Files
PFVCluster/perf/deploy-check.sh
T
mrcharles bfb2d44183 docs(audit): RAM/CPU placement audit + codify capacity goals
Add CAPACITY-GOALS.md (80% RAM / ~50% idle CPU targets + workload placement
model) and AUDIT-2026-07-29.md (live end-to-end analysis of all 7 hosts). Key
findings: memory goal met fleet-wide (highest 72%); CPU idles 0-6% (chase via
workloads, not hardware); tsys5 is single-CPU not dual; cnode4/5 are stale
running VMs. Answers the wnode-sizing question (grow tsys9, not tsys1), confirms
tsys6/7 for RackRental, and gives a DDR3 compatibility decision tree (RDIMM-only
servers reject consumer UDIMM). Refreshed STATUS/docmap and the returned audit
logs, and fixed an empty-array abort in deploy-check.sh.

💘 Generated with Crush

Assisted-by: Crush:glm-5.2
2026-07-29 05:32:12 -05:00

175 lines
6.2 KiB
Bash
Executable File

#!/bin/bash
###############################################################################
# deploy-check.sh
#
# Deploys scripts/check.sh to each reachable Proxmox host, executes it
# read-only, and pulls the resulting log back to returned-logs/.
#
# EXPLICITLY SKIPS:
# - pfv-tsys2 (off the air per user)
# - pfv-tsys9 (off the air per user; also not in original inventory)
#
# Safety features:
# - BatchMode=yes : never hang on a password prompt
# - ConnectTimeout=8 : fail fast on dead hosts
# - per-host try/skip : one bad host never aborts the run
# - ServerAliveInterval : detect hung connections
# - read-only script : check.sh modifies nothing on the target
###############################################################################
set -uo pipefail
SELF_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
ROOT_DIR="$(cd "$SELF_DIR/.." && pwd)"
CHECK_SH="$SELF_DIR/scripts/check.sh"
LOG_DIR="$ROOT_DIR/returned-logs"
mkdir -p "$LOG_DIR"
# ONLY the hosts the user told us are alive.
HOSTS=(pfv-tsys1 pfv-tsys3 pfv-tsys4 pfv-tsys5 pfv-tsys6 pfv-tsys7 pfv-tsys9)
# Common ssh options: non-interactive, fail-fast, no host-key prompt blocking.
SSH_OPTS=(-o BatchMode=yes
-o ConnectTimeout=8
-o ServerAliveInterval=10
-o ServerAliveCountMax=3
-o StrictHostKeyChecking=accept-new)
log() { printf '[%s] %s\n' "$(date +%H:%M:%S)" "$*"; }
if [ ! -r "$CHECK_SH" ]; then
echo "FATAL: $CHECK_SH not found" >&2
exit 1
fi
# Sanity-check shellcheck clean before shipping (best-effort, non-blocking)
if command -v docker >/dev/null 2>&1; then
log "pre-flight: shellcheck on check.sh"
if ! docker run --rm -v "$SELF_DIR:/mnt" -w /mnt \
koalaman/shellcheck:stable --severity=style --format=gcc scripts/check.sh \
>"$LOG_DIR/_shellcheck.preflight.txt" 2>&1; then
log "WARNING: shellcheck reported issues — see _shellcheck.preflight.txt"
log " aborting deploy to avoid shipping a broken script"
exit 1
fi
log "pre-flight: shellcheck clean"
fi
summary_pass=()
summary_fail=()
declare -A HOST_PID # host -> background pid
declare -A HOST_MARKER # host -> per-host marker file
# Per-host worker — runs in background, one per host, all in parallel.
# Writes status into a marker file consumed by the parent.
worker() {
local host="$1"
local marker="$LOG_DIR/_marker.$host"
: > "$marker" # truncate
echo "running" >> "$marker"
local short=""
if ! ssh "${SSH_OPTS[@]}" "root@$host" 'echo ok' >/dev/null 2>&1; then
echo "fail unreachable" >> "$marker"
return
fi
if ! scp "${SSH_OPTS[@]}" "$CHECK_SH" "root@$host:/root/check.sh" >/dev/null 2>&1; then
echo "fail scp-upload-failed" >> "$marker"
return
fi
local remote_size
remote_size=$(ssh "${SSH_OPTS[@]}" "root@$host" 'wc -c < /root/check.sh' 2>/dev/null || echo 0)
if [ "${remote_size:-0}" -lt 1000 ]; then
echo "fail upload-corrupt" >> "$marker"
return
fi
local remote_stdout
remote_stdout=$(ssh "${SSH_OPTS[@]}" "root@$host" \
'chmod +x /root/check.sh && bash /root/check.sh' 2>&1)
local rc=$?
if [ "$rc" -ne 0 ]; then
echo "fail check-exit-$rc" >> "$marker"
# don't return - still try to pull whatever log got produced
fi
short=$(printf '%s\n' "$remote_stdout" | grep -oE 'Wrote: /root/[a-zA-Z0-9_-]+\.log' | head -n1 | awk '{print $2}')
if [ -z "$short" ]; then
short=$(ssh "${SSH_OPTS[@]}" "root@$host" 'echo "/root/$(hostname -s).log"' 2>/dev/null)
fi
if [ -z "$short" ]; then
echo "fail no-log-path" >> "$marker"
return
fi
if ! scp "${SSH_OPTS[@]}" "root@$host:$short" "$LOG_DIR/" >/dev/null 2>&1; then
echo "fail scp-download-failed" >> "$marker"
return
fi
local local_name local_path
local_name="$(basename "$short")"
local_path="$LOG_DIR/$local_name"
if [ ! -s "$local_path" ]; then
echo "fail local-empty" >> "$marker"
return
fi
echo "ok $local_name $(wc -c < "$local_path") $(wc -l < "$local_path")" >> "$marker"
}
# ---- launch all workers in parallel --------------------------------------
log "launching ${#HOSTS[@]} hosts in parallel..."
for host in "${HOSTS[@]}"; do
rm -f "$LOG_DIR/_marker.$host"
worker "$host" &
HOST_PID[$host]=$!
HOST_MARKER[$host]="$LOG_DIR/_marker.$host"
log " launched $host (pid ${HOST_PID[$host]})"
done
# ---- wait for all, with periodic progress --------------------------------
remaining=("${HOSTS[@]}")
while [ "${#remaining[@]}" -gt 0 ]; do
sleep 10
new_remaining=()
for host in "${remaining[@]}"; do
if ! kill -0 "${HOST_PID[$host]}" 2>/dev/null; then
# process finished
wait "${HOST_PID[$host]}" 2>/dev/null || true
marker="${HOST_MARKER[$host]}"
if [ -r "$marker" ]; then
status_line="$(tail -n1 "$marker")"
log "[$host] done: $status_line"
case "$status_line" in
ok*) summary_pass+=("$host:$status_line") ;;
fail*) summary_fail+=("$host:$status_line") ;;
*) summary_fail+=("$host:unknown") ;;
esac
else
log "[$host] done but marker missing"
summary_fail+=("$host:no-marker")
fi
else
new_remaining+=("$host")
fi
done
if [ "${#new_remaining[@]}" -gt 0 ]; then
remaining=("${new_remaining[@]}")
else
remaining=()
fi
if [ "${#remaining[@]}" -gt 0 ]; then
log "still running: ${remaining[*]} (${#remaining[@]} hosts)"
fi
done
# Final summary
log "============================================================"
log "DEPLOY SUMMARY"
log "============================================================"
log "Passed (${#summary_pass[@]}):"
for p in "${summary_pass[@]:-}"; do [ -n "$p" ] && log " ✓ $p"; done
log "Failed (${#summary_fail[@]}):"
for f in "${summary_fail[@]:-}"; do [ -n "$f" ] && log " ✗ $f"; done
log ""
log "Contents of $LOG_DIR:"
ls -la "$LOG_DIR"
# Clean up marker files
rm -f "$LOG_DIR"/_marker.* 2>/dev/null