On your Nutanix and vSphere cluster you may have recently seen this alert stating “Storage Containers are not mounted on all nodes”. This may have started showing up after some cluster modifications or maintenance, or could have appeared after some recent upgrades.
You’re probably wanting to know, what is this alert for? How big of a problem is it and what do we need to do to fix it? I’d like to answer those questions for you.
This alert is generated through a recurring NCC check “Storage Container Mount Configuration”. First off it’s useful to know that the main purpose of this check is to validate the configuration of a Metro Availability pair of Nutanix/ESXi clusters. That’s why this check is categorized under the data protection checks. The command to run this check from the CVM CLI is “ncc health_checks data_protection_checks protection_domain_checks storage_container_mount_check”.
As for the impact of the problem being reported I’d have to say it’s conditional. The briefest way of putting it is “a host which doesn’t have the container mounted cannot start VMs in that container.” In most configurations we would expect any presented container to be mounted for all nodes in the cluster. Without getting too far into detail on Metro Availability configurations, just know that it’s required to have the protected containers mounted on all hosts in the cluster and thus we have an NCC check verifying this was done.
Since the Metro Availability setup is currently only supported on VMware vSphere we shouldn’t be seeing this check run on Hyper-V or AHV clusters. The check identifies containers included in protection domains and then checks whether they are mounted on all the ESXi hosts in the cluster.
So what do we do about this alert? The simplest answer is to ensure any containers presented to any ESXi hosts in the cluster is mounted in the same manner (same NFS path, same datastore name) on all ESXi hosts in the cluster.
If you’re seeing this alert and you don’t know why, the article “NCC Health Check: storage_container_mount_check” gives the steps to manually validate the conditions reviewed by NCC.
As sometimes happens, a false-positive issue was identified with this check. This has already been resolved in the latest release of NCC as shown in the release notes. Look for “storage_container_mount_check” in the resolved issues section. If you’re running an NCC version prior to 188.8.131.52 I would recommend to upgrade, clear the alerts, and then run a full health check to see if the alert comes back. For details on upgrading NCC look here.