This page covers capacity alerts and advice, the storage gauges and Storage Details views in the web interface, and how to manage uploaded ISO files.
Overview
SC//HyperCore™ HyperVisor software gives you several places in the web interface to check storage capacity. The numbers can be hard to read without some context. That is because the SCRIBE storage engine uses a block reference count model, with thin provisioning and deduplication. This page walks through each storage indicator and what it actually tells you.
For how snapshots and clones affect storage use, see the capacity implications covered in SC//HyperCore snapshots.
The SC//HyperCore system is built to be highly available and to heal itself. But it needs enough free capacity and resources to check and heal itself.
SC//HyperCore alert systems always send email alerts about capacity. The thresholds are:
| Free Capacity | Alert Level |
|---|---|
| 40% free | Informational |
| 30% free | Minor |
| 20% free | Major — Scale Computing Support recommends freeing capacity or adding a node at this level |
| 10% free | Critical — potential stability concerns |
At 90% consumed capacity, this message appears: “Free storage capacity extremely low. VM Creation, Snapshots, Cloning, Importing and Replication are disabled”. If needed, and if Scale Computing Support approves, this block can be bypassed.
Contact Scale Computing Support if you need help finding and freeing capacity.
Scale Computing Support always recommends keeping at least one node’s worth of RAM free. That way the system can still run all VMs while one node is unavailable.
SC//HyperCore VMs use thin provisioning to save storage space. Once space is allocated on a VM disk, it stays used on the system even if the VM’s OS later deletes that data, unless the guest OS sends unmap (TRIM) commands. From SC//HyperCore 9.6, the system reclaims that space. See Unmap Support for Storage Reclamation. Deduplication helps with any data overlap, but it does not change how thin provisioning works.
Individual VM Disk Gauges
Each VM card in the SC//HyperCore web interface has its own gauges for CPU load, network load, IOPS, and used capacity.
The Disk gauge on a VM card shows the combined used and total capacity of all of the VM’s virtual disks. So you can see, for each VM, how much of its provisioned disk space is actually in use.
Use this gauge to find VMs that take up a lot of storage. With the Storage Details dialog, it helps you narrow down where most of the system’s capacity is going.
The HUD DISK Gauge
The Heads Up Display (HUD) at the top of the SC//HyperCore web interface has a DISK gauge. It gives a quick view of the system’s total used capacity, along with IOPS and Latency (LAT) indicators.
Click the DISK gauge to open the Storage Details dialog.
Storage Details
Storage Details gives the most complete view of system storage in the web interface. It has three tabs: Utilization, Efficiency, and Provisioned.
Two terms help when you read these views:
- Physical is the base hardware storage capacity: the RAW capacity of the system.
- Logical is the storage capacity as defined in software. The system keeps one duplicate copy of each piece of data for redundancy. So logical capacity is about half the physical capacity, before deduplication is taken into account. You may see a small offset from this figure because of internal SCRIBE optimizations.
All figures in Storage Details change in real time as the system changes. This is expected, given how SCRIBE’s reference count architecture works.
Utilization Tab
The Utilization view gives an overview of:
- Free space left on the system now
- The current data efficiency ratio, as set by the system’s deduplication feature
- Estimated logical free space left
The Logical side of the graph splits used space into VM Data Usage and Snapshot Usage, so you can see where capacity is going.
Efficiency Tab
The Efficiency view shows how much deduplication is saving on the system right now. It shows logical usage as if there were no deduplication, so you can compare it with the true physical used space on the right side. Space saved by deduplication is shown in green.
Use this view to see how much your deduplication ratio is actually helping. It also helps you estimate what capacity use would look like on a system without deduplication.
Provisioned Tab
The Provisioned view focuses on long-term storage health. It shows:
- Logical view: The capacity provisioned for VM drives, the capacity actually used on those drives, and the capacity used by snapshots.
- Physical view: A projection of long-term storage needs. It is based on current provisioned capacity and uses the current deduplication efficiency ratio.
The Over Provisioned box (shown in red when present) is the extra storage space you would need if every current VM drive filled up. It is a forward-looking indicator for capacity planning.
ISO Files and Media Management
The SC//HyperCore web interface also tracks the size of each uploaded ISO file. ISO files use storage on the system, and they are easy to miss when you look for space to free up.
To view and manage uploaded ISOs, go to the Control Center and select the Media tab. There you can find and remove ISO files you no longer need.
SCRIBE’s reference count model is complex, so all storage figures in the web interface are best-effort estimates, not exact measurements. The figures in Storage Details are always changing. There is no reliable way to know in advance exactly how much space you will free by deleting a given VM or snapshot.
The VM disk gauge and the Storage Details dialog together give the best picture of where capacity is going. They are most useful for finding VMs with large storage footprints. Snapshot capacity is still the hardest to estimate precisely.
For capacity planning, treat the Over Provisioned indicator on the Provisioned tab as your clearest signal that the system is nearing a storage limit.