Stack guide
Back Up Multisite on Compute Engine to Cloud Storage
How this hosting and storage combination fits together with Multisite Migrate: setup, limits to plan around, and a restore test before you rely on it.
Short answer
This stack guide covers Back Up Multisite on Compute Engine to Cloud Storage as a concrete Multisite operations path. Combine hosting-class limits (PHP time/memory, cron, disk) with Multisite Migrate’s Jobs in Chunks and the correct destination: native Plus targets are Google Drive, Dropbox, S3-compatible storage, and SFTP.
Wasabi, Backblaze B2, Cloudflare R2, DigitalOcean Spaces, and MinIO are used through S3-compatible endpoints — not as separate proprietary vaults. Free keeps local schedules; Plus adds cloud schedules and upload; Pro adds remote import, empty-server installer, Staging, and network split/extract. For migrations between hosts, take a verified archive first, move with the matching restore path, then confirm URLs, media, and users on Staging before DNS cutover. Tier focus: Plus. Keep the previous host reachable until the new Staging drill passes twice, and document which Plus destination holds the rollback object so on-call staff are not hunting buckets during an outage.
Host constraints and chunked jobs
Prepare PHP limits, real cron where wp-cron is delayed, and free disk for Multisite-sized archives. Jobs in Chunks resume after interrupts so host timeouts do not force a full restart.
Exclude cache/tmp noise. Confirm Network Admin vs subsite scope before the first production schedule.
Destination truth for this stack
Connect Plus destinations you actually have: Google Drive, Dropbox, S3-compatible, or SFTP. Test the connection with a small job before relying on overnight schedules. OneDrive is not native. S3-compatible services share endpoint/credential patterns — label them clearly in ops docs.
Retention: keep enough generations for rollback. A stack named “Back Up Multisite on Compute Engine to Cloud Storage” still fails if you only keep one corrupt overnight object.
Migration and cutover
Archive → transfer → restore → verify before DNS cutover. Use import, remote import, or empty-server flows as your tier allows. Staging is the place to catch HTTPS, mapping, and media issues — not the live cutover window.
If moving between hosts, keep the old archive until the new Staging drill passes twice.
Steps
- Prepare the host side
Confirm PHP limits, cron, and free disk for Multisite archives before the first full-network job. - Connect the destination
Drive/Dropbox/S3-compatible/SFTP on Plus; test the connection with a small upload. - Run schedule or one-shot
Use Jobs in Chunks; retain enough generations for rollback; avoid overlapping locks. - Restore drill
Staging import/restore; check HTTPS, mapping, and Woo/users if relevant. - Cutover only after verify
Change DNS or primary hostnames only after Staging checks pass and a rollback archive remains available.
Frequently asked questions
Is OneDrive supported?
No native OneDrive destination. Use Google Drive, Dropbox, S3-compatible, or SFTP.
Can I migrate between hosts with this stack?
Yes — verified Multisite archive, transfer, restore path matching your tier, then DNS cutover.
Do I still need host snapshots?
They are a useful extra layer, not a replacement for Multisite-aware plugin archives.
How do B2/R2/Wasabi fit?
They are not separate Multisite Migrate vaults. Configure S3-compatible credentials and test upload.
Sources
Related guides and comparisons
Make your first restorable network backup for free.