Ensure servers reboot successfully
This article shares tips that help you ensure that server reboots are fast and minimally disruptive.
Ensure that server backups are configured and running
We recommend that you use backups to keep your data up to date. Rackspace Cloud Backups run a differential backup on a set frequency. You can set a cloud backup to run on any number of directories.
Important: On performance cloud servers that have a data disc, the data partition is not included in any image snapshots that you take. Ensure that these data partitions are included in your backup scheme.
Back up Linux
If you are running a Linux® server, you might need to back up the following directories:
- /etc (This directory contains most of your configuration files.)
- /var/www (This directory often contains your websites and files.)
- /var/lib/mysqlbackup (Servers built by using Rackspace Managed
Operations have an automated process that automatically runs a
mysql dumpto this folder.)
Back up Windows
If you are running a Windows® server, we recommend that you back up locations where your data might be stored, such as the following locations:
- Any additional drives such D: and E:
Back up live databases
Cloud Backup does not back up live databases. You must back these databases up by using Microsoft SQL Server Management Studio.
We recommend that you carefully consider your specific applications and their backup needs.
Ensure that services are configured to start after boot
When you install a service, it is not automatically configured to restart after the server is rebooted. You must configure services to restart after the server reboots. To learn how to perform this task, select the following resource that corresponds to your server’s operating system (OS):
- Windows: Prepare to migrate a Windows server
- Ubuntu® operating system: UPDATE-RC.D(8)
- RHEL® and CentOS®: Using the chkconfig utility
Ensure that iptables or Windows Firewall rules are saved and configured to restart on reboot
It is important to ensure that the firewall rules that you configure stay active upon reboot. This section shows you how perform this task.
We do not recommend using a passphrase when you generate a Secure Sockets Layer (SSL) certificate. However, if you already have a passphrase in place for your SSL certificate, you need to input that into the server when it reboots. The services on the server cannot start until you enter that passphrase.
Ensure that Cloud Block Storage volumes attach on reboot
If you have data on a Cloud Block Storage volume attached to a next generation cloud server, you must ensure that any volumes are properly connected after a reboot.
To do this, you need to add your volume to the static file system information
fstab file. You cannot attach Cloud Block Storage volumes to a first
generation cloud server.
Note: If you are using Windows, mounted block storage should remain mounted after the server reboots.
FSCK (File system consistency check)
A file system consistency check (
fsck operation) generally runs
automatically at boot time. Two common triggers automatically
fsck. Either the OS detects that a file system is in an
inconsistent state (due to a non-graceful shutdown such as a crash or loss of
power), or after the number of times that the system is mounted passes a
After you reboot your server, this check might happen automatically. If it does, it might increase the delay for your server coming back online. While delays are usually negative experiences, in this case the delay might save your server. We recommend that you allow the file system check to complete even though it might cause delays. If you attempt to reboot the server again, it resumes the file system check and extends the delay.
So that you know how your servers and other cloud products react during and after a reboot, we strongly recommend that you take a few minutes to test the entire process of getting your environment back up and running.
We recommend that you perform testing during the development phase or on separate servers to limit any customer impact.
Mitigate the impact of a reboots
Use the tips in this section to reduce the impact of server reboots.
One of the best ways to prevent the prolonged impacts of a reboot is to distribute your application over multiple redundant, tiered servers. This approach, which is called Horizontal scaling, is a great way to minimize the risk of downtime due to a single server going down.
Custom error pages
Using a cloud load balancer gives you the ability to set a custom error page in the event that a server connected to the load balancer is offline or unresponsive. When you proactively configure that error page, a visitor to your site receives an error message that is specific to your unique application.
©2019 Rackspace US, Inc.
Except where otherwise noted, content on this site is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License