Sankofa/docs/ceph/PERC_CONTROLLER_CHECK.md

# Checking PERC Controller for 250GB Drives

**Date**: 2025-12-13  
**Issue**: 6x 250GB drives not visible to OS, need to check PERC controller

---

## Current Status

- **OS Detected**: Only 2x 300GB drives (sda, sdb)
- **Expected**: 6x 250GB drives should be visible
- **Likely Cause**: Drives not configured in PERC controller or in RAID array

---

## Check PERC Controller

### Option 1: Install Dell OpenManage Tools

```bash
# Install Dell OpenManage Server Administrator
apt-get update
apt-get install -y srvadmin-all

# Start services
systemctl start dsm_sa_datamgrd
systemctl start dsm_sa_snmpd
systemctl start dsm_sa_eventmgrd

# Check physical disks
omreport storage pdisk controller=0

# Check virtual disks (RAID arrays)
omreport storage vdisk controller=0
```

### Option 2: Use PERC CLI (if available)

```bash
# Check if perccli is installed
which perccli

# If not installed, download from Dell
# Then check physical disks
perccli /c0 show

# Check virtual disks
perccli /c0/vall show
```

### Option 3: Check via /proc or /sys

```bash
# Check for PERC controller info
cat /proc/scsi/scsi

# Check block devices
ls -la /dev/sd*

# Check for additional controllers
lspci | grep -i raid
lspci | grep -i perc
lspci | grep -i lsi
```

### Option 4: Check dmesg for disk detection

```bash
# Check kernel messages for disk detection
dmesg | grep -i "sd[a-z]"
dmesg | grep -i "disk"
dmesg | grep -i "perc"
dmesg | grep -i "raid"

# Check recent disk events
journalctl -k | grep -i disk | tail -50
```

---

## What to Look For

### If Drives Are in PERC Controller

You should see:
- 8 physical disks total
- 2x 300GB drives (already configured)
- 6x 250GB drives (may be unconfigured or in RAID)

### If Drives Are Unconfigured

They may show as:
- "Unconfigured Good" or "Ready"
- Not assigned to any virtual disk
- Available for configuration

### If Drives Are in RAID

They may be:
- Part of a RAID array (virtual disk)
- Not exposed as individual disks to OS
- Need to be removed from RAID to use individually

---

## Next Steps Based on Findings

### Scenario 1: Drives Detected in PERC but Unconfigured

**Action**: Configure drives as individual disks (non-RAID) or create separate virtual disks

```bash
# Using perccli (example - adjust for your setup)
# Create individual virtual disks (one per physical disk)
# This exposes each drive to the OS

# Or configure PERC in HBA mode (if supported)
# This passes drives directly to OS without RAID
```

### Scenario 2: Drives Not Detected in PERC

**Possible causes**:
- Drives not physically installed
- Drive backplane issue
- Controller issue

**Action**: 
1. Check physical installation
2. Check drive bay LEDs
3. Check PERC controller status
4. May need to reseat drives

### Scenario 3: Drives in RAID Array

**Action**: 
1. Check RAID configuration
2. Decide if RAID array can be broken up
3. May need to backup data and reconfigure
4. Create individual virtual disks for Ceph OSDs

---

## Recommended Configuration for Ceph

For Ceph OSDs, you want:
- **Individual disks** (not in RAID)
- **Direct access** to each physical disk
- **PERC in HBA mode** (if supported) or individual virtual disks

**Why**: Ceph handles redundancy, so RAID is not needed and can reduce performance.

---

## Quick Check Commands

Run these on R630-01:

```bash
# 1. Check for PERC controller
lspci | grep -iE "raid|perc|lsi|megaraid"

# 2. Check kernel messages
dmesg | grep -iE "sd[a-z]|disk|perc" | tail -20

# 3. Check /proc for SCSI devices
cat /proc/scsi/scsi

# 4. Check for additional block devices
lsblk
fdisk -l

# 5. Check if OpenManage is available
which omreport
which perccli
```

---

## Summary

The 6x 250GB drives are likely:
1. **In PERC controller** but not configured/exposed
2. **In a RAID array** that needs to be reconfigured
3. **Not physically installed** (less likely if user confirmed they exist)

**Next Step**: Check PERC controller status to see where the drives are.

---

**Last Updated**: 2025-12-13