Server errors are no fun.

And it is especially frustrating when the error is cryptic, like “Unknown instance“.

What does it mean, really?

A common error people face in VPS, Cloud and Dedicated servers is MySQL failing with “stop: Unknown instance:“

# service mysql restart
stop: Unknown instance:
start: Job failed to start

What is MySQL “stop: Unknown instance:” error?

When MySQL is restarted, the service scripts will attempt to terminate running MySQL programs, and start the program afresh with the config reloaded.

This goes without a hitch 99.999% of the time, but it is known to fail after server upgrades, migrations or config changes.

When the service script is either unable to kill the existing MySQL processes, or restart the MySQL processes with the new config, the restart will fail with the message:

stop: Unknown instance

Here at Bobcares, we often see this error as part of our Server Management Services for web hosts, digital marketers and other online service providers.

We’ve seen causes as varied as resource bottlenecks to security restrictions for this error.

Here are the top 5 causes for this error, and how we fix them:

1. Out of memory

MySQL uses something called global_buffers and thread_buffers to procure memory from the operating system.

When MySQL doesn’t get the requested memory from the system, it fails to start up, causing the restart error.

Solution : We resolve this by optimizing the web and MySQL servers to use lesser memory. We adjust concurrent connection settings, and buffer sizes to reduce the memory footprint, and allow MySQL to start again.

2. Permission errors

MySQL needs full read and write permissions to all directories in which it stores data files.

During server migrations or upgrades we’ve seen the permissions and ownership of these folders accidentally change, causing MySQL self-health checks to fail.

Solution : To fix this, we reset the right permissions and ownership (mysql:mysql ownership, 640 permissions for files and 750 for folders) recursively on all MySQL data directories.

3. Undead MySQL program

For a MySQL restart to be successful, the existing program should terminate and release all locked files (eg. Socket file, data files, etc.).

During high server load or during resource bottlenecks (eg. disk, memory, net, etc.) the MySQL program won’t exit normally, and this will prevent a proper MySQL restart.

Solution : To solve this, MySQL can be hard killed (using command kill -9), which will forcefully liberate all resources from the process, and terminate it. If that doesn’t fix it, renaming or deleting the MySQL socket file will force the program to quit.

4. IP or port binding issues

MySQL will need to listen to an IP address to listen for queries.

We’ve seen firewall updates or server upgrades causing IP and port binding of MySQL to be blocked, thereby causing the MySQL startup to fail.

Solution : There are two ways to fix this:

• Fix the firewall rules – If it is indeed a firewall rule issue, we track the erring rule down, and change it so as to allow MySQL connections again.
• Change binding IP – In case the external IP binding is no longer possible, we change the IP to the localhost (aka 127.0.0.1) to allow MySQL start up.

5. MySQL config changes

Usually after an upgrade or a migration, we’ve seen MySQL config files to be rest to the default version.

This causes MySQL to not recognize InnoDB databases, look for log files in the wrong folder, and more. Any of this will lead to MySQL startup failure, and thereby cause the restart error.

Solution : We keep a system and config file backup in all servers we maintain. So, in such cases, we restore the MySQL config file so that everything works as before.

Security restrictions, file system errors, and more..

MySQL startup errors can be caused by many other issues that includes, AppArmor security restrictions, file system errors like read-only mount, and more.

Here at Bobcares, we find solutions to this in minutes by following the evidence trail left by MySQL programs and log files.

If you need assistance to fix the MySQL error, click here to talk to our MySQL experts. We are online 24/7.

Conclusion

MySQL’s “stop: Unknown instance” error can be caused by resource limits, permission errors, security limits and more. Today we discussed the top 5 reasons for this error and how our Support Engineers fix them.