Compare systems against a golden image

[ Applies to ] StorageGuard 9.2.8 and later  /  All supported storage, backup, SAN and cloud systems  /  Windows and Linux deployments

A golden image is a system whose configuration you've approved as the standard for its type: the PowerScale cluster that's set up exactly right, the Data Domain built from the hardening checklist. Comparison groups check other systems against it, line by line, so drift shows up as a difference you can read, not as a finding you have to interpret. Baseline policies catch violations of rules; golden image comparison catches anything that differs.

Choose the golden image

• One per system type. A comparison group covers one platform, so pick a golden image for each platform you want to control.
• Make it clean first. Fix the golden image's own findings before you use it as the reference, or you'll propagate its mistakes as the standard.
• Keep it stable. Changes to the golden image change what every other system is compared with. Treat it as change-controlled.

Create a comparison group

1. Go to the Comparison tab and click Add Comparison Group.
2. Enter a Name and Description, for example PowerScale – production standard.
3. Select the System type.
4. Select the Golden Image: the system all others are compared with.
5. Select the Assets or CI groups to compare. A CI group keeps the comparison current as systems are added.
6. Optionally, select a default Text Normalizer, or create one from the form. See below.
7. Save the group.

Use Edit and Delete on the Comparison tab to maintain groups.

Review the results

• Comparison details (the ... menu) gives a summary: which systems match the golden image and which differ, and by how much.
• Detailed View shows the full comparison, command by command or API call by API call, with differences highlighted. Select Show Differences Only to hide everything that matches, and switch between Normalized, Original, Split and Unified views.

Ignore expected differences with text normalizers

Some values differ on every system by design: host names, serial numbers, IP addresses, cluster IDs. Without help, every comparison flags them. A text normalizer masks that kind of value before the comparison, so only meaningful differences remain.

1. Open the Text Normalizers page.
2. Create a normalizer with a pattern for the value to mask, for example the site name or the IP address format.
3. Attach the normalizer to a comparison group as its default, or select it when you create the group.

Start with the differences that appear on every system in the detailed view; those are the candidates for normalization.

Related articles

• Track configuration changes
• Define your sites and CI groups (optional)
• Get evidence for an audit