ORA-15025: Could Not Open Disk - Fix ASM Disk Access Errors

ORA-15025: Could Not Open Disk

Error Overview

Error Text: ORA-15025: could not open disk "string"

The ORA-15025 error occurs when Oracle ASM cannot open a disk device during diskgroup mount, rebalance, or disk discovery operations. ASM requires direct read/write access to the underlying block devices, and any interruption in that access — from a permissions change, a multipath failover, a missing udev rule, or a stale ASM_DISKSTRING — will trigger this error. The diskgroup may fail to mount entirely, or a running rebalance may abort mid-operation.

Common Causes

1. Incorrect Device File Permissions

/dev/oracleasm/disks/ device files owned by root instead of oracle/grid
Block device permissions on /dev/sd* or /dev/mapper/* set to 640 instead of 660
ORACLE_BASE or grid home ownership changed after patching
SELinux or AppArmor blocking the ASM instance from opening the device

2. ASM_DISKSTRING Misconfiguration

Discovery string does not match the actual device path after storage reconfiguration
Blank or missing ASM_DISKSTRING forcing a full OS scan that cannot locate devices
Mismatch between AFD (ASM Filter Driver) discovery string and legacy ASMLib paths
Multiple discovery strings defined but one path has become stale

3. Multipath Device Mapping Issues

Device mapper paths renamed after a SAN zoning change
Multipath daemon (multipathd) not running, exposing raw paths instead of /dev/mapper/ aliases
DM-Multipath blacklist misconfigured, causing device churn
WWID change on a replaced LUN leaving the old alias orphaned

4. ASMLib / udev Rule Problems

oracleasm scandisks not run after adding or replacing LUNs
Missing or corrupted ASMLib kernel module (oracleasm module not loaded)
udev rule for Oracle ASM (99-oracle-asmdevices.rules) missing after OS upgrade
/etc/udev/rules.d/ rules not applied because udevadm trigger was not executed

5. Disk Hardware or SAN Failure

LUN removed or masked on the SAN array without first dropping the disk from ASM
Disk returning I/O errors detected before ORA-15025 appears in the alert log
HBA firmware issue causing transient path loss
Storage path failover incomplete, leaving one path unavailable at mount time

6. Grid Infrastructure Owner Mismatch

ASM instance running as grid user but devices owned by oracle
Group membership of asmadmin / asmdba not correctly set
/dev/oracleasm/ directory recreated by a script with wrong ownership

Diagnostic Queries

Identify the Failed Disk

-- List all ASM disks and their current status
SELECT
  dg.name          AS diskgroup,
  d.disk_number,
  d.name           AS disk_name,
  d.path,
  d.header_status,
  d.mount_status,
  d.mode_status,
  d.state,
  ROUND(d.total_mb / 1024, 2)  AS total_gb,
  ROUND(d.free_mb  / 1024, 2)  AS free_gb,
  d.os_mb
FROM v$asm_disk d
LEFT JOIN v$asm_diskgroup dg ON d.group_number = dg.group_number
ORDER BY dg.name, d.disk_number;

-- Find disks with MISSING or error states
SELECT
  disk_number,
  name,
  path,
  header_status,
  mount_status,
  mode_status,
  state,
  reads,
  writes,
  read_errs,
  write_errs
FROM v$asm_disk
WHERE mount_status != 'CACHED'
   OR mode_status  != 'ONLINE'
   OR state        != 'NORMAL'
ORDER BY disk_number;

Check ASM Discovery String

-- View current ASM_DISKSTRING parameter
SHOW PARAMETER asm_diskstring;

-- View all ASM parameters from within the ASM instance
SELECT name, value, description
FROM v$parameter
WHERE name IN ('asm_diskstring', 'asm_diskgroups', 'asm_power_limit')
ORDER BY name;

-- Verify which disks were discovered with the current string
SELECT
  path,
  header_status,
  mount_status,
  name,
  failgroup,
  ROUND(total_mb/1024, 2) AS total_gb
FROM v$asm_disk
WHERE header_status = 'MEMBER'
ORDER BY path;

Check Diskgroup Mount Status

-- Overall diskgroup health
SELECT
  group_number,
  name,
  state,
  type,
  total_mb,
  free_mb,
  ROUND((1 - free_mb/NULLIF(total_mb,0))*100, 1) AS pct_used,
  usable_file_mb,
  offline_disks,
  compatibility,
  database_compatibility
FROM v$asm_diskgroup
ORDER BY name;

-- Check recent ASM alert log entries via external table (if configured)
-- Review $ORACLE_BASE/diag/asm/+asm/<instance>/trace/alert_+ASM*.log
-- Look for: ORA-15025, ORA-15040, ORA-15042, kfed, kfdhdb

-- Check for disks in OFFLINE state inside diskgroup
SELECT
  dg.name          AS diskgroup,
  d.name           AS disk_name,
  d.path,
  d.mode_status,
  d.state,
  d.repair_timer
FROM v$asm_disk d
JOIN v$asm_diskgroup dg ON d.group_number = dg.group_number
WHERE d.mode_status = 'OFFLINE'
   OR d.state IN ('DROPPING', 'HUNG');

Check Active ASM Operations

-- Monitor ongoing rebalance or mount operations
SELECT
  group_number,
  operation,
  state,
  power,
  actual,
  sofar,
  est_work,
  est_rate,
  est_minutes
FROM v$asm_operation
ORDER BY group_number;

Step-by-Step Resolution

Step 1: Check the Alert Log for the Exact Device Path

The ORA-15025 message names the specific disk path that failed. Locate the alert log:

# Find the ASM alert log
find $ORACLE_BASE/diag/asm -name "alert_+ASM*.log" | head -5

# Tail the log for recent errors
tail -200 $ORACLE_BASE/diag/asm/+asm/+ASM1/trace/alert_+ASM1.log | grep -E "ORA-|15025|15040|15042"

Step 2: Verify Device Existence and Permissions

# List ASMLib managed disks
/usr/sbin/oracleasm listdisks

# Scan for new or changed disks
/usr/sbin/oracleasm scandisks

# Query disk attributes
/usr/sbin/oracleasm querydisk -d DATA01

# Check raw block device permissions (non-ASMLib / udev setup)
ls -la /dev/sd* /dev/mapper/asm* 2>/dev/null

# Check ownership of /dev/oracleasm/ disk entries
ls -la /dev/oracleasm/disks/

Expected ownership for a grid/ASMLib setup:

crw-rw---- 1 grid asmadmin ... /dev/oracleasm/disks/DATA01

Step 3: Fix Device Permissions

# For raw block devices — set correct ownership and mode
chown grid:asmadmin /dev/sdc /dev/sdd
chmod 660 /dev/sdc /dev/sdd

# Reload udev rules if using udev-based ASM disk management
udevadm control --reload-rules
udevadm trigger --type=devices --action=change

# For ASMLib — reinitialize if needed
/usr/sbin/oracleasm configure -i
# Answer prompts: ASM user = grid, ASM group = asmadmin, etc.
/usr/sbin/oracleasm init
/usr/sbin/oracleasm scandisks

Step 4: Verify and Correct ASM_DISKSTRING

-- Connect to ASM instance as SYSASM
-- sqlplus / as sysasm

-- Update the discovery string to match device paths
ALTER SYSTEM SET asm_diskstring = '/dev/oracleasm/disks/*', '/dev/mapper/asm*' SCOPE=BOTH;

-- After updating, force rediscovery by attempting to mount
ALTER DISKGROUP data MOUNT;

Step 5: Check Multipath Configuration

# Verify multipathd is running
systemctl status multipathd

# List multipath devices and confirm aliases
multipath -ll

# Show which device corresponds to which WWID
multipathd show paths
multipathd show maps

# Reload multipath configuration if aliases changed
systemctl reload multipathd

Step 6: Mount the Diskgroup

-- After resolving the device access issue, mount the diskgroup
ALTER DISKGROUP data MOUNT;

-- If MOUNT fails, try FORCE mount (use with caution — requires all disks visible)
ALTER DISKGROUP data MOUNT FORCE;

-- Check mount result
SELECT name, state FROM v$asm_diskgroup;

Step 7: Mark Offline Disks Online (if disk temporarily disappeared)

-- If the disk came back but is OFFLINE
ALTER DISKGROUP data ONLINE DISK DATA_0003;

-- Or online all disks in the diskgroup
ALTER DISKGROUP data ONLINE ALL;

Prevention Strategies

1. Persistent udev Rules

# Example udev rule for ASM disks — /etc/udev/rules.d/99-oracle-asmdevices.rules
# Identify disks by WWID to survive device rename

KERNEL=="sd*", ENV{ID_WWN}=="0x600508b1001c123456789abcdef01234", \
  SYMLINK+="oracleasm/disks/DATA01", OWNER="grid", GROUP="asmadmin", MODE="0660"

# Apply after creating or editing the file
udevadm control --reload-rules && udevadm trigger

2. Monitor Disk Availability with a Scheduled Query

-- Create a procedure to alert on disk access issues
CREATE OR REPLACE PROCEDURE check_asm_disk_access AS
  v_count NUMBER;
BEGIN
  SELECT COUNT(*)
  INTO   v_count
  FROM   v$asm_disk
  WHERE  mount_status != 'CACHED'
     OR  mode_status  != 'ONLINE'
     OR  state        != 'NORMAL';

  IF v_count > 0 THEN
    -- Replace with your preferred alerting mechanism
    RAISE_APPLICATION_ERROR(-20001,
      'ASM disk access issue detected: ' || v_count || ' disk(s) not healthy.');
  END IF;
END;
/

-- Schedule the check every 5 minutes
BEGIN
  DBMS_SCHEDULER.CREATE_JOB(
    job_name        => 'CHECK_ASM_DISK_ACCESS',
    job_type        => 'STORED_PROCEDURE',
    job_action      => 'check_asm_disk_access',
    repeat_interval => 'FREQ=MINUTELY;INTERVAL=5',
    enabled         => TRUE
  );
END;
/

3. ASM Disk Compatibility Checks

-- Ensure diskgroup and database compatibility levels are current
SELECT
  name,
  compatibility,
  database_compatibility
FROM v$asm_diskgroup
ORDER BY name;

-- Check for disks with errors accumulating (early warning)
SELECT
  name, path, read_errs, write_errs,
  ROUND(read_errs + write_errs) AS total_errors
FROM v$asm_disk
WHERE read_errs + write_errs > 0
ORDER BY total_errors DESC;

4. Best Practices

Always run oracleasm scandisks after any SAN LUN presentation or replacement
Use persistent device names (WWID-based udev aliases or ASMLib) instead of /dev/sd* paths which can change across reboots
Test disk visibility from all cluster nodes before mounting a diskgroup
Document the ASM_DISKSTRING value and store it in your runbook; an empty string triggers a full OS scan and can be very slow on large servers
Use Oracle’s kfed utility to verify disk headers before attempting forced mounts

ORA-15001 — Diskgroup does not exist or is not mounted
ORA-15017 — Diskgroup cannot be mounted
ORA-15040 — Diskgroup is incomplete (not all disks found)
ORA-15032 — Not all alterations performed (partial disk operations)
ORA-15041 — Diskgroup space is not sufficient
ORA-15042 — ASM disk is missing from diskgroup
ORA-15063 — ASM insufficient number of disks in diskgroup

Emergency Response

Quick Checks

Confirm the disk path from the alert log

grep "15025\|could not open disk" \
  $ORACLE_BASE/diag/asm/+asm/+ASM1/trace/alert_+ASM1.log | tail -20

Re-scan ASMLib disks immediately

/usr/sbin/oracleasm scandisks && /usr/sbin/oracleasm listdisks

Attempt diskgroup mount after fixing permissions

ALTER DISKGROUP data MOUNT;
SELECT name, state FROM v$asm_diskgroup;

Post-Resolution Cleanup

-- Verify all disks are healthy after recovery
SELECT name, path, header_status, mount_status, mode_status, state
FROM v$asm_disk
ORDER BY name;

-- Check for any orphaned OFFLINE disks that may have repair timers running
SELECT name, path, mode_status, state, repair_timer
FROM v$asm_disk
WHERE repair_timer > 0;

-- If repair timer is running and disk is confirmed unavailable, drop it
-- ALTER DISKGROUP data DROP DISK DATA_0003 FORCE;
-- WARNING: Only run the above if ASM has enough redundancy to survive the drop