A staged upload always won over a server path, so an old upload that
was never imported could be restored instead of the chosen file.
- Record the selected source (checked server file or finished upload)
and import only from it
- Delete a stale staged upload when another source is used
An all-databases MySQL/MariaDB restore also restored the mysql system
database, so after the next restart the root and app passwords were the
source ones and the MariaDB health check failed. The restore now skips mysql,
sys, performance_schema and information_schema by default. The new option
"Restore users and privileges" (restore_mysql_users in the API) keeps the
old behaviour.
Before Coolify kept `driver`, `driver_opts` and `labels` of renamed
Compose volumes, Docker created those volumes without them. Applying the
options now would make Docker Compose ask to recreate the volumes on
every deployment.
- Add `ignores_compose_driver_options` column to local_persistent_volumes
and mark all volumes that exist at upgrade time
- Parsers now use composeRenamedVolumeDeclarationFor() so flagged volumes
keep the old name-only top-level declaration
- Legacy application parsers (v1/v2) always emit name-only declarations
- Reset the flag when a volume is replicated, since a clone is a new
Docker volume
- Include the flag in server transfer export/import; bundles without it
default to true
- Storages UI shows which volumes ignore driver options and allows
deleting them so the options apply on recreation
- Add tests for parsers, cloning and server transfer
- database import: add keep_owners (restore PostgreSQL owners and
privileges) and sqlite_database (pick the target SQLite file) to the
API, OpenAPI spec, and import form
- oauth: require password confirmation and support team selection on
OAuth login
- remote process: send team broadcast events to the team of the server
that ran the process instead of the causer's user id
- servers API: handle the legacy build server setting
- compose: support parent bind paths and volume driver options
- dns: improve record hints and provider management
- notifications: fix throttling and missing backup notifications
- ui: rework the confirmation modal
- add feature and unit tests for the new behavior
- Detect the backup format before any database is changed. "All databases"
imports no longer drop everything when the file cannot be restored.
- PostgreSQL: restore custom and tar archives with pg_restore and SQL with
psql in both modes, including gzip; restore the #11481 custom-archive fix.
Replacing existing data recreates the database for SQL backups.
- MySQL/MariaDB: accept a tar with one dump; reject dumps with more than one
database in single mode instead of restoring them partially.
- MongoDB: restore plain and gzip archives and dump directories packed as
tar; "Replace collections" maps to --drop.
- Prepare bz2, xz, and zip backups in the helper image, because database
images do not ship those tools; stream S3 backups from the S3 helper.
- Show the exact restore script in the import form.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Move database import temp/container cleanup into a queued listener with retries, and clear the import running state on failed restore starts. Tighten Cloudflare replace/delete to the current server IP and owning resource, notify when DNS jobs fail, and pass domain-removal confirmation through remove-by-key. Document import request oneOf schemas without extra properties.
PostgreSQL single-database restores no longer always replace existing
objects. Update the file and S3 confirmation dialogs so they match the
callout instead of claiming all data will be replaced.
Allow single-database PostgreSQL imports to drop matching objects before restore. The API and import form accept replace_existing, which adds --clean --if-exists to pg_restore, and pg_restore now uses --exit-on-error.
Design tokens
- Fix WCAG AA text failures: darken light muted text and lighten dark
tertiary text so 13-14px copy clears 4.5:1 on every surface.
- Rebuild the dark surface ladder in correct values (oklch compresses to
near-black below ~15%), giving visible steps: content, chrome, cards.
- Establish a 3-layer shell (deep content, lighter sidebar/topbar chrome,
lifted cards) with crisp hairline rings and a restrained card shadow.
- Unify every surface to one neutral temperature (pure gray); neutralize
the cool-tinted sidebar text and legacy surface tokens.
- Convert all color tokens to oklch (neutrals + brand).
- Unify inputs to the recessed token; fix invisible dark placeholders.
- One shared --shadow-dropdown for all menus/listboxes/command palette.
Settings sidebar
- Sticky contained-card rail with subtle scrollbar; collapsible sub-section
groups (active open by default), cross-page section links.
- Mobile: collapsible disclosure with animated open, outside/Escape dismiss,
press feedback, and reduced-motion support.
Mobile
- Consolidate the resource header (title + status + links on one row).
- Tighten settings-page vertical rhythm.
- Rebuild the main sidebar as an animated shadcn-style sheet (slide/fade,
Escape, scroll lock, in-panel close, close-on-navigate).
Fixes
- #11532: selected server invisible in the light-mode terminal toolbar.
Docs: update DESIGN.md tokens, shell layering, and temperature rules.
Gate public TCP proxy until a port is set, show loading while saving
access, and open proxy logs in process-dialog. Compact backup
execution actions, add Postgres general section anchors, and tint
resource Actions icons with warning color.
Extract the database import form into its own component and add realtime
status refresh components for application server badges and service resource
cards.
Andras BacsaiandClaude Opus 5.5
peaklabs-dev
Aditya Tripathi
