immich-app/immich
1
0
Fork 0
mirror of https://github.com/immich-app/immich.git synced 2026-03-01 11:20:12 +03:00
Code Issues 783 Packages Projects Releases 12 Wiki Activity

[PR #25876] fix: development containers init race conditions #18379

New Issue
Open
opened 2026-02-05 16:37:20 +03:00 by OVERLORD · 0 comments
OVERLORD commented 2026-02-05 16:37:20 +03:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/immich-app/immich/pull/25876
Author: @midzelis
Created: 2/4/2026
Status: 🔄 Open

Base: main ← Head: push-olwpzvrxnomt

📝 Commits (1)

  • a6cffcc fix: development containers init race conditions

📊 Changes

7 files changed (+134 additions, -24 deletions)

View changed files

📝 cli/package.json (+1 -1)
📝 docker/docker-compose.dev.yml (+55 -3)
📝 e2e/docker-compose.dev.yml (+61 -3)
📝 e2e/package.json (+3 -3)
📝 pnpm-lock.yaml (+5 -5)
📝 server/Dockerfile.dev (+8 -8)
📝 web/package.json (+1 -1)

📄 Description

pnpm fetch was useless in the development containers - it wasn't even populating the correct location

but, pnpm advises that stores should not be shared across multiple containers due to file consistency issues. technically, it would probably be ok as long as only a single pnpm process was active - but with multiple containers that was not true.

instead, share the tarball cache (pnpm-cache) so that downloads can be shared pnpm-store will be unique/isolated per container.

to give a change for pnpm hardlinks to work, the store should be a volume too to match the node_modules all being volumes.

however, the node_modules are volumes that are shared between the containers, which also had race conditions when both containers were starting up linking/copying packages from the store to node_modules. instead of using isolated node_modules between containers (which would use up a lot more space) - this can be solved by ensuring that all the package installs are performed before either container starts in a single container (preventing races)

so, a tiny init-container is added that just runs pnpm install - the server/web containers depend on it. if the node_modules are already there, it exits quickly.

see also #25398

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/immich-app/immich/pull/25876 **Author:** [@midzelis](https://github.com/midzelis) **Created:** 2/4/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `push-olwpzvrxnomt` --- ### 📝 Commits (1) - [`a6cffcc`](https://github.com/immich-app/immich/commit/a6cffcc3889a87fba8737e5ba610462c54f17d9a) fix: development containers init race conditions ### 📊 Changes **7 files changed** (+134 additions, -24 deletions) <details> <summary>View changed files</summary> 📝 `cli/package.json` (+1 -1) 📝 `docker/docker-compose.dev.yml` (+55 -3) 📝 `e2e/docker-compose.dev.yml` (+61 -3) 📝 `e2e/package.json` (+3 -3) 📝 `pnpm-lock.yaml` (+5 -5) 📝 `server/Dockerfile.dev` (+8 -8) 📝 `web/package.json` (+1 -1) </details> ### 📄 Description pnpm fetch was useless in the development containers - it wasn't even populating the correct location but, pnpm advises that stores should not be shared across multiple containers due to file consistency issues. technically, it would probably be ok as long as only a single pnpm process was active - but with multiple containers that was not true. instead, share the tarball cache (pnpm-cache) so that downloads can be shared pnpm-store will be unique/isolated per container. to give a change for pnpm hardlinks to work, the store should be a volume too to match the node_modules all being volumes. however, the node_modules are volumes that are **shared** between the containers, which also had race conditions when both containers were starting up linking/copying packages from the store to node_modules. instead of using isolated node_modules between containers (which would use up a lot more space) - this can be solved by ensuring that all the package installs are performed before either container starts in a single container (preventing races) so, a tiny init-container is added that just runs `pnpm install` - the server/web containers depend on it. if the node_modules are already there, it exits quickly. see also #25398 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
OVERLORD added the pull-request label 2026-02-05 16:37:20 +03:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
accessibility
changelog:enhancement
changelog:security
changelog:skip
changelog:translation
cli
date-time
dependencies
documentation
external-library
format
good first issue
mobile-beta
mobile-beta
mobile-beta
needs-answer
nice to have
pull-request
Mirrored from GitHub Pull Request
 sharing
tech-debt
📱mobile
🖥️web
🗄️server
🧠machine-learning
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
OVERLORD
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: immich-app/immich#18379
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?