Each quarter EMC Isilon produces a technical newsletter called the Uptime Bulletin. This newsletter provides product tips and announcements. You can access our first issue of the EMC Isilon Uptime Bulletin (PDF) on the EMC Online Support site (login required).

The first issue includes the following Isilon OneFS tip about upgrading.

Reboot your cluster before an upgrade

Before upgrading to the latest version of OneFS, it helps to be prepared. Assessing compatibility, gathering cluster information, and following best practices during the pre-upgrade window will ensure that the upgrade process goes smoothly.

One of the pre-upgrade steps that EMC Isilon recommends is to reboot your clusters. A reboot ensures a clean state for the upgrade process, and it usually takes fewer than fifteen minutes for simultaneous reboots (where all the nodes in a cluster are restarted at the same time). This recommendation is particularly applicable to clusters that have run a long time without a cluster-wide reboot event.

A cluster reboot can help you to:

  • Identify failing hardware
  • Ensure that cluster configurations, such as share and export configurations, are saved and persist through a reboot
  • Distinguish among problems that might occur either during the reboot process or during the upgrade process

Overall, a pre-upgrade reboot can make it easier for EMC Customer Services to diagnose and resolve any upgrade-related issues quickly. In an ideal upgrade window, you should also test client services after a reboot and before the upgrade. This gives you the opportunity to validate the upgrade verification tests.

For more information about the OneFS upgrade process

Visit the EMC Online Support site to access the following PDF guides (login required), which help you prepare for the upgrade process for OneFS 6.5–7.0:

If you have questions about the upgrade process, please visit the Contact EMC Technical Support site (login required) for contact information.

 

Kirsten Gantenbein

Kirsten Gantenbein

Principal Content Strategist at EMC Isilon Storage Division
Kirsten Gantenbein
Kirsten Gantenbein
Tags: , ,

6 Comments

  1. Vlad says:

    Reboot cluster before an upgrade doesn’t comply to OneFS nature: upgrade shouldn’t bring cluster offline.
    – Rebooting cluster – all clients are disconnected. Cluster is offline, but upgrade shouldn’t interrupt service and cluster has to stay online.

    • Hi Vlad. It’s true that a rolling upgrade will not bring a cluster offline. However, this article suggests that rebooting your cluster — for example two weeks before an upgrade — can help identify any issues that might hinder the upgrade process. Also, upgrades to major releases of OneFS require a simultaneous upgrade, which will bring the cluster offline. Therefore, it’s a best practice to plan maintenance and upgrade windows to help the upgrade process go seamlessly.

  2. Jake says:

    Are you serious with this advice? The sales folks are quick to claim zero-downtime as all updates and maintenance can be done in rolling fashions, but this advice is 100% conflicting.

    So who should us customers trust – the sales people or the technical people who have to make it work? Sadly my experience is the later – seldom do OneFS upgrades go smoothly and without downtime.

    • Hi Jake. Sorry about the confusion around rolling upgrades. Rolling upgrades and rolling reboots are available, but it’s best to refer to release notes for the version of OneFS that you plan to install to confirm that you can do a rolling upgrade. In the future, we’ll also add information about rolling and simultaneous upgrades to the Isilon Support & Compatibility Guide.

Leave a Comment

Comments are moderated. Dell EMC reserves the right to remove any content it deems inappropriate, including but not limited to spam, promotional and offensive comments.