Licensing Across Data Centers

What does the Oracle 10 day rule cover?

The Oracle 10 day rule covers a single unlicensed failover node in a cluster that can take over for a failed licensed node for up to a total of 10 separate days in any calendar year. It is the one genuine piece of disaster recovery relief in standard Oracle licensing, and it is narrow. The rule applies to one node, the failover must be to that node, and the days are counted as separate calendar days regardless of how many hours the node ran on each. On the eleventh day, full licensing applies.

Buyers misread this rule constantly. It does not give you a free standby that you can run whenever you like. It does not cover a second node that is routinely active. It does not stack across multiple failover targets. It is a single cold node, in the same cluster, used only when the primary fails, capped at 10 separate days a year. Read against your specific cluster topology, the rule either applies cleanly or it does not, and the difference is the difference between zero and a full second licensing position. The broader metric mechanics sit in the Oracle database licensing guide.

Does an Oracle standby database need a license?

An Oracle standby database needs a license whenever it is open, active, or otherwise capable of running beyond the narrow failover relief, which covers the great majority of Data Guard configurations. The software is installed on the standby host and can run, so it is in scope under the same metric as the primary. Active Data Guard, where the standby is open for read or reporting, always needs licensing, and the Active Data Guard option itself is separately licensable on top of the base database.

The common error is to treat any standby as covered by the 10 day rule. It is not. The rule is for a passive failover node, not for a replicated standby that the database keeps current and that serves a real operational purpose. If your standby is part of normal resilience rather than a cold node waiting for a single failure, plan to license it fully. The metric that applies is the same one you chose for production, explained in processor versus Named User Plus.

Does remote mirroring need an Oracle license?

Remote mirroring generally needs an Oracle license when the mirror lands on storage attached to a host where the database is installed or could run, because the test is capability to run, not present activity. Storage level replication feels like a pure copy operation, and teams assume a copy is not a deployment. Oracle reads it differently. If the mirrored data can be mounted and the database started at the remote site, that site is a place Oracle can run, and a finding will treat it as licensable.

The buyer move is to understand exactly what sits at the far end of the mirror. A mirror to dark storage with no Oracle binaries and no host that could start the database is a different case from a mirror to a warm site with the software installed and ready. Whether a given mirror is licensable is contract dependent and configuration dependent, so it is one of the points where reading your specific agreement and inspecting the actual topology matters most.

Do backups and cold copies need a license?

Backups and genuinely cold copies usually do not need a license, because there is no installed, runnable database, only stored data. A tape, an object store, or a backup file is data at rest, not a deployment, and a finding that tries to license pure backup storage is usually overreaching. The line is whether the software is present and capable of running. A backup that simply holds blocks is on the safe side of that line.

The error appears when a backup environment quietly becomes something more, a restore target kept permanently mounted, a recovery host with the binaries installed and ready, a reporting copy that someone started. Once the database can run there, the environment crosses from backup to deployment. Documenting that your backups are genuinely offline, with no installed runnable instance, is part of the evidence that defends the position, the kind of documentation set out in building your own deployment inventory.

What disaster recovery mistakes do auditors find?

The disaster recovery mistakes auditors find most often are standby nodes treated as free, failover nodes run beyond 10 days, and warm DR sites assumed to be backups. Each turns on the same misunderstanding, that the purpose of an environment changes its licensing, when in fact the test is whether the software is installed and capable of running. Auditors do not need to prove you used the DR site. They only need to show it could run.

• A second standby treated as covered by the 10 day rule when it runs continuously
• A failover node that exceeded 10 separate days in a year during an extended outage
• A warm DR site with the binaries installed, assumed to be a passive backup
• An Active Data Guard reporting replica missing the separate option license
• A remote mirror to a host that could start the database, counted as storage only

The defense in every case is the same sequence: read the agreement, map the real topology, document what can and cannot run, and reconcile the script output to that documented reality before it leaves your organisation. Submitting raw script output, which sweeps every host with the binaries installed, concedes the DR estate before any conversation starts. Reviewing first is covered in what data to share and what to withhold.

A worked example

Consider an anonymized insurer running a primary site of sixteen licensed cores with a remote DR site of the same size. The DR site held a physical standby kept continuously current, which the insurer had assumed was covered by the 10 day rule.

The continuously current standby did not qualify for the 10 day rule, so part of the finding stood. The insurer reconfigured a single node as a genuine cold failover target, documented the change, and licensed the remainder honestly, cutting the defended position by roughly 72 percent, within the 60 to 80 percent range a line by line review typically achieves. This example is illustrative and anonymized, and outcomes depend on your estate, your contract and your evidence.

Your next step

Disaster recovery licensing is where good intentions meet a strict rule, and the gap between them is expensive. An independent buyer side review maps your DR topology against the 10 day rule and your specific contract, then gives you a position you can defend or a reconfiguration that saves real money. Our advisors work on a Fixed Fee or Gainshare basis with no risk to you, and we reduce your Oracle exposure or we reimburse our service fee.