Der Wert 2048 aus v1.119.3 stammte aus der Zeit, als der Container 3 GB
hatte. Er war zu niedrig gewaehlt: Der Build starb punktgenau bei 2045 MB,
also an meiner eigenen Grenze. Gemessener Bedarf sind rund 2,6 GB, deshalb
jetzt 4096. Der Container hat inzwischen 6 GB und damit genug Luft.
v1.119.4
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Der Deploy starb mit "Ineffective mark-compacts near heap limit". Nodes
Standard-Heap richtet sich nach dem Arbeitsspeicher und liegt auf dem
Container bei rund 1,5 GB -- zu wenig fuer next build. update-app.sh setzt
jetzt --max-old-space-size=2048.
Ich hatte zuerst ESLint im Build verdaechtigt (mein ignoreDuringBuilds:
false aus v1.117.4) und die Aenderung zurueckgenommen. Eine Messung
widerlegte das: Spitzenspeicher mit ESLint 2.590 MB, ohne 2.595 MB -- kein
Unterschied. Die Ruecknahme ist damit hinfaellig, ESLint bleibt im Build.
Der Build braucht rund 2,6 GB. Bei 3 GB Container-RAM bleibt das knapp;
NODE_OPTIONS ist die Sofortmassnahme, mehr RAM fuer LXC 202 die eigentliche
Loesung. Steht als Block J in TODO.md, zusammen mit der Alternative, den
Build auf den Spark zu verlagern.
v1.119.3
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
There were two near-identical update scripts and they drifted: the npm ci fix
from 1.114.1 only ever landed in update-app.sh, so the boot path kept running
npm install. my-weekly-todo-update.service now runs update-app.sh, and
boot-update.sh is a shim kept only so a not-yet-recopied unit still works.
To be usable at boot the script needed three things it did not have:
- Skip the rebuild when HEAD already matches origin/main and a standalone
build exists, so a reboot is not a full rebuild every time (--force
overrides).
- Do not restart PM2 when its daemon is not up yet. At boot this unit runs
before pm2-root.service, and probing with `pm2 describe` would spawn a
stray daemon that races the real service, so it checks the pid file.
- A flock single-instance lock, so a boot run and a manual run cannot install
into the same tree at once.
It also picks up prisma migrate deploy / generate and the .env copy that only
the boot script had. The log moved to /var/log/my-weekly-todo-update.log.
v1.115.2
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
next.config.js sets output: 'standalone', but Next.js does not copy public/
or .next/static/ into .next/standalone — that is a documented manual step.
Skipping it shipped new server code referencing new chunk hashes while the
old chunks stayed in the standalone tree, so those requests 404'd, React
never hydrated, and pages hung on "Loading session..." while the API kept
returning 200.
v1.114.2
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Moves the deploy script off the server and into the repo as
scripts/update-app.sh, with the two fixes made during the v1.114.0 deploy:
- set -euo pipefail, so a failed git pull aborts instead of rebuilding and
restarting PM2 on unchanged code while printing "Update Complete"
- npm ci instead of npm install, which rewrote package-lock.json and left the
tree dirty so the next git pull aborted
The body is wrapped in a main() function because the script lives inside the
checkout it updates, and git pull rewriting it mid-run would leave bash
reading from a stale file offset.
v1.114.1
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
