Bulk migration stalled
Use this page when the Bulk Offload progress on the Dashboard stops moving or never starts. It explains how the background run works, how to find the cause on the Health page, and how to finish with WP-CLI.
How the bulk offload runs#
When you click Offload All Media, WP Media Cloud runs the work in the background, not in your browser.
- Your site sends a request to itself at
wp-admin/admin-ajax.php. This is called a loopback request. - That request uploads files for up to 20 seconds, or until PHP has used 90% of its memory limit. It then saves its place and sends the next loopback request.
- A WP-Cron event runs every 5 minutes as a backup. If nothing is running but work is left, it uploads a batch itself, then tries the loopback request again.
The Dashboard checks progress every few seconds while the run is active. Newest media is offloaded first.
If your server cannot reach itself, the WP-Cron event still uploads a batch every 5 minutes, so the run continues more slowly. This needs WP Media Cloud 1.1.9 or newer and a working WP-Cron. For large libraries, WP-CLI is much faster.
Step 1: Run the Health checks#

Go to Media Cloud > Health and click Run All Checks. Three rows matter for the bulk offload.
| Row | What you see | What it means | How to fix it |
|---|---|---|---|
| Loopback Request | “Loopback requests working correctly.” | Your site can call itself. | Nothing to do. |
| Loopback Request | “Loopback may be restricted. Bulk offload falls back to WP-Cron.” | WordPress could not complete its own loopback test. The run continues through WP-Cron, more slowly. | See Step 2 to make it faster, and Step 3 to make sure WP-Cron runs. |
| WP-Cron | “Cron active. Next run in …” | A run is active and the backup event is scheduled. | Nothing to do. |
| WP-Cron | “Bulk offload queued but cron healthcheck not scheduled.” | Work is waiting, but the backup event is missing. | Click Stop Offload, then Offload All Media again. |
| WP-Cron | “No bulk offload running, so cron is idle as expected.” | No run is active. | Start the run from the Dashboard. |
| Upload Test | “Plugin not ready: storage credentials are incomplete. Check Settings or run the Setup Wizard.” | A required storage setting is empty, so nothing can upload. | See New uploads not being offloaded. |
Step 2: Fix a blocked loopback request#
These are the usual reasons a site cannot call itself.
- Password protection on staging. HTTP basic authentication blocks the request. Allow requests from the server’s own IP address, or use WP-CLI.
- A security plugin or firewall. Some block requests to
admin-ajax.phpor requests from the server’s own IP. Allow them, or turn the rule off while the run completes. - The host blocks self-requests. Ask your host to allow the server to reach its own domain over HTTP and HTTPS.
- DNS points elsewhere. If the domain resolves to a different server from inside the host, the request goes to the wrong place. Ask your host to check this.
After a fix, run the Health checks again and click Offload All Media or Continue Offload.
Step 3: Check WP-Cron#
The WP-Cron backup only works when WP-Cron runs. When loopback requests are blocked, it is what moves the run forward.
- If
wp-config.phphasdefine( 'DISABLE_WP_CRON', true );, WordPress does not run cron on page loads. Make sure a real server cron callswp-cron.phpat least every 5 minutes, or remove the line. - On a site with very little traffic, WP-Cron only runs when someone visits. A server cron job fixes this.
Step 4: Check the PHP memory limit#
WP Media Cloud stops each batch at 90% of the PHP memory_limit. Very low limits make batches very short.
If your host sets memory_limit to -1 (unlimited), use WP Media Cloud 1.1.6 or newer. Older versions read -1 as a limit of minus one byte and stopped before every file, so the run never moved.
You can see the current limit in Tools > Site Health > Info > Server.
Step 5: Restart once after updating to 1.1.6 (multisite)#
On multisite, version 1.1.6 moved each subsite’s queue into that subsite. A run that was in progress during the update is cleared.
Start the run again once on each affected subsite. Anything already offloaded is kept and is not uploaded again.
Step 6: Look at failed items#
A run does not stop because one file fails. Failed files are recorded with the reason and the run moves on.
- On the Dashboard, the Bulk Offload card shows “N items failed to offload”. Click it to see each file and its error.
- Fix the cause, then click Retry all. The result shows how many were offloaded and how many still fail.
- Clear empties the list. The files stay on your server and the next bulk offload tries them again.
You can also run wp wpmc failed to list them, and wp wpmc offload --retry-failed to retry them.
| Error | What it means | How to fix it |
|---|---|---|
| “Missing file: /path/to/file.jpg” | The file is in the Media Library but not on disk. | Put the file back in that folder, or delete the attachment. |
| “HTTP 403 AccessDenied: Access Denied” | The keys cannot write to this bucket. | Give the key write access to the bucket. |
| “HTTP 404 NoSuchBucket: …” | The bucket name is wrong. | Run the setup wizard and check the Bucket name. |
| “No response from storage provider” or “cURL error 28: …” | The request timed out or was blocked. | Ask your host to allow outbound HTTPS to the provider. Then retry. |
| “Bunny Storage HTTP 401: …” | The Bunny storage zone password is wrong. | Run the setup wizard and paste the Zone Password again. |
For more error strings, see Connection test failing.
Resume or restart a run#

- Click Continue Offload to pick up again. It only uploads attachments that are not offloaded yet.
- Click Stop Offload to cancel. Files already offloaded stay offloaded.
- Large files such as videos can take a while. Wait a few minutes before you decide a run has stalled.
Use WP-CLI instead#
WP-CLI runs on the server, so it does not need loopback requests or WP-Cron. It is the fastest option for large libraries.
- On the Dashboard, click Stop Offload so the two runs do not overlap.
- Check the current state with
wp wpmc status. - Start the offload with
wp wpmc offload. Run it again at any time and it continues where it stopped.
Useful options:
| Command | What it does |
|---|---|
wp wpmc offload --limit=20 | Offloads only 20 attachments, as a quick test. |
wp wpmc offload --batch=100 | Fetches 100 attachments per batch instead of 50. |
wp wpmc offload --retry-failed | Retries only the items that failed before. |
wp wpmc offload --network | Runs on every site in a multisite network. |
nohup wp wpmc offload --network > wpmc-offload.log 2>&1 & | Keeps running after you close your SSH session. |
The WP-CLI tab in Media Cloud lists every command with a copy button.