Oracle DR: Complete Windows VM implementation and cleanup

Major changes:
- Implemented Windows VM 109 as DR target (replaces Linux LXC)
- Tested RMAN restore successfully (12-15 min RTO, 24h RPO)
- Added comprehensive DR documentation:
  * DR_WINDOWS_VM_STATUS_2025-10-09.md - Current implementation status
  * DR_UPGRADE_TO_CUMULATIVE_PLAN.md - Plan for cumulative incremental backups
  * DR_VM_MIGRATION_GUIDE.md - Guide for VM migration between Proxmox nodes
- Updated DR_WINDOWS_VM_IMPLEMENTATION_PLAN.md with completed phases

New scripts:
- add_system_key_dr.ps1 - SSH key setup for automated transfers
- configure_listener_dr.ps1 - Oracle Listener configuration
- fix_ssh_via_service.ps1 - SSH authentication fix
- rman_restore_final.cmd - Working RMAN restore script (tested)
- transfer_to_dr.ps1 - FULL backup transfer (renamed from 02_*)
- transfer_incremental.ps1 - Incremental backup transfer (renamed from 02b_*)

Cleanup:
- Removed 19 obsolete scripts for Linux LXC DR
- Removed 8 outdated documentation files
- Organized project structure

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Marius
2025-10-09 18:54:08 +03:00
parent 6a6ffe84af
commit ac2340c967
26 changed files with 2309 additions and 6208 deletions

View File

@@ -0,0 +1,699 @@
# Oracle DR - Upgrade to Cumulative Incremental Backup Strategy
**Generated:** 2025-10-09
**Objective:** Implement cumulative incremental backups with Proxmox host storage for optimal RPO/RTO
**Target RPO:** 3-4 hours (vs current 24 hours)
**Target RTO:** 12-15 minutes (unchanged)
---
## 📋 EXECUTIVE SUMMARY
### Current State
- **Backup Strategy:** FULL daily (02:30), DIFFERENTIAL incremental (14:00)
- **Storage:** Backups transferred to VM 109 (powered OFF most of time)
- **RPO:** 24 hours (only FULL backup used for restore)
- **Issue:** DIFFERENTIAL incremental caused UNDO corruption during restore
### Proposed State
- **Backup Strategy:** FULL daily (02:30), CUMULATIVE incremental (13:00 + 18:00)
- **Storage:** Backups on Proxmox host (pveelite), mounted in VM 109 when needed
- **RPO:** 3-4 hours (using FULL + latest CUMULATIVE)
- **Benefit:** Simple, reliable restore without UNDO/SCN issues
### Why CUMULATIVE?
-**Simple restore:** FULL + last cumulative (no dependency chain)
-**No UNDO corruption:** Each cumulative is independent from Level 0
-**Better RPO:** Max 5 hours data loss (vs 24 hours)
-**Reliable:** No issues with missing intermediate backups
---
## 🎯 IMPLEMENTATION PHASES
### PHASE 1: Configure Proxmox Host Storage (15 minutes)
**Objective:** Create backup storage on pveelite host, accessible by VM 109 via mount point
**Steps:**
#### 1.1 Create backup directory on pveelite (SSH to host)
```bash
# On pveelite (10.0.20.202)
ssh root@10.0.20.202
# Create directory structure
mkdir -p /mnt/pve/oracle-backups/ROA/autobackup
chmod 755 /mnt/pve/oracle-backups
chmod 755 /mnt/pve/oracle-backups/ROA
chmod 755 /mnt/pve/oracle-backups/ROA/autobackup
# Verify
ls -la /mnt/pve/oracle-backups/ROA/autobackup
```
#### 1.2 Add mount point to VM 109 (Proxmox CLI)
```bash
# Stop VM 109 if running
qm stop 109
# Add mount point as additional storage
# This creates a VirtIO-9p mount point
qm set 109 -mp0 /mnt/pve/oracle-backups,mp=/mnt/oracle-backups
# Or via Proxmox Web UI:
# VM 109 → Hardware → Add → Mount Point
# - Source: /mnt/pve/oracle-backups
# - Mount point: /mnt/oracle-backups
# - Read-only: NO
# Start VM to test
qm start 109
```
#### 1.3 Verify mount in Windows VM
```powershell
# SSH to VM 109
ssh -p 22122 romfast@10.0.20.37
# Check if mount point appears as drive
# ⚠️ IMPORTANT: E:\ is already used in VM 109
# Mount will appear as F:\ (next available drive letter)
Get-PSDrive -PSProvider FileSystem
# Expected: C:, D:, E: (existing), F: (new mount from host)
# Verify mount path accessible
Test-Path F:\ROA\autobackup
# Create test file
New-Item -ItemType Directory -Path F:\ROA\autobackup -Force
echo "test" > F:\ROA\autobackup\test.txt
# Verify from host
exit
ssh root@10.0.20.202 "ls -la /mnt/pve/oracle-backups/ROA/autobackup/test.txt"
# Should show the test file - mount is working!
```
**⚠️ CRITICAL NOTE:**
- VM 109 already has E:\ partition
- Mount point will be **F:\** (not E:\)
- Update all scripts to use **F:\** instead of E:\
---
### PHASE 2: Modify RMAN Backup Scripts on PRIMARY (20 minutes)
**Objective:** Change incremental backups from DIFFERENTIAL to CUMULATIVE, add second daily incremental
#### 2.1 Găsește scriptul RMAN incremental existent
```powershell
# SSH to PRIMARY
ssh -p 22122 Administrator@10.0.20.36
cd D:\rman_backup
# Găsește scriptul incremental existent
Get-ChildItem *incr*.txt, *incr*.rman
# Ar trebui să vezi ceva gen:
# rman_backup_incremental.txt SAU
# rman_incremental.rman SAU similar
```
#### 2.2 Modifică scriptul EXISTENT - adaugă doar un cuvânt
**Fișier:** Scriptul incremental găsit la pasul 2.1 (ex: `D:\rman_backup\rman_backup_incremental.txt`)
**Modificare:** Găsește linia cu `INCREMENTAL LEVEL 1` și adaugă `CUMULATIVE`
**ÎNAINTE:**
```
BACKUP INCREMENTAL LEVEL 1 ...
```
**DUPĂ:**
```
BACKUP INCREMENTAL LEVEL 1 CUMULATIVE ...
```
**Asta e tot!** Un singur cuvânt adăugat.
**Exemplu complet (dacă scriptul arată așa):**
```
ÎNAINTE:
BACKUP INCREMENTAL LEVEL 1 AS COMPRESSED BACKUPSET DATABASE ...
DUPĂ:
BACKUP INCREMENTAL LEVEL 1 CUMULATIVE AS COMPRESSED BACKUPSET DATABASE ...
```
#### 2.3 Test manual
```powershell
# On PRIMARY
cd D:\rman_backup
# Rulează scriptul modificat
# Folosește numele scriptului tău existent!
rman cmdfile=rman_backup_incremental.txt log=logs\test_cumulative_$(Get-Date -Format 'yyyyMMdd_HHmmss').log
# Verifică că s-a creat backup
Get-ChildItem C:\Users\oracle\recovery_area\ROA\autobackup\*.bkp | Sort-Object LastWriteTime -Descending | Select-Object -First 3
```
---
### PHASE 3: Update Transfer Scripts (30 minutes)
**Objective:** Update transfer scripts to send backups to Proxmox host instead of VM
#### 3.1 Găsește scripturile de transfer existente
```powershell
# SSH to PRIMARY
ssh -p 22122 Administrator@10.0.20.36
cd D:\rman_backup
# Găsește scripturile de transfer
Get-ChildItem *transfer*.ps1
# Ar trebui să vezi:
# - transfer_to_dr.ps1 (pentru FULL)
# - transfer_incremental.ps1 SAU 02b_transfer_incremental_to_dr.ps1 (pentru INCREMENTAL)
```
#### 3.2 Modifică scripturile EXISTENTE - schimbă doar destinația
**Găsește în fiecare script aceste linii și modifică-le:**
**ÎNAINTE (transfer la VM):**
```powershell
$DRHost = "10.0.20.37" # VM-ul
$DRPort = "22122" # SSH pe VM
$DRUser = "romfast" # User din VM
$DRPath = "D:/oracle/backups/primary" # Path în VM
```
**DUPĂ (transfer la Proxmox host):**
```powershell
$DRHost = "10.0.20.202" # pveelite HOST
$DRPort = "22" # SSH standard pe host
$DRUser = "root" # Root pe Proxmox
$DRPath = "/mnt/pve/oracle-backups/ROA/autobackup" # Path pe host
```
**Asta e tot!** Doar 4 linii modificate în fiecare script.
#### 3.2 Setup SSH key for Proxmox host access
```powershell
# On PRIMARY (10.0.20.36)
# Generate SSH key for Proxmox host (if not exists)
ssh-keygen -t rsa -b 4096 -f C:\Users\Administrator\.ssh\id_rsa_pveelite -N ""
# Copy public key to Proxmox host
type C:\Users\Administrator\.ssh\id_rsa_pveelite.pub | ssh root@10.0.20.202 "mkdir -p ~/.ssh && cat >> ~/.ssh/authorized_keys"
# Test connection
ssh -i C:\Users\Administrator\.ssh\id_rsa_pveelite root@10.0.20.202 "echo SSH_OK"
```
#### 3.3 Test transfer script
```powershell
# On PRIMARY
cd D:\rman_backup
# Test FULL backup transfer
.\02_transfer_to_pveelite_host.ps1 -BackupType FULL
# Verify on Proxmox host
ssh root@10.0.20.202 "ls -lh /mnt/pve/oracle-backups/ROA/autobackup/*.bkp"
# Test INCREMENTAL backup transfer
.\02_transfer_to_pveelite_host.ps1 -BackupType INCREMENTAL
```
---
### PHASE 4: Update Scheduled Tasks on PRIMARY (20 minutes)
**Objective:** Create/update scheduled tasks for 2 cumulative incremental backups per day
#### 4.1 View current scheduled tasks
```powershell
# On PRIMARY
Get-ScheduledTask | Where-Object {$_.TaskName -like "*Oracle*"} | Select-Object TaskName, State, @{N='NextRun';E={(Get-ScheduledTaskInfo $_).NextRunTime}}
```
#### 4.2 Găsește task-ul incremental existent (14:00)
```powershell
# On PRIMARY
Get-ScheduledTask | Where-Object {$_.TaskName -like "*incr*" -or $_.TaskName -like "*14*"} | Select-Object TaskName, State
# Notează numele exact al task-ului
```
#### 4.3 Modifică task-ul 14:00 → 13:00 (primul incremental)
```powershell
# Folosește numele găsit mai sus
$taskName = "Oracle RMAN Incremental Backup" # ÎNLOCUIEȘTE cu numele real!
# Schimbă doar ora: 14:00 → 13:00
$trigger = New-ScheduledTaskTrigger -Daily -At "13:00"
$task = Get-ScheduledTask -TaskName $taskName
Set-ScheduledTask -TaskName $taskName -Trigger $trigger
```
#### 4.4 Clonează task-ul pentru al doilea incremental (18:00)
```powershell
# Exportă task-ul existent
$task = Get-ScheduledTask -TaskName $taskName
$xml = [xml](Export-ScheduledTask -TaskName $taskName)
# Modifică ora în XML
$xml.Task.Triggers.CalendarTrigger.StartBoundary = $xml.Task.Triggers.CalendarTrigger.StartBoundary -replace "T13:00:", "T18:00:"
# Importă ca task nou
Register-ScheduledTask -TaskName "$taskName 1800" -Xml $xml.OuterXml
# Sau mai simplu - copiază task-ul din Task Scheduler GUI și schimbă ora
```
#### 4.5 Verifică toate task-urile
```powershell
# Ar trebui să vezi 3 task-uri Oracle:
# 1. FULL (02:30) - neschimbat
# 2. INCREMENTAL (13:00) - modificat din 14:00
# 3. INCREMENTAL (18:00) - clonat din 13:00
Get-ScheduledTask | Where-Object {$_.TaskName -like "*Oracle*"} |
Select-Object TaskName, State, @{N='NextRun';E={(Get-ScheduledTaskInfo $_).NextRunTime}} |
Format-Table -AutoSize
```
#### 4.5 Verify all tasks
```powershell
# List all Oracle tasks
Get-ScheduledTask | Where-Object {$_.TaskName -like "*Oracle*"} |
Select-Object TaskName, State, @{N='NextRun';E={(Get-ScheduledTaskInfo $_).NextRunTime}} |
Format-Table -AutoSize
# Expected tasks:
# 1. Oracle RMAN Full Backup 0230 - Daily 02:30
# 2. Oracle RMAN Cumulative Backup 1300 - Daily 13:00
# 3. Oracle RMAN Cumulative Backup 1800 - Daily 18:00
```
---
### PHASE 5: Update DR Restore Script (30 minutes)
**Objective:** Update restore script to read backups from mount point (E:\) and handle cumulative backups
#### 5.1 Modifică scriptul de restore existent pentru cumulative backups
**Fișier:** `D:\oracle\scripts\rman_restore_final.cmd` (scriptul tău existent)
**Modificări necesare:**
**1. Schimbă locația backup-urilor:**
```cmd
REM ÎNAINTE:
set BACKUP_DIR=C:/Users/oracle/recovery_area/ROA/autobackup
REM DUPĂ (⚠️ F:\ nu E:\ - E:\ e deja folosit în VM!):
set BACKUP_DIR=F:/ROA/autobackup
```
**2. Verifică că mount point-ul e accesibil:**
Adaugă la început:
```cmd
REM Verifică mount point
if not exist F:\ROA\autobackup (
echo ERROR: Mount point F:\ not accessible!
echo Make sure VM has mount point configured and host is reachable
exit /b 1
)
set PFILE=C:\Users\oracle\admin\ROA\pfile\initROA.ora
set LOG_FILE=D:\oracle\logs\restore_cumulative_%date:~-4%%date:~3,2%%date:~0,2%_%time:~0,2%%time:~3,2%%time:~6,2%.log
echo ============================================================================
echo Oracle DR Restore - FULL + CUMULATIVE Incremental
echo ============================================================================
echo DBID: %DBID%
echo Backup Location: %BACKUP_DIR% (mount from Proxmox host)
echo Log: %LOG_FILE%
echo ============================================================================
REM Step 1: Shutdown database if running
echo.
echo [STEP 1/8] Shutting down database...
echo SHUTDOWN ABORT; > D:\oracle\temp\shutdown.sql
echo EXIT; >> D:\oracle\temp\shutdown.sql
sqlplus / as sysdba @D:\oracle\temp\shutdown.sql 2>nul
timeout /t 5 /nobreak >nul
REM Step 2: Startup NOMOUNT
echo.
echo [STEP 2/8] Starting instance NOMOUNT...
echo STARTUP NOMOUNT PFILE='%PFILE%'; > D:\oracle\temp\nomount.sql
echo EXIT; >> D:\oracle\temp\nomount.sql
sqlplus / as sysdba @D:\oracle\temp\nomount.sql
if %errorlevel% neq 0 (
echo ERROR: Failed to startup NOMOUNT
exit /b 1
)
REM Step 3: Restore control file
echo.
echo [STEP 3/8] Restoring control file...
echo SET DBID %DBID%; > D:\oracle\temp\restore_ctl.rman
echo. >> D:\oracle\temp\restore_ctl.rman
echo RUN { >> D:\oracle\temp\restore_ctl.rman
echo ALLOCATE CHANNEL ch1 DEVICE TYPE DISK; >> D:\oracle\temp\restore_ctl.rman
echo # Find latest control file backup >> D:\oracle\temp\restore_ctl.rman
echo RESTORE CONTROLFILE FROM '%BACKUP_DIR%/ctl*.bkp'; >> D:\oracle\temp\restore_ctl.rman
echo RELEASE CHANNEL ch1; >> D:\oracle\temp\restore_ctl.rman
echo } >> D:\oracle\temp\restore_ctl.rman
echo EXIT; >> D:\oracle\temp\restore_ctl.rman
rman target / cmdfile=D:\oracle\temp\restore_ctl.rman
if %errorlevel% neq 0 (
echo ERROR: Control file restore failed
exit /b 1
)
REM Step 4: Mount database
echo.
echo [STEP 4/8] Mounting database...
echo ALTER DATABASE MOUNT; > D:\oracle\temp\mount.sql
echo EXIT; >> D:\oracle\temp\mount.sql
sqlplus / as sysdba @D:\oracle\temp\mount.sql
REM Step 5: Catalog all backups
echo.
echo [STEP 5/8] Cataloging backups from mount point...
echo CATALOG START WITH '%BACKUP_DIR%/' NOPROMPT; > D:\oracle\temp\catalog.rman
echo LIST BACKUP SUMMARY; >> D:\oracle\temp\catalog.rman
echo EXIT; >> D:\oracle\temp\catalog.rman
rman target / cmdfile=D:\oracle\temp\catalog.rman
REM Step 6: Restore and recover database
echo.
echo [STEP 6/8] Restoring FULL + latest CUMULATIVE...
echo RUN { > D:\oracle\temp\restore_db.rman
echo ALLOCATE CHANNEL ch1 DEVICE TYPE DISK; >> D:\oracle\temp\restore_db.rman
echo ALLOCATE CHANNEL ch2 DEVICE TYPE DISK; >> D:\oracle\temp\restore_db.rman
echo. >> D:\oracle\temp\restore_db.rman
echo # RMAN will automatically select: >> D:\oracle\temp\restore_db.rman
echo # 1. Level 0 (FULL from 02:30) >> D:\oracle\temp\restore_db.rman
echo # 2. Latest Level 1 CUMULATIVE (from 13:00 or 18:00) >> D:\oracle\temp\restore_db.rman
echo. >> D:\oracle\temp\restore_db.rman
echo RESTORE DATABASE; >> D:\oracle\temp\restore_db.rman
echo RECOVER DATABASE; >> D:\oracle\temp\restore_db.rman
echo. >> D:\oracle\temp\restore_db.rman
echo RELEASE CHANNEL ch1; >> D:\oracle\temp\restore_db.rman
echo RELEASE CHANNEL ch2; >> D:\oracle\temp\restore_db.rman
echo } >> D:\oracle\temp\restore_db.rman
echo EXIT; >> D:\oracle\temp\restore_db.rman
rman target / cmdfile=D:\oracle\temp\restore_db.rman
if %errorlevel% neq 0 (
echo ERROR: Database restore/recovery failed
exit /b 1
)
REM Step 7: Open database with RESETLOGS
echo.
echo [STEP 7/8] Opening database with RESETLOGS...
echo ALTER DATABASE OPEN RESETLOGS; > D:\oracle\temp\open.sql
echo EXIT; >> D:\oracle\temp\open.sql
sqlplus / as sysdba @D:\oracle\temp\open.sql
REM Step 8: Create TEMP and verify
echo.
echo [STEP 8/8] Creating TEMP tablespace and verifying...
echo ALTER TABLESPACE TEMP ADD TEMPFILE 'C:\Users\oracle\oradata\ROA\temp01.dbf' > D:\oracle\temp\verify.sql
echo SIZE 567M REUSE AUTOEXTEND ON NEXT 640K MAXSIZE 32767M; >> D:\oracle\temp\verify.sql
echo. >> D:\oracle\temp\verify.sql
echo SET LINESIZE 200 >> D:\oracle\temp\verify.sql
echo SELECT NAME, OPEN_MODE FROM V$DATABASE; >> D:\oracle\temp\verify.sql
echo SELECT TABLESPACE_NAME, STATUS FROM DBA_TABLESPACES ORDER BY 1; >> D:\oracle\temp\verify.sql
echo EXIT; >> D:\oracle\temp\verify.sql
sqlplus / as sysdba @D:\oracle\temp\verify.sql
echo.
echo ============================================================================
echo DR RESTORE COMPLETED SUCCESSFULLY!
echo ============================================================================
echo Database is OPEN and ready
echo.
endlocal
exit /b 0
```
---
### PHASE 6: Weekly Test Procedure (1 hour first time, 30 min ongoing)
**Objective:** Document weekly test procedure using new cumulative backup strategy
#### 6.1 Test procedure (run on Saturday morning)
```bash
# On Linux workstation or any machine with SSH to Proxmox
# Step 1: Verify latest backups on host (5 min)
ssh root@10.0.20.202 "ls -lth /mnt/pve/oracle-backups/ROA/autobackup/*.bkp | head -10"
# Expected to see:
# - FULL backup from this morning (02:30)
# - CUMULATIVE from yesterday 18:00
# - CUMULATIVE from yesterday 13:00
# - Older files...
# Step 2: Start DR VM (2 min)
ssh root@10.0.20.202 "qm start 109"
# Wait for Windows boot
sleep 180
# Verify VM is up
ping -c 3 10.0.20.37
# Step 3: Verify mount point in VM (2 min)
ssh -p 22122 romfast@10.0.20.37 "Get-ChildItem E:\oracle-backups\ROA\autobackup\*.bkp | Measure-Object"
# Should show ~10-15 backup files
# Step 4: Run restore (15-20 min)
ssh -p 22122 romfast@10.0.20.37 "D:\oracle\scripts\rman_restore_cumulative.cmd"
# Monitor restore progress
ssh -p 22122 romfast@10.0.20.37 "Get-Content D:\oracle\logs\restore_cumulative_*.log -Wait"
# Step 5: Verify database (5 min)
ssh -p 22122 romfast@10.0.20.37 "cmd /c 'set ORACLE_HOME=C:\Users\Administrator\Downloads\WINDOWS.X64_193000_db_home&& set ORACLE_SID=ROA&& set PATH=%ORACLE_HOME%\bin;%PATH%&& sqlplus -s / as sysdba @D:\oracle\scripts\verify_restore.sql'"
# Step 6: Shutdown VM (2 min)
ssh -p 22122 romfast@10.0.20.37 "shutdown /s /t 60"
# Or force from Proxmox:
ssh root@10.0.20.202 "qm shutdown 109"
# Verify VM stopped
ssh root@10.0.20.202 "qm status 109"
```
#### 6.2 Create automated test script
```bash
#!/bin/bash
# File: /root/scripts/test_oracle_dr.sh
# Run on Linux workstation or Proxmox host
LOG_FILE="/root/scripts/logs/dr_test_$(date +%Y%m%d_%H%M%S).log"
PVEHOST="10.0.20.202"
DRVM="10.0.20.37"
DRVM_PORT="22122"
log() {
echo "[$(date '+%Y-%m-%d %H:%M:%S')] $1" | tee -a "$LOG_FILE"
}
log "==================================================================="
log "Oracle DR Weekly Test - Started"
log "==================================================================="
# Step 1: Check backups on host
log "Step 1: Verifying backups on Proxmox host..."
ssh root@$PVEHOST "ls -lh /mnt/pve/oracle-backups/ROA/autobackup/*.bkp | wc -l" | tee -a "$LOG_FILE"
# Step 2: Start DR VM
log "Step 2: Starting DR VM 109..."
ssh root@$PVEHOST "qm start 109"
sleep 180
# Step 3: Verify mount
log "Step 3: Verifying mount point in VM..."
ssh -p $DRVM_PORT romfast@$DRVM "powershell -Command 'Get-ChildItem E:\oracle-backups\ROA\autobackup\*.bkp | Measure-Object'" | tee -a "$LOG_FILE"
# Step 4: Run restore
log "Step 4: Running RMAN restore (this will take 15-20 minutes)..."
ssh -p $DRVM_PORT romfast@$DRVM "D:\oracle\scripts\rman_restore_cumulative.cmd" | tee -a "$LOG_FILE"
if [ $? -eq 0 ]; then
log "Restore completed successfully"
else
log "ERROR: Restore failed"
exit 1
fi
# Step 5: Verify database
log "Step 5: Verifying database..."
ssh -p $DRVM_PORT romfast@$DRVM "cmd /c 'set ORACLE_HOME=C:\Users\Administrator\Downloads\WINDOWS.X64_193000_db_home&& sqlplus -s / as sysdba @D:\oracle\scripts\verify_restore.sql'" | tee -a "$LOG_FILE"
# Step 6: Shutdown VM
log "Step 6: Shutting down DR VM..."
ssh root@$PVEHOST "qm shutdown 109"
sleep 60
log "==================================================================="
log "Oracle DR Weekly Test - Completed Successfully"
log "==================================================================="
```
---
## 📊 EXPECTED RESULTS
### Backup Schedule (after implementation)
| Time | Type | Size | Retention | Transfer to |
|------|------|------|-----------|-------------|
| 02:30 | Level 0 FULL | 6-7 GB | 2 days | Proxmox host |
| 13:00 | Level 1 CUMULATIVE | 150-300 MB | 2 days | Proxmox host |
| 18:00 | Level 1 CUMULATIVE | 200-400 MB | 2 days | Proxmox host |
### RPO Analysis
| Disaster Time | Backup Used | Data Loss |
|---------------|-------------|-----------|
| 03:00-13:00 | FULL (02:30) | Max 10.5 hours |
| 13:00-18:00 | FULL + CUMULATIVE (13:00) | Max 5 hours |
| 18:00-02:30 | FULL + CUMULATIVE (18:00) | Max 8.5 hours |
| **Average RPO** | | **~4-5 hours** |
### Storage Requirements
- **Proxmox host:** ~15 GB (2 days × 7.5 GB/day)
- **VM 109 disk:** 500 GB (unchanged, backups not stored in VM)
- **Daily transfer:** ~7.5 GB (FULL + 2× CUMULATIVE)
### RTO (unchanged)
- Start VM: 2 minutes
- Restore FULL + CUMULATIVE: 12-15 minutes
- Verify & open: 1 minute
- **Total: ~15-18 minutes**
---
## 🚨 ROLLBACK PLAN
If any issues during implementation:
### Rollback Step 1: Restaurează scripturile originale
```powershell
# On PRIMARY
cd D:\rman_backup
Copy-Item rman_backup_incremental_ORIGINAL.txt rman_backup_incremental.txt -Force
Copy-Item transfer_incremental_ORIGINAL.ps1 transfer_incremental.ps1 -Force
Copy-Item transfer_to_dr_ORIGINAL.ps1 transfer_to_dr.ps1 -Force
# Verifică că s-au restaurat
Get-Content rman_backup_incremental.txt | Select-String "CUMULATIVE"
# Nu ar trebui să găsească nimic dacă restaurarea a reușit
```
### Rollback Step 2: Restaurează task-urile originale
```powershell
# Șterge task-ul nou de la 18:00
Unregister-ScheduledTask -TaskName "Oracle RMAN Incremental Backup 1800" -Confirm:$false
# Restaurează task-ul de la 13:00 înapoi la 14:00
$taskName = "Oracle RMAN Incremental Backup" # Numele task-ului tău
$trigger = New-ScheduledTaskTrigger -Daily -At "14:00"
Set-ScheduledTask -TaskName $taskName -Trigger $trigger
# SAU restaurează din backup XML
Register-ScheduledTask -Xml (Get-Content "D:\rman_backup\backup_tasks\Oracle RMAN Incremental Backup.xml") -Force
```
---
## ✅ VALIDATION CHECKLIST
After completing implementation:
- [ ] Proxmox host directory created: `/mnt/pve/oracle-backups/ROA/autobackup`
- [ ] VM 109 mount point configured and tested (E:\ visible in Windows)
- [ ] RMAN script modified to CUMULATIVE (keyword added)
- [ ] New transfer script created (`02_transfer_to_pveelite_host.ps1`)
- [ ] SSH key for Proxmox host created and tested
- [ ] Scheduled task created for 13:00 CUMULATIVE backup
- [ ] Scheduled task created for 18:00 CUMULATIVE backup
- [ ] Existing 02:30 FULL task updated to use new transfer script
- [ ] Manual test of CUMULATIVE backup successful
- [ ] Manual test of backup transfer to host successful
- [ ] DR restore script updated (`rman_restore_cumulative.cmd`)
- [ ] Full end-to-end restore test successful
- [ ] Weekly test script created and tested
- [ ] Documentation updated (STATUS and IMPLEMENTATION_PLAN docs)
---
## 📞 NEXT SESSION HANDOFF
**Status:** PLAN COMPLETE - Ready for implementation
**Estimated Implementation Time:** 2-3 hours
**Recommended Schedule:** Saturday morning (low activity time)
**Context for next session:**
1. Primary server: 10.0.20.36 (Windows, Oracle 19c, database ROA)
2. DR VM: 109 on pveelite (10.0.20.37, currently working with FULL-only restore)
3. Proxmox host: pveelite (10.0.20.202)
4. Goal: Implement CUMULATIVE incremental backups (13:00 + 18:00) for better RPO (4-5 hours vs 24 hours)
5. Key change: Backups stored on Proxmox host, mounted in VM 109 when needed
**Start implementation with:**
```bash
# Phase 1 - Proxmox host storage setup (15 min)
ssh root@10.0.20.202
mkdir -p /mnt/pve/oracle-backups/ROA/autobackup
# ... follow Phase 1 steps
```
**IMPORTANT - Backup manual înainte de modificări:**
Fă backup MANUAL la fișierele pe care le vei modifica:
```powershell
# Pe PRIMARY, copiază fișierele EXISTENTE înainte de modificare:
cd D:\rman_backup
Copy-Item rman_backup_incremental.txt rman_backup_incremental_ORIGINAL.txt
Copy-Item transfer_incremental.ps1 transfer_incremental_ORIGINAL.ps1
Copy-Item transfer_to_dr.ps1 transfer_to_dr_ORIGINAL.ps1
# Exportă task-urile
Get-ScheduledTask | Where-Object {$_.TaskName -like "*Oracle*"} | ForEach-Object {
Export-ScheduledTask -TaskName $_.TaskName | Out-File "D:\rman_backup\backup_tasks\$($_.TaskName).xml"
}
```
**Dacă ceva nu merge, restaurezi din aceste copii!**
---
**Generated:** 2025-10-09
**Version:** 1.0
**Author:** Claude Code (Sonnet 4.5)
**Status:** ✅ PLAN COMPLETE - Ready for next session implementation