feat(bootstrap): Velocity 堆可由 FELIS_VELOCITY_XMX 配置，全新安装预建 k3s 存储根目录消除 reaper 警告

2026-09-25 01:20:52 +08:00
1 parent 3e61fde06d
commit 989be55b4a
@@ -42,6 +42,8 @@
 #                     through the handshake address instead of modern forwarding
 #                     (default: legacy18). Read once at Velocity start, so changing it
 #                     means re-running this script and restarting the proxy.
+#   FELIS_VELOCITY_XMX maximum heap of the Velocity proxy, as <n>M or <n>G (default: 1G;
+#                     at least 256M). docs/operations.md sizes it by player count.
 #   FELIS_VELOCITY_FORK_JAR path to a Felis-Legacy Velocity fork build to install as the
 #                     proxy instead of the stock download (default: unset, stock).
 #   FELIS_VELOCITY_FORK_JAR_SHA256 expected sha256 of that jar. REQUIRED whenever the jar
@@ -91,9 +93,9 @@
 #   FELIS_WORLDS_HOST_PATH node directory holding the world volumes (on the k3s this
 #                     installer provisions: /var/lib/rancher/k3s/storage). Setting it
 #                     enables the daily retention reaper, which archives and then deletes
-#                     worlds idle beyond the retention window, and grants the reaper's
-#                     uid (1000) traverse access to that root — k3s ships it 0700
-#                     root:root (default: unset = no reaper)
+#                     worlds idle beyond the retention window; the reaper reads that
+#                     root as root, so it keeps k3s's own 0700 root:root
+#                     (default: unset = no reaper)
 #   FELIS_REGISTRY_STORAGE / FELIS_UPLOADS_STORAGE / FELIS_BACKUP_STORAGE capacity the
 #                     registry, uploads and world-archive PVCs request on first install
 #                     (defaults: 10Gi, 5Gi, 10Gi). An existing claim keeps its size; on
@@ -170,6 +172,9 @@ FELIS_BACKUP_STORAGE="${FELIS_BACKUP_STORAGE:-}"
 # from its volumeName. Left unset, no reaper CronJob renders and archives accumulate until
 # the backup PVC fills (then backups fail loudly; nothing is deleted).
 FELIS_WORLDS_HOST_PATH="${FELIS_WORLDS_HOST_PATH:-}"
+# k3s's local-path provisioner root. It appears with the first volume the provisioner
+# creates, which on a fresh install is after the reaper's PV has been applied.
+K3S_STORAGE_ROOT="/var/lib/rancher/k3s/storage"
 # Control-plane database backups (felis db backup): a daily timer bundles pg_dump with the
 # /etc/felis state a rebuild needs, and every upgrade that has migrations to apply snapshots
 # the database first (felis migrate up). The directory sits outside /var/lib/rancher on
@@ -211,6 +216,7 @@ FELIS_NANO_PROXY_CIDR="${FELIS_NANO_PROXY_CIDR:-}"
 # needs this. Overridable because adding a second 1.8 backend otherwise means editing this
 # script; it is still a restart-time list, not one that follows the CRs.
 FELIS_LEGACY_FORWARDING_SERVERS="${FELIS_LEGACY_FORWARDING_SERVERS:-legacy18}"
+FELIS_VELOCITY_XMX="${FELIS_VELOCITY_XMX:-1G}"
 # The Go tarball is unpacked and run as root, so the default version is pinned by the sha256
 # go.dev/dl publishes for each architecture install_go_toolchain handles. Move all three
 # together; any other FELIS_GO_VERSION has to bring its own FELIS_GO_SHA256.
@@ -703,6 +709,24 @@ validate_settings() {
    pinned|latest) ;;
    *) die "FELIS_GAME_STACK must be pinned or latest (got '${FELIS_GAME_STACK}')" ;;
  esac
+  [ "$(heap_megabytes "$FELIS_VELOCITY_XMX")" -ge 256 ] \
+    || die "FELIS_VELOCITY_XMX must be a heap size of at least 256M, written <n>M or <n>G (got '${FELIS_VELOCITY_XMX}')"
+}
+
+# heap_megabytes prints a JVM heap size written <n>M or <n>G in megabytes, or 0 for any
+# other spelling.
+heap_megabytes() {
+  local n="${1%?}"
+  case "$n" in
+    ""|*[!0-9]*|0*) echo 0; return ;;
+  esac
+  # Six digits of gigabytes is far past any host; the cap keeps the arithmetic in range.
+  [ "${#n}" -le 6 ] || { echo 0; return; }
+  case "$1" in
+    *[Mm]) echo "$n" ;;
+    *[Gg]) echo "$((n * 1024))" ;;
+    *) echo 0 ;;
+  esac
 }

 # validate_offsite_settings checks the FELIS_OFFSITE_* inputs before anything is
@@ -2371,6 +2395,10 @@ install_velocity_service() {
  # sees it -- unquoted, that spelling would hand java a stray "legacy112" argument and the unit
  # would not start. Quoting keeps the whole property one argv item.
  local legacy_forwarding_servers="${FELIS_LEGACY_FORWARDING_SERVERS}"
+  # -Xms stays at 512M so a small proxy does not reserve its whole ceiling up front, unless
+  # the ceiling itself is lower (the JVM refuses an initial heap above the maximum).
+  local xmx="$FELIS_VELOCITY_XMX" xms="512M"
+  [ "$(heap_megabytes "$xmx")" -ge 512 ] || xms="$xmx"
  cat > "$VELOCITY_SERVICE" <<EOF
 [Unit]
 Description=Felis Velocity proxy (Mojang authentication + modern forwarding)
@@ -2382,7 +2410,7 @@ Type=simple
 User=${VELOCITY_USER}
 Group=${VELOCITY_USER}
 WorkingDirectory=${VELOCITY_DIR}
-ExecStart=${JRE_DIR}/bin/java -Xms512M -Xmx1G -XX:+UseG1GC -XX:+ParallelRefProcEnabled -XX:+AlwaysPreTouch -Dmojang.sessionserver=http://${api_ip}:8081/session/minecraft/hasJoined "-Dfelis.legacy-forwarding.servers=${legacy_forwarding_servers}" -jar ${VELOCITY_DIR}/velocity.jar
+ExecStart=${JRE_DIR}/bin/java -Xms${xms} -Xmx${xmx} -XX:+UseG1GC -XX:+ParallelRefProcEnabled -XX:+AlwaysPreTouch -Dmojang.sessionserver=http://${api_ip}:8081/session/minecraft/hasJoined "-Dfelis.legacy-forwarding.servers=${legacy_forwarding_servers}" -jar ${VELOCITY_DIR}/velocity.jar
 Restart=on-failure
 RestartSec=5
 NoNewPrivileges=yes
@@ -3330,7 +3358,13 @@ deploy_bundle() {
    # through a static hostPath PV, so the host directory keeps k3s's own 0700 root:root and
    # needs no extra grant. It must exist, though: the PV declares type Directory.
    if [ ! -d "$FELIS_WORLDS_HOST_PATH" ]; then
-      warn "worlds root ${FELIS_WORLDS_HOST_PATH} does not exist yet; the reaper CronJob cannot start until it does (hostPath type Directory)"
+      if [ "$FELIS_WORLDS_HOST_PATH" = "$K3S_STORAGE_ROOT" ]; then
+        # The provisioner would create it moments later with this same 0700 root:root;
+        # creating it now keeps a fresh install from warning about its own default.
+        install -d -m 0700 -o root -g root "$FELIS_WORLDS_HOST_PATH"
+      else
+        warn "worlds root ${FELIS_WORLDS_HOST_PATH} does not exist yet; the reaper CronJob cannot start until it does (hostPath type Directory)"
+      fi
    fi
    manifest_args+=(--worlds-host-path "$FELIS_WORLDS_HOST_PATH" --archive-local-path "$FELIS_ARCHIVE_LOCAL_PATH")
  fi
@@ -875,6 +875,8 @@ run_bundle_flags() { # backup-pvc worlds-host-path
    myManifests() { printf "%s\n" "$@"; }
    setfacl() { printf "SETFACL %s\n" "$*"; }
    chmod() { printf "CHMOD %s\n" "$*"; }
+    install() { printf "INSTALL %s\n" "$*"; }
+    K3S_STORAGE_ROOT="${K3S_STORAGE_ROOT_T:-/var/lib/rancher/k3s/storage}"
    node_global_cidrs() { printf "203.0.113.7/32\n2001:db8::7/128\n"; }
    pvc_size() { case "$2" in registry) printf "20Gi\n" ;; esac; }
    run_bundle() {
@@ -935,6 +937,14 @@ missing="/tmp/felis-worlds-root-must-not-exist-$$"
 out="$(run_bundle_flags felis-backups "$missing")"
 expect "a missing worlds root is warned about, not silently skipped" "WARN: worlds root $missing does not exist yet" "$out"

+# On a fresh install the k3s default root does not exist until the provisioner's first
+# volume; the installer creates it with k3s's own mode instead of warning about its default.
+out="$(K3S_STORAGE_ROOT_T="$missing" run_bundle_flags felis-backups "$missing")"
+expect "a missing k3s storage root is created with k3s's mode" "INSTALL -d -m 0700 -o root -g root $missing" "$out"
+case "$out" in
+  *WARN*) echo "FAIL: the installer's own default root must not be warned about: $out"; fails=$((fails + 1)) ;;
+esac
+
 # The reaper reads the root as root with DAC_OVERRIDE through a static PV, so an existing
 # root is left exactly as k3s shipped it: uid 1000 is now the game servers' uid, and a
 # traverse grant for it on the node's storage root would serve nothing but them.
@@ -1709,15 +1719,15 @@ esac
 # felis-api's transactions) when restarted, so a rerun that changed none of them must leave
 # them running, and one that changed a thing must still restart it.

-vsblock="$(awk '/^install_velocity_service\(\) \{/,/^}/' "$BS"; awk '/^velocity_fingerprint\(\) \{/,/^}/' "$BS")"
+vsblock="$(awk '/^install_velocity_service\(\) \{/,/^}/' "$BS"; awk '/^velocity_fingerprint\(\) \{/,/^}/' "$BS"; awk '/^heap_megabytes\(\) \{/,/^}/' "$BS")"
 [ -n "$vsblock" ] || { echo "FAIL: no install_velocity_service found in $BS"; exit 1; }
 vdir="$(mktemp -d)"
 mkdir -p "$vdir/v/plugins/felis-link" "$vdir/jre"
 printf 'jar\n' > "$vdir/v/velocity.jar"
 printf 'plugin\n' > "$vdir/v/plugins/felis-velocity.jar"
 printf 'JAVA_VERSION="25"\n' > "$vdir/jre/release"
-run_velocity_service() { # is-active(0|1)
-  ACTIVE="$1" VELOCITY_SERVICE="$vdir/unit" VELOCITY_DIR="$vdir/v" JRE_DIR="$vdir/jre" \
+run_velocity_service() { # is-active(0|1) [heap]
+  ACTIVE="$1" FELIS_VELOCITY_XMX="${2:-1G}" VELOCITY_SERVICE="$vdir/unit" VELOCITY_DIR="$vdir/v" JRE_DIR="$vdir/jre" \
  VELOCITY_FINGERPRINT="$vdir/fp" VELOCITY_USER=felis-velocity FELIS_GAME_PORT=25565 \
  FELIS_LEGACY_FORWARDING_SERVERS='' bash -c '
    set -Eeuo pipefail
@@ -1734,6 +1744,7 @@ run_velocity_service() { # is-active(0|1)
 }
 out="$(run_velocity_service 1)"
 expect "a proxy with no recorded start is restarted" "SYSTEMCTL restart felis-velocity" "$out"
+expect "the default heap is 512M..1G" "java -Xms512M -Xmx1G " "$(cat "$vdir/unit")"
 [ -s "$vdir/fp" ] && echo "PASS the restart records what the proxy runs" \
  || { echo "FAIL no fingerprint was recorded after the restart"; fails=$((fails + 1)); }
 out="$(run_velocity_service 1)"
@@ -1747,6 +1758,10 @@ expect "a changed plugin jar restarts the proxy" "SYSTEMCTL restart felis-veloci
 expect "a stopped proxy is started whatever the fingerprint" "SYSTEMCTL restart felis-velocity" "$(run_velocity_service 0)"
 printf 'JAVA_VERSION="25.0.1"\n' > "$vdir/jre/release"
 expect "a patched JRE restarts the proxy" "SYSTEMCTL restart felis-velocity" "$(run_velocity_service 1)"
+expect "a new heap size restarts the proxy" "SYSTEMCTL restart felis-velocity" "$(run_velocity_service 1 3G)"
+expect "the unit carries the new ceiling" "java -Xms512M -Xmx3G " "$(cat "$vdir/unit")"
+run_velocity_service 1 384M >/dev/null
+expect "a ceiling below 512M is also the initial heap" "java -Xms384M -Xmx384M " "$(cat "$vdir/unit")"
 rm -rf "$vdir"

 ssblock="$(awk '/^restart_existing_system_servers\(\) \{/,/^}/' "$BS")"
@@ -1903,6 +1918,21 @@ expect "a v6 node address gets a v6 rule" "tcp dport 5432 ip6 saddr 2001:db8::7
 rm -rf "$fwdir"


+# --- the proxy heap -----------------------------------------------------------------------
+hblock="$(awk '/^heap_megabytes\(\) \{/,/^}/' "$BS")"
+[ -n "$hblock" ] || { echo "FAIL: no heap_megabytes found in $BS"; exit 1; }
+heap() { bash -c "$hblock"'
+heap_megabytes "$1"' _ "$1"; }
+expect "a heap in gigabytes converts to megabytes" "2048" "$(heap 2G)"
+expect "a heap in megabytes is kept" "768" "$(heap 768m)"
+for bad in 1 1K 0G 01G G -1G 1.5G 9999999G; do
+  expect "the heap spelling '$bad' is refused" "0" "$(heap "$bad")"
+done
+case "$(awk '/^install_velocity_service\(\) \{/,/^}/' "$BS")" in
+  *'-Xms${xms} -Xmx${xmx} '*) echo "PASS the proxy unit takes its heap from FELIS_VELOCITY_XMX" ;;
+  *) echo "FAIL the proxy unit's heap is not FELIS_VELOCITY_XMX"; fails=$((fails + 1)) ;;
+esac
+
 # --- reproducible image ids ---------------------------------------------------------------
 # restart_existing_system_servers compares image ids across runs; a default BuildKit
 # provenance attestation (it carries a timestamp) would make every rebuild look new.