Bug: SQLITE_BUSY database locked error when updating site or child documents — possibly related to concurrent node activity (related to error #402)

[dbokpaw]

When updating a site (or a child document), I encounter a SQLITE_BUSY: database is locked error. The update fails with the following error message:

"failed to publish. Failed to begin immediate transaction: original error: sqlite.Exec: Stmt.Step: SQLITE_BUSY: database is locked (BEGIN IMMEDIATE TRANSACTIONS;)"

On a separate occasion around the same time, I also received:

"failed to publish: HTTP 500 from PublishBlobs: Internal Server Error"

This has occurred both when updating the main site and when updating child documents. I have not yet found a reliable way to reproduce it consistently, but support previously mentioned that this often happens when something else is happening on the node at the same time.

Steps to reproduce (sporadic):

1. Open a site or child document in the desktop app.
2. Make changes and attempt to publish/update.
3. Occasionally, the update fails with the SQLITE_BUSY error or HTTP 500 error.
4. The issue seems unpredictable and may occur when other background activity is taking place.

Expected behavior:
The update should complete successfully without database locking or server errors.

Actual behavior:
Updates intermittently fail with SQLITE_BUSY: database is locked or HTTP 500 from PublishBlobs: Internal Server Error, even when no obvious concurrent actions are being performed.

Environment:
App version: 2026.3.4
OS: Windows 11
Device: Desktop

[juligasa]

Some fixes were introduced in 2026.3.5. After using the latest versions, have you found any error of this type? @dbokpaw ?

[dbokpaw]

I haven't encountered any errors like that in the latest version. It works perfectly now.

[juligasa]

Great, i also pushed some changes to speed up syncing here b0b508a If this comes out again, feel free to reopen