Expected behaviors and known issues in supported SC//HyperCore releases. Entries apply to all supported releases unless they name a version.
Product: SC//HyperCore · Releases: all supported releases
SC//HyperCore expected behavior
- Updates may include CPU microcode updates and changes to virtualized CPU features. If a CPU feature is unavailable to the whole cluster or is being deprecated, VMs that have access to it need to be power cycled to keep the cluster reliable. Affected VMs are tagged as needing a power cycle, and may have difficulty live migrating until they are.
- From SC//HyperCore 9.6, REST API HTTPS requests require TLS 1.3. TLS 1.2 is deprecated to improve security. This may affect scripts, older PowerShell versions on Windows Server 2019 and earlier, and third-party integrations.
- From SC//HyperCore 9.1.24, SSL certificates under 2048 bits are deprecated. If no custom certificate is in place, a new default certificate is created during the update. If a custom certificate is in use, it must be greater than 2048 bits, or the update fails its pre-check and won’t run until the certificate meets the standard.
- Taking a snapshot with a node offline in the cluster won’t succeed. Snapshots are blocked until all nodes have rejoined the cluster.
- Shutting down a VM from within the guest at certain points in an update can affect its progress. Don’t change a VM’s state from within the guest while an update is in progress.
- Replication from SC//HyperCore versions earlier than 9.2 isn’t supported. If you have questions about replicating from systems on 9.1 or earlier, contact Scale Computing Support.
- Two-port NUC-based nodes have issues with DHCP assigning leases over LAN0. The default network configuration for these nodes sets LAN1 as primary to avoid this. Failing over to LAN0 brings the DHCP issues back.
Guest operating systems
- Disconnecting the performance drivers from CentOS 5 blocks incoming connections. Restart the network service on the VM to restore inbound traffic.
- FreeBSD 10, 11 and 12 don’t recognize live-added virtual drives without a reboot.
- Very rarely, Windows 10 may crash during startup. Rebooting the VM resolves the issue.
- Windows drives sometimes show “0” as the serial ID. This is cosmetic and doesn’t affect the system.
- Windows 11 release 22H2 isn’t compatible with SC//HyperCore 9.2. Windows 11 releases before 22H2 work as expected. Listed in the 9.6 release notes.
- Windows 11 needs a clean installation on SC//HyperCore 9.2 and later. Upgrades from Windows 10 aren’t supported and won’t complete. Listed in the 9.6 release notes.
- Windows 11 may show an alert that “Memory Integrity is disabled.” Enabling memory integrity severely degrades performance in Windows 11 VMs on SC//Platform. Scale Computing recommends keeping it disabled.
- Windows 11 and Windows Server 2022 may log Error 27 in the Windows Event Log, indicating TPM issues. This doesn’t affect performance or the TPM implementation.
Alerts
- Squelched alerts are logged in the cluster log. These are typically CRITICAL or WARNING conditions created during events such as a rolling update, which aren’t alerted to the user.
- SC//HyperCore no longer sends email alerts for snapshot quiesce failures. Internal analysis found these alerts were sent excessively without giving actionable insight.
Web interface
- Connecting to the web interface through a node that has lost its backplane connection makes the sign-in prompt hang. The prompt, or a “Disconnected…” message, displays indefinitely. Connect through an online node instead.
- The network activity gauge on a VM card may show activity when the guest OS shows none, or when the NIC is disconnected. The gauge correctly reports all traffic sent to the VM’s vnet interface, whether or not the NIC is connected or the VM is sending traffic. Excessive broadcast traffic shows as activity but is dropped at the vnet interface before it reaches the guest.
Snapshots and replication
- The initial replication of a cloned VM doesn’t show progress in the web interface. It reports “transmission idle” while replication is active.
- After you delete a remote cluster connection, you’re prompted to create a new one even if another exists. If you have more than one remote cluster connection, refresh the browser after deleting one, before setting up new replication.
- The initial replication of a VM clone isn’t cloneable. The next snapshot is.
- Replication doesn’t alert you right away when it fails to create new snapshots. An alert is sent only after a snapshot hasn’t replicated for 24 hours.
Virtualization
- Replicating or importing a VM to a cluster with fewer resources can produce clones with higher CPU or RAM than the target cluster supports. Edit cloned VMs before starting them to bring CPU and RAM within the target cluster’s limits.
- Windows VMs may log Event ID 122, “access to drivers on Windows update was blocked by policy.” Installing the VirtIO serial drivers from the Scale Tools installation wizard on the CD drive prevents the event.
- Only the default NIC on a Cloud Unity VM can be disconnected.
- Very rarely, the hypervisor fails to delete a drive within the allotted time when you remove it. Shut down and restart the affected VM to resolve the issue.
- Windows Server 2012 Standard (non-R2) can crash when you live-add drives in quick succession or live-delete a drive. Wait 30 seconds before adding another drive. If the VM crashes, rebooting it resolves the issue.
- VMs using nested virtualization may crash or hang during live migration. Power down nested-virtualization VMs before system updates.
- VMs with the Hyper-V role enabled before SC//HyperCore 8.9, for nested virtualization or containers, might fail to boot. Listed in the 9.6 release notes.
- UEFI VMs freeze or crash when a disk is added or removed while the VM is running. Power down UEFI VMs before adding or removing disks. Power cycling an affected VM restores it.
- Scale Guest Agent Tools may have trouble installing drivers and running the guest agent on Windows VMs. The agent may fail to start, and some drivers may not install correctly. Contact Scale Computing Support for help.
- Compressed disks uploaded in SC//HyperCore 9.4 can’t be expanded. Make sure media you upload is uncompressed. Listed in the 9.6 release notes.