Clarity regarding when a reboot of a Unitrends appliance may be required.
Requests often come to Unitrends support asking "How often do I need to reboot my Unitrends Appliance?" or "Should I reboot after an upgrade?"
This article will clarify Unitrends best practices on this topic.
The Unitrends Appliance, whether virtual or physical, runs on a custom tuned and configured Linux OS. This OS is designed for long uptime and high stability.
It is generally NOT recommended to reboot Linux outside of specific circumstances.
Rebooting a Unitrends appliance prior to or after an update is not required, unless otherwise instructed to in the release notes (which is rare). There is also no recommendation to reboot simply because of how long a unit has been running.
A Unitrends appliance should only be rebooted under the following conditions:
The release notes indicate a reboot is specifically required after upgrading.
This is rare, and is conditional on the unit receiving not just Unitrends software updates but also OS Kernel updates. In order make the new OS kernel effective, a reboot is required. However, the replacement of this kernel does not require an immediate reboot. Existing backups will continue, but at your convenience the reboot should be done to install the updated kernel.
A UI alert indicates a reboot is required.
If you do not reboot after one of these updates noted above, the UI will notify you with an alert the next day or in some cases your next logon, and this alert will also appear in your daily summary report. If you are not receiving this alert, then there is no reboot pending and no action is required. Commonly this alert may occur if you update from a significantly older release to a newer release skipping several intermediary versions. The release notes for a current release may not indicate that a reboot is required, but one of the prior releases you skipped may have included that requirement. This UI prompt is a failsafe to ensure you receive this notification where appropriate. Again, backups should continue normally until you have the opportunity to reboot.
Unitrends Support indicates your unit is hung and non-responsive and a reboot may be needed to return the unit to stability.
This can occur most commonly when a Unitrends Appliance has storage bus issues. Most commonly, an SATA or USB archive drive or iSCSI mount was connected that was unstable. In some cases VM host interruption, storage failover, or other environmental impacts may also cause a virtual appliance to experience similar issues. There will typically be indications of this in the linux /var/log/messages log or dmesg output in the shell console. If your unit is hung or unresponsive, you shoudl have unitrends support review it's status before rebooting. If you are experiencing environmental issues that result in our appliance instability, a reboot may resolve that symptom, but we want to be sure we address the cause of that symptom where possible to prevent a repeat of the issue.
Planned maintenance in your environment.
If you are planning to bring infrastructure down in your environment that would directly impact a unitrends appliance ability to run, we recommend to power down the appliance. Typically for physical appliances this would only be when you intend to disconnect power from the unit. For Virtual appliances this may be due to host maintenance or planned interruptions between the running UB and attached storage devices.
To safely shut down or reboot a unitrends appliance when necessary, reference: Shutdown or restart a Recovery Series appliance
If you are unable to access the UI, it is recommended to contact Unitrends support for further diagnostics. Rebooting can lose precious status information that could be used to troubleshoot the system if it is hung or unresponsive.
If it is necessary to reboot, the following commands can be run to safely shut down Unitrends services and the unit itself:
/etc/init.d/bp_rcscript stop shutdown now