Shutdown/Reboot hangs Indefinitely - Ubuntu 16.04

Question

After upgrading my laptop (Dell Inspiron 5447) to Ubuntu 16.04 I saw that I can't shut down or restart after a few hours using the system. It hangs in Plymouth and stays that way until I force the shutdown by pressing the On/Off button.

As this is a common problem (there are tons of questions on the Internet) I tried to fix it, but none of my attempts were effective. Already tried:

1. Boot with acpi=force, acpi=noirq and pci=noacpi;
2. Disabling Swap;
3. Run 'sync' before reboot or shutdown;
4. Use another kernel version (4.7, 4.8, 4.9 and 4.10 from mainline);
5. Disabling USB 3.0 in BIOS setting;
6. Disabling TLP;

The most interesting thing is:

1. I can shutdown/reboot normally using 14.04 and the kernel from Xenial HWE;
2. The system is shutting down/rebooting normally if it doesn't have a considerable uptime (< 2 hours for example);

My Machine Specs:

Dell Inspiron 5447 - BIOS A10 Intel Core i5-4210 Processor 8 GB RAM 480 GB SSD Sandisk Hybrid Graphics (i915/amdgpu - Radeon R7 M265)

Does anyone know how to solve this problem?

Answer 1

I got exactly the same problem on Ubuntu 16.04 LTS. After installing amdgpu-pro drivers ver. 492261 system hangs at reboot. Last message printed is: Restarting System

I solved the issue by adding reboot=pci to the list of parameters passed to the kernel:

• Edit /etc/default/grub
• Add reboot=pci to the end of GRUB_CMDLINE_LINUX_DEFAULT variable
• run as root: update-grub
• reboot (system will hang, turn off/on manually)
• reboot (this time it will work)

I tried all the suggestions in the other answers and they didn't solve the issue for me.

Answer 2

The problem occurred when the PCI device controlled by the mei_me module was with Runtime Power Management activated.

I got it fixed by adding mei_me module to the RUNTIME_PM_DRIVER_BLACKLIST configuration of TLP.

So, If you have this same problem. It's probably some PCI device with Power Management issues..

Answer 3

I had a similar issue running Ubuntu Server 16.04LTS on Vmware. On shutdown it would "hang" for some time and then proceed to shutdown or reboot. Turns out my host was not configured to do NTP sync, while ubuntu does by default. The host clock was out. As soon as I let the host sync too problem disappeared.

Answer 4

I had the same issue after an update and AMD driver update. Try the following:

sudo ubuntu-drivers autoinstall

Answer 5

In my case the linux boot parameter "noefi" solved the problem on my thin client wyse Z90D8 with librelex.

Answer 6

I had the same issue in 18.04 with an MSI apache laptop (GE62VR GRF Apache Pro, GTX1060, intel core 7 H series).

My solution (I have a dual boot with win10)

For testing: When the computer starts and you have the boot menu press the e key.

I observed that the apci=off option was present. I removed this. Had a successful boot (note this may be a problem for some other hardware configurations that have a tendency to hang when the acpi=off is removed).

Press f10 to boot. This will only be valid for this boot and is not saved or permanent.

After login I then rebooted the system and it worked fine. No hang just went through a standard reboot cycle (pew!).

To make it permanent in grub:

In terminal type sudo nano /etc/default/grub

Locate the line: GRUB_CMDLINE_LINUX=“acpi=off” and comment it out as follows:

#GRUB_CMDLINE_LINUX=“acpi=off”

I suppose you could also delete the line but I like to track my changes.

Save and close this file.

Then update in terminal with:

sudo update-grub

Next time you start or reboot the system will use these parameters at boot time without you having to change it on the fly.

I hope this is useful to someone out there.