Skip to main content

Upgrade

# 자동 (META-INF/VERSION.JSON 의 최신 build_date)
sudo bash /opt/kopens/install/upgrade.sh

# 특정 tag
sudo bash /opt/kopens/install/upgrade.sh --to 2026-20260601

# 직전 tag 로 rollback
sudo bash /opt/kopens/install/upgrade.sh --rollback

# 새 빌드 존재 여부만 확인
sudo bash /opt/kopens/install/upgrade.sh --check

Execution sequence (automatic):

  1. Compare build_date in META-INF/VERSION.JSON
  2. pre-snapshot → /data1/pp-backups/pe-backup-<ts>.tar.zst (calls backup.sh; can be bypassed with SKIP_BACKUP=1)
  3. docker pull <new-tag>
  4. /etc/kopens/version.env update + systemctl restart
  5. /api/v1/system/health 90-second probe
  6. all UP → success / DEGRADED → automatic rollback + snapshot path guidance

Progress status: cat /etc/kopens/upgrade.status — JSON (idle / running / success / rolling-back / rolled-back / error + snapshot path).

UI / API trigger:

# 컨테이너 안 (ConfigController) 에서 자동
echo '{"target_tag":"2026-20260601"}' > /etc/kopens/upgrade.request
# 호스트의 kopens-upgrade.path (inotify) 감지 → upgrade.sh --from-trigger 실행

Detailed guide: the "OTA Upgrade" section of Container Mode Operations Guide.

(legacy) Native Mode​

$PE_HOME/bin/upgrade.sh

Execution sequence:

  1. Call backup.sh (automatic)
  2. Download the latest install.sh from product.kopens.io
  3. install.sh downloads the artifact (KOPENS_EDGE_V<VER>.tar.zstd) → swaps in /opt/kopens
  4. Call the new start.sh

Normal output:

[1] BACKUP ... done (/data1/pp-backup/20260508-040215/)
[2] DOWNLOAD install.sh ... 22 KB OK
[3] EXEC install.sh ...
   - downloading KOPENS_EDGE_V2026.tar.zstd ... 727 MB OK
   - extracting ... OK
   - finalizing ...
[4] START ... (start.sh 호출됨)
UPGRADE COMPLETED.

Duration depends on artifact size + line speed + decompression (typically 5–15 minutes).

The gateway is down during the upgrade

While upgrade.sh is running, the gateway is stopped. Notify external SCADA / Sparkplug receivers in advance. Run it during a night-time inspection window whenever possible.

2. Pre-checks​

Before starting the upgrade:

# 1. 외부 망이 충분히 빠른지 — 회선이 느리면 download 만 수십 분
$PE_HOME/bin/network-speed-test.sh

# 2. 디스크 여유 (산출물 크기의 2–3배는 있어야)
df -h /opt /data1

# 3. 살아있는 백업이 있는지
ls -lt /data1/pp-backup | head -3

3. Verifying the Upgrade Result​

$PE_HOME/bin/ps.sh                              # 모든 컴포넌트 살아있는지
curl -s http://127.0.0.1/api/v1/edge | jq      # 버전/엣지 ID 새 값 확인
curl -s http://127.0.0.1/ui/opcua/tree | jq '.data.tree | length'  # 사이트/태그 정상
$PE_HOME/bin/log-viewer.sh                      # 30초 정도 봐서 SEVERE 없는지
# (Ctrl+C)

Also confirm on the external SCADA / Sparkplug receiver side that data is arriving normally.

4. Common Pitfalls​

SymptomCause / Resolution
Stalls at download install.shCheck the external network / product.kopens.io. Verify directly with curl -v http://product.kopens.io/...
The new artifact is corrupted after extractingDownload was interrupted. Check disk space and retry
Only some components are alive after the upgradeStep-by-step failure in start.sh — use log-viewer.sh to identify where it stopped
Roll back to a previous version after the upgradeRestore from backup → Backup / Restore

5. firmware.sh — OS Package Update​

$PE_HOME/bin/firmware.sh

Internally:

dnf update -y
dnf clean all

Use this when applying security patches on a quarterly basis. If a kernel update is included, a reboot is required.

$PE_HOME/bin/firmware.sh
needs-restarting -r              # 새 커널이 깔렸는지 확인
# 결과가 "Reboot is required" 이면
$PE_HOME/bin/stop.sh
$PE_HOME/bin/reboot.sh

6. Learn More​

  • Integrated upgrade scenario (safe procedure) → Operational Scenarios — Deploying a New Version
  • Artifacts are produced by the GitLab release chain (edge:libs → edge:deploy-dev → edge:image → edge:staging → edge:version).
  • Installation / deployment (for developers): Installation / Deployment