tcp_ack+633 Recurring Error

[Platypus69]

Hi all,

I have a recurring error on a RN428 running in the following configuration:

• RN428
• Firmware 6.10.8
• 6 x 12GB IronWolf Seagate HDDs (ST12000VN0007)
• RAID 6
• Volume is full, so like 10-100GB free

So every so often NAS hangs with tcp_ack+633 error.

I have to reboot the NAS, and then most of the time it starts doing a re-sync which takes 35+ hours.

Otherwise, everything is fine.

From memory I do no recall ever seeing this error with earlier firmware, so am wondering whether there is a bug in version 6.10.x

Not sure what is triggering this... So whether it only happens during a scrub or some other CPU intensive task.

So:

1. What does this error mean? It seems to imply TCP stack???
2. Any thoughts on what is causing this?
   1. Which logs should I look at?
      1. Should I post any logs here
3. How can I prevent this?
   1. Switch to other LAN port
   2. Downgrade firmware
   3. Ensure volume has a specific percentage of free space

TIA as always

[StephenB]

---

@Platypus69 wrote:

• Volume is full, so like 10-100GB free

---

Very risky, as you can lose all your data if the volume gets too full.  With BTRFS you want at least 10% free space.

---

@Platypus69 wrote:

So every so often NAS hangs with tcp_ack+633 error.

1. What does this error mean? It seems to imply TCP stack???

---

The NAS crashed at that memory location.  It might not be the TCP stack at all though.  

The logs might give more clues, hard to say.  You shouldn't post them publicly though.  You can start by looking in system.log, kernel.log, and systemd-journal.log around the time of the failure.