WARNING you have Transparent Huge Pages (THP) support enabled Redis Origin: Data OSI layer 7 · Application Last updated 5 Nov 2026

Redis: # WARNING you have Transparent Huge Pages (THP) support enabled in your kernel. This will create latency and memory usage issues with Redis

THP is enabled, which can cause latency spikes and high memory usage with Redis.

Exact message you may see WARNING you have Transparent Huge Pages (THP) support enabled in your kernelThis will create latency and memory usage issues with RedisTo fix this issue run the command 'echo madvise > /sys/kernel/mm/transparent_hugepage/enabled' as rootTHP support enabled Redis latency
Explore the map of this error Full error map →
Seen on: Redis Linux

Meaning

Disable it: echo never | sudo tee /sys/kernel/mm/transparent_hugepage/enabled, persist via a systemd unit or tuned, and restart Redis. On Kubernetes/cloud hosts set it at the node level. It is a startup warning, but THP is a top cause of fork latency and RSS bloat.

Common causes

  • Default THP=always/madvise
  • Cloud image defaults
  • Containers sharing the host kernel setting
  • Not persisted across reboots
  • Latency issues during BGSAVE

⚡ Quick fix

  1. echo never > /sys/kernel/mm/transparent_hugepage/enabled (persist at boot)
  2. Restart Redis
  3. Verify the setting after reboot

Detailed fix by platform

Linux

  1. cat /sys/kernel/mm/transparent_hugepage/enabled

How to diagnose

  1. Setting — never?
  2. Persist — Boot unit?
  3. Latency — After disable?

🔧 Still not fixed?

Many errors look alike. If the steps above didn’t solve it, one of these is probably what you’re facing:

Where this error sits

Origin: Data. Errors that are often confused with or caused by this one (click to open):

BullMQ: Missing lock for job / job stalled more than allowable limit — BullMQ: Error: Missing lock for job 42. moveToFinished / job stalled more than allowable limit (Redis)BullMQ: Missing lock …BUSY Redis is busy running a script — Redis: (error) BUSY Redis is busy running a script. You can only call SCRIPT KILL or SHUTDOWN NOSAVE (Redis)BUSY Redis is busy ru…Celery with Redis: tasks run twice (visibility_timeout) — Celery: Task executed twice / Restoring unacked messages — Redis broker visibility_timeout shorter than task ETA (Redis)Celery with Redis: ta…Client output buffer limit reached (scheduled to be closed) — Redis: Client id=12 addr=10.0.0.5:51234 scheduled to be closed ASAP for overcoming of output buffer limits (Redis)Client output buffer …mem_fragmentation_ratio high / memory not released — Redis: INFO memory — mem_fragmentation_ratio:2.8 (RSS much larger than used_memory) (Redis)mem_fragmentation_rat…Redis::CannotConnectError (Sidekiq/Rails) — Redis::CannotConnectError: Error connecting to Redis on localhost:6379 (Errno::ECONNREFUSED) (Ruby / Rails)Redis::CannotConnectE…WARNING: /sys/kernel/mm/transparent_hugepage/enabled is 'always' — MongoDB startup: WARNING: /sys/kernel/mm/transparent_hugepage/enabled is 'always'. We suggest setting it to 'never' (MongoDB)WARNING: /sys/kernel/…WARNING: The TCP backlog setting of 511 cannot be enforced — Redis: # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128 (Redis)WARNING: The TCP back…WARNING y…

Explore the map of this error Full error map →

🧠 Still stuck? Analyze your error

Paste the full message, response headers or stack trace — we'll detect the platform and point to the most likely cause.

💬 What worked for others

No tips yet — did something else fix this for you?

Share what worked for you

Tips are checked by an editor before they appear. Links are removed.