deploy.sh era documentat doar in README-ul VM 109; nu aparea nicaieri in
proxmox/README.md, deci nu il gaseai daca nu stiai deja ca exista. Iar
`deploy.sh --help` raspundea cu o singura linie pe stderr si exit 2 - o
unealta documentata trebuie sa se poata explica singura.
- deploy.sh: --help / -h cu utilizare completa (optiuni, ce verifica inainte
sa copieze, distributia pe noduri si de ce e asimetrica, ce face --guest cu
VM 109 oprit, si faptul ca nu atinge cron-ul). Exit 0 pentru --help, exit 2
cu mesaj pe stderr pentru optiune necunoscuta.
- proxmox/README.md: deploy.sh, vm109-patch-window.sh si check_servicing.ps1
in tabelul de fisiere, bloc "Deploy modificari" in Quick Start, plus intrari
in "Vreau sa..." pentru deploy, pentru patching si pentru simptomul
"Restore failed fara log RMAN".
- README VM 109: --help in lista de comenzi si un exemplu de output --check,
ca sa se vada diferenta dintre "DIFERA" si "LIPSESTE pe nod".
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BhQBTegE4PiMPPaapLHjkc
Pana acum deploy-ul se facea manual (`cp ... /opt/scripts/`), nedocumentat
nicaieri - de unde si fisierele weekly-dr-test-proxmox.sh.bak-* ramase pe
pveelite. deploy.sh face aceiasi pasi, dar refuza situatiile care au costat
deja timp:
- suprascrierea unui script AFLAT IN EXECUTIE: bash citeste scriptul
incremental de pe disc, iar suprascrierea unui test DR in curs (~18 min)
corupe executia;
- CRLF: bash pe Linux raspunde "$'\r': command not found";
- transfer trunchiat: bash -n local si inca o data pe nod dupa copiere;
- flag-ul vm109-debug.flag lasat in urma dupa deploy pe guest (dezarmeaza
permanent watchdog-ul, exact apararea care a prins incidentul 04-20).
Distributia pe noduri e asimetrica intentionat si e codificata explicit in
script: scripturile cluster-aware (test DR, patch window, watchdog) pe ambele
noduri fiindca urmaresc VM 109 dupa failover HA; ZFS/mirror doar pe pveelite
unde traieste datasetul; alertele si failover-ul doar pe pvemini, fiindca
reactioneaza la caderea pveelite - puse pe pveelite ar fi inutile si ar
inlesni exact split-brain-ul pe care incearca sa-l previna.
Verificarea "ruleaza acum" foloseste pgrep -f '[/]opt/scripts/x.sh'.
Parantezele nu sunt cosmetice: comanda trimisa prin ssh apare ea insasi in
lista de procese, deci forma fara paranteze se gaseste pe sine si raporteaza
orice script ca fiind in executie (verificat: exit 0 vs exit 1).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BhQBTegE4PiMPPaapLHjkc
Testul DR din 2026-08-08 a raportat "Restore failed" dupa 11 secunde, fara
niciun log RMAN. Cauza nu a fost restore-ul: KB5101001 fusese descarcat in
timpul testului din 2026-08-01 (singurul moment in care VM 109 e pornit),
a ramas staged dupa qm stop si s-a finalizat la boot-ul testului urmator.
Cronologie din Event Log-ul guest-ului:
06:00:58 RestartManager 10010 - nu poate reporni powershell.exe (restore-ul)
06:01:04 SCM 7034 - OpenSSH SSH Server terminat neasteptat
06:01:06 pveelite: client_loop: send disconnect: Broken pipe -> FAILED
06:01:38 VM-ul se reboteaza singur
Fereastra testului (Sambata 06:00) era in afara Active Hours (08:00-17:00),
deci pentru Windows era fereastra de mentenanta valida - iar VM 109 fiind
pornit doar in timpul testului, aceea era singura fereastra posibila.
Agravant: sshd nu avea acsiuni de recovery (RESET_PERIOD 0), deci dupa ce a
murit a ramas mort si au esuat si colectarea logului si shutdown-ul gratios.
Masuri:
- NoAutoUpdate=1 + AUOptions=2 pe VM 109 (aplicat direct in registry)
- actiuni de recovery pentru sshd: restart la 5s/10s/30s, reset=86400
- guard "STEP 3b: Windows servicing" inainte de restore (check_servicing.ps1):
asteapta idle 300s, consuma controlat un reboot in asteptare, altfel
abandoneaza cu "ABORTED - Windows servicing" in loc de un "Restore failed"
inselator. Fail-open daca checkul lipseste - nu are voie sa pice testul.
- fereastra lunara de patching (vm109-patch-window.sh + install_updates.ps1),
prima duminica 03:00, cu re-armare NoAutoUpdate=1 indiferent de rezultat
Adaugat si .gitattributes: cu core.autocrlf=true scripturile .sh ajungeau in
working tree cu CRLF, iar ele se deployeaza prin scp direct pe Proxmox.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BhQBTegE4PiMPPaapLHjkc
Adds end-to-end procedure for moving production back from DR (10.0.20.37)
to a repaired/reinstalled PRIMARY (10.0.20.36): final RMAN backup on DR
in restricted/read-only mode, RMAN restore on PRIMARY, app connection
switch, scheduled-task reactivation, VM 109 stop. Companion PowerShell
script handles the restore with sanity checks (IP, NFS, backup freshness)
and aborts if Oracle major version != 19, since failback to 21c would
need an extra dictionary upgrade step (~30-60 min) that adds untested
risk during the critical window — recommended path is 19c failback then
upgrade later in a planned window.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
VM 109 returned to its original home on pveelite, co-located with
oracle-backups NFS storage. The README is updated to reflect that:
the VM is now in HA (ha-prefer-pveelite, state=stopped, nofailback=1)
rather than excluded from HA, and the new layered defences (trap
guard, watchdog cron, dynamic memory pre-flight, max_restart caps)
are documented alongside the original 8a0c557 trap.
Adds a Storage Failover section describing the pveelite -> pvemini
manual failover flow: email alert from pveelite-down-alert.sh,
failover-dr-to-pvemini.sh on the surviving node, failback when
pveelite returns. The pve1 nightly mirror is the third copy.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Following the 2026-04-20 cluster outage, the cluster README now covers
HA resource limits, corosync token tuning (10s tolerance for USB glitches),
rasdaemon/netconsole/kdump diagnostic stack on pvemini, mail relay via
mail.romfast.ro with SMTP auth, OOM alerting via cron, and swap on pveelite.
VM 109 README now clearly states it was removed from HA and is only
started by the weekly DR test script.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Move oracle/migration-scripts/ to proxmox/lxc108-oracle/migration/
- Move oracle/roa/ and oracle/roa-romconstruct/ to proxmox/lxc108-oracle/sql/
- Move oracle/standby-server-scripts/ to proxmox/vm109-windows-dr/
- Move chatbot/ to proxmox/lxc104-flowise/
- Update proxmox/README.md with new structure and navigation
- Update all documentation with correct directory references
- Remove unused input/claude-agent-sdk/ files
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
