How to migrate an OP5 Monitor appliance from EL6 to EL7.

The purpose of this document is to serve as a guide for customers who are running appliances with EL6 and wish to upgrade their operating system to EL7, through migration to a virtual machine or a new installation on the same machine.

Note that it is not possible to upgrade an installation of CentOS6 in place. A new installation of CentOS7 needs to be set up on either another server (virtual or physical) or replacing the current installation on the same server.

Note on clean EL7 installations and Monitor compatibility

In some situations, you might want to install a clean version of Monitor on EL7 to migrate from EL6 -> EL7, and then perform the upgrade on the new system instead.

Before doing this, please note that Monitor versions older than ~7.3.21 do not support EL7, which means that these versions should always be upgraded in-place before migrating.

Step 1: Upgrade

If you are running an old version of OP5 Monitor, first use yum to update to the latest version available through the repositories. Automatic updates stay in the same mainline: this means that if you are on version 7.x there will not be an automatic upgrade to version 8.

Run:

# yum update

If you are on an older version, it may not support EL7, so updating OP5 Monitor first is always recommended.

Step 2: Verify your current OP5 Monitor version

# cat /etc/op5-monitor-release
VERSION=X.Y.Z

Important: When making and restoring op5-backup files, the versions have to match exactly. Do not import a backup file that was created on another version of OP5 Monitor.

Step 3: Disable notifications and stop OP5 Monitor

You can do this through the Web GUI:
“Manage > Process Information > Options (right corner) > Operations > Disable Notifications”

However, this is a "runtime setting", so to be safe, we recommend you disable notifications directly in naemon.cfg followed by a restart of Monitor:

enable_notifications=0

For more information about naemon.cfg, see Main Configuration File Options.

Step 4: Run the backup

# mon stop
# op5-backup --migrate

From the help:

-g|--migrate

System migration to the next major version of CentOS release
No other options is needed. If you are using this on a 32 bit
system you will automaticly perform an archetecture change.
This is the only time when the 'migrate' module is in use.
If you do not use the 'migrate' module you need to specify
mode manually with -m charch if you are moving from a 32 bit
system

(...)

To make a backup to be used when migrating to a new major system
where you want to restore on the same server after you have
reinstalled the operating system. This backup will include all 
backup modules available. Which means that you will have not only
the op5 products in the backup but also a system backup that will
be used in the restore. This will not include all kinds of system
settings. Just the ones needed by op5 products

For more information on what the different modules back up, inspect the scripts under:

/etc/op5-backup/modules

Step 5: Inspect the backup log and verify the backup

The logs from op5-backup runs can be found in:

/var/log/op5-backup/

For example, something like:

# grep -i error backup-20200311-094153.log

Verify the backup file:

# tar tvf <backup-file>
# op5-restore -l -b <backup-file>

Step 6: Transfer the backup file to your new server

Step 7: Restore the backup file

If you are migrating to a new server, shut down the old server before running the restore procedure, as it will re-use the IP configuration of the old server.

Before running the restore procedure, manually verify that the new server has the IP address you expect and that the required ports are open. OP5 Monitor normally requires two-way communication on ports 22 (SSH), possibly 161 (SNMP checks) and 15551 (Merlin) to all nodes.

Run op5-restore:

# op5-restore -b <backup-file>

When the restore is complete, reboot the server.

Verify functionality in the GUI. Verify e-mail and SMS notifications, if configured.

If checks are failing, make sure that you have installed any necessary dependencies that may have existed on the old server, for example, the VMware SDK, Dell OMSA, or any custom Perl modules.

If everything looks all right, go ahead and re-enable notifications again:

Ninja Web GUI: “Manage > Process Information > Options (right corner) > Operations > Enable Notifications” 

You may also need to re-enable notifications in naemon.cfg.