Release notes / 2026.10.10

Early access release

Moorhold 2026.10.10 Replaced

Published

Release state

Replaced #

This was the Latest release until 2026.10.12 replaced it. Before it became Latest, we updated our own Moorhold server to it from 2026.10.9, and it passed a restore drill there. Its files stay on get.moorhold.dev and still work. We haven't tested an update from this release to 2026.11.4. If your server runs it, email hello@moorhold.dev and we'll help you move it.

We wrote this release's notes, including their Still to be tested line, before we ran it on our own server.

This release makes Trino's tables show in DataHub. 2026.10.9 brought that feature but each read stopped with a certificate error, so no tables showed. This release fixes the reads. 2026.10.9's notes listed the fault as still known. An admin can ask for more memory for an app from the portal, and the server applies the approved request by itself. A rollback no longer stops halfway when you've dropped a kept copy since the rollback point was taken, a restore from the wrong snapshot is refused before anything stops, an update counts the images it pulls, and running the installer again on a server that already runs this release says so plainly.

https://moorhold.dev/releases/2026.10.10

Before you update #

What changes
Trino's tables now show in DataHub. An admin can ask for more memory for an app from the portal, a rollback no longer stops halfway when you've dropped a kept copy, and a restore from the wrong snapshot is refused before anything stops.
Downtime
Moorhold is unavailable from when the update stops the running release until the new one has started. That was seen on this release's update from 2026.10.9 by the installer.
Backup
Yes: before it changes anything, the update takes a rollback point, a backup of every database on this server. If the new release doesn't start, it goes back to that point by itself.
Manual steps
None for the update. If the sign-in page afterwards shows Keycloak's own design, see Still known.
Still to be tested
Asking for more memory for an app is still to be tested end to end, as are the setup's first step explaining the two accounts, "Open the dashboard" going straight to the sample's dashboard, a failed restore cleaning up its copy, a failed pull's reason, the Trino reads coming back after trinoIngestion: false is removed, and the Updates page offering 2026.10.10 once LATEST moves.

Updating #

On a server running 2026.10.9, update from the server's command line with:

sudo moorhold upgrade 2026.10.10

It checks the release and the server before it changes anything, and takes the same rollback point as the installer. Once this release is the latest one, an admin can also press "Check for updates" on the Updates page in the portal and ask for the update there.

On a server running 2026.10.9, you can instead download this release's installer and run it on the server as you did to install, but without --hostname:

curl -fsSLo install.sh https://get.moorhold.dev/2026.10.10/install.sh
sudo bash install.sh

The installer sees the release that's running and updates it in place, with the same checks and the same rollback point. On a server that already runs this release, it says there's nothing to update and changes nothing.

New #

  • Ask for more memory for an app. When an app on your server is close to its memory limit, the portal's home page shows a card that says so. An admin can ask to give the app more memory from there, choosing the time it happens, in the server's own time zone. With two admins, another admin approves the request first. At the time chosen, the server applies the new size by itself, and if the app isn't healthy afterwards it puts the old size back. The portal asks again later, so a request can be left until you're ready.
  • The launchpad's tools, named and checked. Each tool on the portal's home page now has a health line of its own. The tools are named the way people say them, the page links to your account, and DataHub no longer shows its own product-update pop-over or fetches anything from DataHub's own project. Superset, DataHub and your account's tile can show Unknown: see Still known.
  • Which account to use. Moorhold's own sign-in and signed-out pages now say that the account you make on your server is that server's own account, separate from your Moorhold account at account.moorhold.dev, which is only for downloads and access keys. The setup's first step, where you set your admin password, says the same. On a server updated from an earlier release, the sign-in page that comes next may not say it yet: see Still known.
  • "Open the dashboard" goes straight to the sample's dashboard. At the end of the setup, the "Open the dashboard" button now opens the sample's own dashboard in Superset when the dashboard is there, signing you in on the way without a page in between. While it's still being made, the button says so and opens Superset's home page instead.

Fixed #

  • Trino's tables in DataHub. 2026.10.9's feature now works: each read of a catalog no longer stops with a certificate error, so your tables and views show in DataHub with their columns, on the schedule the DataHub component's settings say. If you stopped the reads with the workaround the last release's notes gave, setting trinoIngestion: false in the DataHub component's overrides in your Stack, remove that line and the reads start again on their next pass. 2026.10.9's notes listed this as still known.
  • A rollback after you've dropped a kept copy. sudo moorhold rollback no longer stops halfway when a database it would restore was dropped after the rollback point was taken, as sudo moorhold backup drop-kept does. The rollback leaves the dropped database out, says so before it changes anything, and finishes the rest. drop-kept now says what a later rollback to a point from before the drop does with it.
  • A restore from the wrong snapshot. sudo moorhold backup restore now refuses a snapshot that isn't a backup of the database you named, before anything stops, and says which database the snapshot is a backup of. Before, it stopped and restarted the stack first and then failed. A restore that still fails no longer leaves an empty <database>_r<time> copy behind: it drops the half-made copy when it's provable and safe to.
  • An update's image pull. An update now counts the images it pulls, one "Image N of M" line at a time, as an install does, instead of printing Docker's own list of addresses. When a pull fails, its reason keeps the registry's address readable. 2026.10.9's notes listed Docker's lines during an update as still known.
  • Running the installer again on this release. Running this release's installer without --hostname on a server that already runs it now says there's nothing to update and how to repair or move the server instead, and changes nothing. Before, it printed the installer's usage text as if you'd done something wrong. 2026.10.9's notes listed this as still known.

Verified #

  • Unit and golden tests: pass.
  • Images: every image passed the licence check, and each image is the one the manifest names.
  • On the test servers: the update from 2026.10.9 on the long-running test server, by the installer, ended on 2026.10.10, took its rollback point, kept the server's settings and data, and passed the restore drill afterwards. On a fresh 2026.10.9 install, the update by the installer and the update by sudo moorhold upgrade each ended on 2026.10.10 with the data marker kept, and sudo moorhold upgrade checked the release and the server before it changed anything. The update counted its images. Running the installer again said there was nothing to update and changed nothing. A rollback after a kept copy was dropped printed a "Left out" line for it and restored every other database. A restore from the wrong snapshot was refused before anything stopped, and left no copy behind. The update by the installer on the long-running test server stopped the running release until the new one had started, and Moorhold came back by itself.
  • In the portal: the launchpad's account link, Moorhold's own sign-in and signed-out pages explaining the two accounts, no product-update pop-over in DataHub and no fetch from DataHub's own project, and Trino's tables in DataHub after their first read since the update, with each Trino source's last run a success.

Images #

Every one of Moorhold's own images is rebuilt from this release's source. No tool changes version, and the upstream images keep the digests they had in 2026.10.9.

Still known #

  • A restore drill that never finishes. Sometimes, a restore drill and a data load, such as the sample data's, wait on each other for ever. That happens when the load starts to replace a table the drill is reading. If it's the drill at the end of an install or update, that install or update never finishes. Neither finishes by itself, so from a second terminal on the server, stop the load's waiting step:
    sudo docker compose -f /opt/moorhold/project/compose.yaml exec -T postgres psql -X -U postgres -d moorhold_warehouse -c "SELECT pg_cancel_backend(pid) FROM pg_stat_activity WHERE usename = 'moorhold_loader' AND wait_event_type = 'Lock'"

    The drill then finishes, and so does the install or update. The load stops with an error and keeps the data it had. It runs again at its next scheduled time, and you can start it sooner in Airflow. 2026.10.8's notes said a later release would fix it. It doesn't yet, and we're working on a fix. On a test server, a fresh install of this release with the sample data hung this way, and the command above cleared the wait.

  • The sign-in page after an update. After an update, the sign-in page can show Keycloak's own design instead of Moorhold's, without the line that explains which account to use. It still signs you in as before. To get Moorhold's page back, run this on the server:
    sudo docker compose -f /opt/moorhold/project/compose.yaml up -d --force-recreate keycloak

    Sign-in is unavailable until Keycloak has started again. If Keycloak's design comes back after you run sudo moorhold apply, run the command again.

  • Signing back in to DataHub. After you sign out of DataHub and sign in again, you land on the portal rather than the DataHub page you left. This release was meant to change that, and on a test server it didn't. Open DataHub again from the launchpad.
  • The launchpad's health lines. Superset's and DataHub's tiles can show Unknown, "No answer within 2 s", while both tools are working, and your account's tile says Moorhold has no health check for it yet. Trino's and Airflow's lines are right.
  • DataHub on a shared computer. If the last person to use the browser didn't sign out, DataHub can still open as them, with their access, until its session runs out, even after you sign in to Moorhold as yourself. Sign out before someone else uses the browser.
  • Users and Groups in DataHub. If your role can't manage users, DataHub's Users and Groups page shows "Failed to load users!". That's DataHub's message when your role isn't allowed to see the list of users.
  • Working only in Airflow or Superset. Time spent in Airflow or Superset alone doesn't keep your Moorhold session going, so if you use only one of them until the session times out, you see the sign-in page.
  • API tokens made in Airflow. They keep working after you sign out, until they expire or are deleted in Airflow.
  • System tables for agents. Agents that search through DataHub's own MCP server still see the system tables, because DataHub applies its default view only on its own pages.
  • A reboot after a rollback. 2026.10.7 and earlier releases can't stop Moorhold at shutdown. So after a rollback to one of them, a reboot shows services failing and restarting until Moorhold has started, as before. They recover by themselves.
  • Airflow's Triggerer. Airflow shows its Triggerer as unhealthy because Moorhold doesn't run one. Nothing in Moorhold uses deferrable tasks, and Moorhold's own health check ignores it.
  • x86_64 (amd64) servers only.

Upgrading #

On a server running 2026.10.7 or later, update from the server's command line with sudo moorhold upgrade <version>, giving the release you're moving to. On a server running 2026.10.6 or later, you can instead download the new release's installer and run it on the server as you did to install, but without --hostname.

This release's notes give its steps under Updating, and say which releases it updates from. If they don't cover the release your server runs, email hello@moorhold.dev and we'll help. More about upgrading

Recovery steps #

There's no automatic rollback yet, so there are no recovery steps to follow on your own. If something has gone wrong with your Moorhold, email hello@moorhold.dev with the version you're on and what you saw, and we'll help. Please don't send passwords or keys.

Download and verify #

The signed release is on get.moorhold.dev. Check the signature on release.lock first, then each file's SHA-256.

FileSHA-256
moorhold-2026.10.10.tar.gz7d24f60d1a47c50f5d09174e378fc2616ec7b4e6fa1417e2d44873f2edf16f2a
manifest.yaml6b3a5db6f414045518de9f9821f108f321d704c956d145ee424be3c0d795664e
release.lock2507e8265386ab28482f826c39cb0cd9f35db3fa544791126cc2e5f994b8033d
install.sh82cb2db404d047605006232037ab064abe974587acca74ca98d3ac61ee95e8fe
cloud-init-ionos.yaml1835c9c974b68c9a2ba66a1197deedb7428f9404a76b5f96d0ceb63cf3cacf0d
cloud-init-hetzner.yaml4e0a7b461771483e742f09ff6ee60c1e24c32980ed262f46d585f85999037891
release.lock.minisigSignature, key 7F6D12F3F6217F9C
SHA256SUMSThe checksums above
minisign -Vm release.lock -p moorhold-release.pub
sha256sum -c SHA256SUMS

Public key: get.moorhold.dev/moorhold-release.pub