql. There are more details here and best practices that I'm not going to elaborate further on.

While there can be good reasons to compile from source, you might reconsider yours, as it may not be necessary for you.

rpm -qa | grep mysql will display all packages currently installed. Probably mysql-server and mysql-client.

Otherwise, you didn't follow the procedure from the documentation. Even if you did, you will have to repeat it after removing the MySQL packages.

From INSTALL-SOURCE, which was included in the mysql-5.1.44 tarball:

2.3.1. Source Installation Overview

   The basic commands that you must execute to install a MySQL source
   distribution are:
shell> groupadd mysql
shell> useradd -g mysql mysql
shell> gunzip < mysql-VERSION.tar.gz | tar -xvf -
shell> cd mysql-VERSION
shell> ./configure --prefix=/usr/local/mysql
shell> make
shell> make install
shell> cp support-files/my-medium.cnf /etc/my.cnf
shell> cd /usr/local/mysql
shell> chown -R mysql .
shell> chgrp -R mysql .
shell> bin/mysql_install_db --user=mysql
shell> chown -R root .
shell> chown -R mysql var
shell> bin/mysqld_safe --user=mysql &

I just got into the same trouble.

In fact, this is definitely a bug in OEM. In my case, the amount of space on the filesystem that oracle was using was 4.5T and the free space was 3.77T. Something was definitely overflowing, but since this was a Solaris 11 on a zfs filesystem, I did this simple trick: I created the separate zfs dataset, set it's refquota to 2GB, mounted it in the ${ORACLE_HOME}/hostname_${ORACLE_SID}/sysman/emd/upload, et voila ! It started seeing the correct free space and started uploading.

I have no idea so far about what I should do on Linux, wich doesn't have a production-ready zfs.