ALLMSP Blog

Backblaze Business Backup Recovery Testing: Retention, Drives, and Large Restores

Backblaze Computer Backup can report a healthy endpoint while a business-critical recovery gap remains. Follow staged checks and verify a durable recovery.

Backblaze Restore App example showing a backup-date control, 42,616 files selected and no restore destination chosen

Backblaze’s published Restore App example shows the backup-date control and 42,616 selected files totaling 8.19 GB, with the destination still set to No selection. The red callouts and blurred fields are present in the vendor file. This older app example demonstrates pre-restore choices, not a completed recovery or the current standalone app layout.[9] View full-size image

A backed-up file list is not a recovery result. First identify the required computer, drive, file and retained date; then choose a restore method, an authorized operator and a safe destination. Finish by opening the recovered sample and checking it against the expected version. A healthy-looking laptop cannot prove that a disconnected external drive or an excluded share is recoverable.

The selected Version History matters before a drive is lost or unplugged. Backblaze’s specific external-drive guidance distinguishes 30-day and one-year behavior, while the Restore App, web ZIP, Save Files to B2 and shipped-drive paths have different limits and handling. The hypothetical table below shows why changing history today cannot repair an already-cleared version.

A monthly restore sample turns those settings into evidence. The test should cover a recent local file, a prior version, a deleted file within the available window, an external-drive file, a large file, and one expected exclusion. A separate annual or semiannual exercise should rehearse a larger recovery using the chosen application, web ZIP, Save Files to B2, or shipped USB path, with enough destination and temporary space and with any required private encryption passphrase available under controlled custody.

Businesses can use ALLMSP for Backblaze Business Backup backup and recovery validation for organizations in Lawrenceville and Suwanee, throughout Gwinnett County and Metro Atlanta, and across Georgia, with local assistance available when device, network, office, or employee work needs an on-site component.

Key decisions at a glance

  • Validate what Computer Backup can actually protect because operating-system files, applications, network shares, NAS devices, Time Machine data, and other exclusions can create silent recovery gaps.
  • Choose thirty-day, one-year, or forever Version History against business recovery requirements and understand that an increase cannot resurrect versions already removed.
  • Track external-drive attachment and backup state independently, and never delete the source merely because files appear in the Backblaze web console.
  • Restore representative local and external-drive files every month, verify contents and metadata, and preserve evidence that names the source, restore point, method, result, and reviewer.
  • Select web ZIP, Save Files to B2, Restore App, or USB drive paths by size, time, network, temporary space, security, and operational constraints before a major incident.

Prove that required files are eligible and present in the backup

Create a recovery coverage matrix from business process to endpoint path. Backblaze automatically selects supported local data across user profiles and can back up internal drives and selected connected USB or FireWire drives. It does not simply capture an image of the whole computer. Operating-system and application files, temporary data, empty folders, caches, several virtual-drive formats, network shares, NAS devices, remotely mounted volumes, Time Machine data, aliases, shortcuts, symlinks, and other documented exclusions need separate handling.

Test files that are easy to miss. Backblaze cannot back up a file while another program has it locked, so validate locally stored email databases, accounting files, creative project packages, and other long-running application data after an appropriate close or supported quiescing procedure. On Mac, FileVault-protected files can be backed up while the computer is unlocked. Record the source application, path, expected close behavior, last changed time, Backblaze evidence, and alternate method for any file that cannot be made eligible.

Keep network and server recovery evidence separate. A NAS share can be protected through a Backblaze B2 integration, but that bucket and integration do not appear as an ordinary Computer Backup endpoint. Its application key, job schedule, retention, alerts, immutability, and restore test belong to the B2 control set. In the Computer Backup report, identify the NAS or server as an excluded workload with a linked alternate-control owner rather than counting it as green because a nearby workstation is backed up.

Use the supported-data reference for exclusions and locked-file behavior.[1] A NAS or server protected through B2 needs its own supported integration and restore evidence, outside the Computer Backup endpoint result.[8]

Align Version History and external-drive behavior with recovery objectives

Check both the purchased entitlement and the actual Version History setting. The current Business product page advertises one-year history included, while the technical Version History reference still documents a 30-day default and a selectable one-year option at no additional storage charge. Do not assume the entitlement changed an existing backup’s setting or retained dates. Forever retains older versions while enabled and the account remains in good standing, with applicable charges for modified or deleted data older than one year. Record the selected setting, effective date, accountable owner and required recovery interval.

Increasing history protects data still retained when the change is enabled, including versions within the prior 30-day window; it cannot resurrect versions already cleared. Decreasing history immediately prevents recovery beyond the shorter setting. Require approval and complete needed restores before a reduction. For example, enabling one-year history today is not evidence that an already-removed file from 70 days ago exists. Verify the actual retained date rather than calculating protection from the new label alone.

Use the specific External Hard Drives guidance for the recorded history. Under 30-day history, the documented missing-drive alerts occur after 14, 21 and 28 days, and data may be removed after more than 30 days disconnected. Under one-year history, retained drive data can remain for up to one year, with alerts at 30, 345 and 360 days; reconnecting within that retained year recognizes and deduplicates previously backed-up files rather than requiring the whole drive to upload again. Keep drives connected frequently for new backups, verify selection after a rename and test a recent sample. The older general best-practices page’s blanket 30-day paragraph must not override this newer, setting-specific guidance.

The Version History reference explains the documented default and non-retroactive recovery boundary.[3] The current Business product page includes one-year history; verify the actual selected setting rather than inferring it from entitlement.[12] The newer external-drive reference supplies the separate retention, alert and deduplication behavior for each setting.[4]

Worked recovery decision: an external drive last connected 70 days ago

These are hypothetical decisions based on recorded settings, not a claim that any particular customer's files remain available. Inspect the actual backup before acting.

Recorded conditionWhat it does and does not establishNext safe check
30-day history; drive absent 70 daysData may already be outside retention; enabling a year now cannot resurrect cleared versionsInspect retained dates immediately and prepare any available required restore; escalate an actual gap
One-year history was already enabled while the files were retainedDisconnected-drive data can remain for the selected year; a current empty view alone is not proof it is goneSelect the appropriate historical date and verify a restored sample
The one-year drive reconnects within its retained yearCurrent guidance says existing data is recognized and deduplicated; a whole-drive reupload is not required just because 30 days passedConfirm drive selection, new-file progress and a recent sample; do not erase the source
A 620 GB recovery is neededIt exceeds one web ZIP request; the native app is a different path, whose current support and capacity must be checkedEvaluate the supported app or split web requests; verify destination and temporary space, integrity and application readability before promising a deadline

Run monthly sample restores and preserve useful evidence

Build a rotating sample that represents the business instead of repeatedly restoring one tiny document. Include a current local file, an earlier version, a deleted item that should still be in history, a selected external-drive file, a large item, a path with meaningful permissions or metadata, and one excluded record expected to be absent. Use synthetic or low-sensitivity samples where possible, but keep file types, sizes, applications, and drive relationships realistic enough to expose operational failures.

Start the test from the same authority and device path expected during an incident. A managed administrator may restore for a member only when group and Enterprise Control settings permit it, a decentralized user may use their own account, a centralized population may depend on IT. If a private encryption passphrase is enabled, retrieve it under the approved custody process. Select the correct computer and restore point, deliver the result to an isolated destination, scan it, and compare hash, size, content, timestamps, and application readability with the expected sample.

Retain evidence without copying sensitive content into the ticket. Record test date, requester, approver, source endpoint, drive class, selected restore point, version-history setting, method, file types and sizes, start and finish time, destination, passphrase-custody participants, integrity result, metadata result, application-open result, exceptions, remediation owner, and independent reviewer. Delete test copies according to policy. Trend failures by cause so recurring locked files, disconnected drives, stale clients, or insufficient space become managed problems.

The best-practices guide recommends monthly local and external-file restore samples.[2] If a passphrase protects the backup, exercise the approved custody path and follow the provider’s current supplied-passphrase handling rather than promising recovery of a lost secret.[7]

Choose and rehearse the large-restore path before the recovery clock starts

Match the method to the retained data, volume and deadline. Current Mac and Windows release notes describe a standalone Restore App that operates independently of the backup client; the older Create Restore manual introduces the earlier built-in app. Use the current supported app for the platform and check the controls actually present. The Backblaze Downloader instead retrieves an already-prepared web ZIP. The web guide caps each ZIP at 500 GB and at most 20 active ZIP downloads, Save Files to B2 at 10 TB and a shipped USB hard drive at 8 TB. Those are web-method limits; do not treat a web ZIP as the only direct-download recovery path. Check availability, cost, authority and capacity for the actual path before committing to an incident deadline.

For a validation test, choose a controlled custom folder so the original location is not overwritten inadvertently. The Restore App needs slightly more free space than the selected restore and may also need a suitable temporary data drive. Check destination permissions, free space, file system, security scanning, network and sleep behavior without broadly weakening folder permissions to hide an error. Same-location skip and overwrite options have different consequences; record the chosen setting. For shipped media, separately arrange authorized receipt, custody, unlocking, return timing and disposal of test copies.

Run a timed large-restore exercise with a representative data set and a written decision tree. Measure request approval, restore preparation, download or shipment, passphrase retrieval, transfer, integrity check, malware scan, permissions, application validation, and user handoff. Compare the result with the recovery-time objective and identify which steps dominate. Keep a documented alternate path in case internet capacity, temporary space, B2 limits, shipment timing, or authentication fails during the real event.

Check web restore limits and preparation using the web-console procedure.[5] The Restore App procedure identifies date, destination and conflict choices,[9] while current Mac[10] and Windows[11] release notes establish the standalone app and its platform-specific fixes. Use the temporary-drive reference for its separate working-space setting.[6]

If the required source is not in the expected account or selection, return to the Backblaze deployment and coverage review. If the operator lacks approved account or key access, use the Backblaze administration and recovery-authority review before broadening access.

Frequently Asked Questions

Does Backblaze Computer Backup create a complete system image?

No. Computer Backup automatically selects supported local data, but it excludes operating-system files, applications, caches, temporary data, network shares, NAS devices, Time Machine data, and other documented items. Build a coverage matrix and use separate methods for workloads that require system, application-consistent, server, or NAS recovery.

Can Backblaze back up a file while an application has it locked?

No. Backblaze says files in use by another program cannot be backed up at that moment. Test locally stored email, accounting databases, creative projects, and similar data with the application safely closed or quiesced, then verify the file’s backup and restore result rather than assuming later capture.

What Version History does Backblaze Computer Backup use by default?

The current Business product page includes one-year history, while the technical Version History reference still documents a 30-day default and a selectable free one-year option. Verify the actual backup setting and when it took effect. A purchased entitlement or newly selected longer history does not prove that versions already removed by the service are recoverable. Forever has charges for older modified or deleted data.

Does increasing Backblaze Version History restore versions already deleted from the service?

No. Increasing history preserves versions still retained when the change is enabled, including those within the previous 30-day window; it does not resurrect versions already cleared. Check the actual retained date and prepare any required available restore promptly. Before reducing history, complete needed recoveries because the shorter setting immediately limits what can be recovered.

What happens if an external drive stays disconnected from Backblaze?

It depends on the recorded Version History. Current guidance gives 30-day history alerts at 14, 21 and 28 days, with possible removal after more than 30 days absent. One-year history retains disconnected-drive data for up to a year and alerts at 30, 345 and 360 days. Reconnecting within that retained year deduplicates existing data without a whole-drive reupload; verify new files and the actual restore date.

Can a business delete source files after Backblaze uploads them?

No. Backblaze describes Computer Backup as a live mirror rather than archival storage. When source data is deleted or a drive disappears, retention follows the configured Version History and device behavior. Preserve the source and maintain a separate local backup instead of treating the online copy as a storage migration.

How often should a business test Backblaze restores?

Backblaze recommends signing in monthly and restoring a few files from local and external drives. Use a rotating sample of current, prior-version, deleted, external, large, metadata-sensitive, and expected-exclusion cases, then verify integrity and application readability while recording scope, timing, authority, result, and remediation.

How large can a free Backblaze web ZIP restore be?

The current web-console guide limits each ZIP restore to 500 GB and allows no more than 20 active ZIP downloads. The Downloader retrieves a ZIP that was already prepared; it is not the same tool as the native Restore App. Choose the appropriate path and test destination capacity, security and elapsed time instead of applying the ZIP cap to every restore method.

What options does Backblaze offer for a larger restore?

Use the native Restore App for a suitable direct-to-destination recovery, or evaluate a web-created B2 restore up to 10 TB or a shipped USB drive up to 8 TB. The native app needs adequate destination and temporary space; its same-location overwrite option can affect existing files. For a test, prefer a controlled custom folder and verify the restored version and application readability.

What evidence should a Backblaze recovery test retain?

Retain the approved requester, source computer and drive class, restore point, retention setting, method, file types and sizes, timing, destination, custody participants, integrity and metadata checks, malware scan, application-open result, exceptions, remediation owner, and reviewer. Do not store private passphrases or sensitive recovered contents in the ticket.

Facebook
LinkedIn
WhatsApp
X
Email
Print
Threads
Reddit

Latest Articles