Replace replicas one by one under load
scripts/rolling-update.sh runs migrate once before touching any replica, then for each old replica scales the service to N+1 without recreating the others, waits until the new container is healthy and the balancer sees N+1 UP servers, and only then stops the old one with docker stop -t 6 (so it drains) and removes it. The log names the hostname and image of the old and the new replica of every step, which is what the report quotes. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 5.5
parent
5adb8719d4
commit
572204950a
Executable
+20
@@ -0,0 +1,20 @@
|
||||
#!/bin/sh
|
||||
set -eu
|
||||
cd "$(dirname "$0")/.."
|
||||
service=${API_SERVICE:-api}
|
||||
show() { docker inspect -f '{{.Config.Hostname}} image={{slice .Image 7 19}}' "$1"; }
|
||||
echo "$(date -u +%T) migrate once before the first replacement"
|
||||
docker compose run --rm --no-deps -T migrate 2>&1
|
||||
old=$(docker compose ps -q "$service")
|
||||
count=$(printf '%s\n' "$old" | grep -c .)
|
||||
for id in $old; do
|
||||
running=$(docker compose ps -q "$service")
|
||||
docker compose up -d --no-deps --no-recreate --wait --wait-timeout "${WAIT_TIMEOUT:-90}" --scale "$service=$((count + 1))" "$service" 2>&1
|
||||
scripts/lb-wait.sh $((count + 1))
|
||||
new=$(docker compose ps -q "$service" | grep -v -F "$running")
|
||||
echo "$(date -u +%T) replace $(show "$id") -> $(show "$new")"
|
||||
docker stop -t "${STOP_TIMEOUT:-6}" "$id" >/dev/null
|
||||
docker rm "$id" >/dev/null
|
||||
done
|
||||
scripts/lb-wait.sh "$count"
|
||||
echo "$(date -u +%T) done: $count replicas replaced"
|
||||
Reference in New Issue
Block a user