You are not logged in.
Hello, I've got my laptop (Razer Blade 15 2021) setup to use hibernation. Most of the time it will hibernate successfully, until eventually it doesn't.
magos% sudo systemctl hibernate
Call to Hibernate failed: Not enough swap space for hibernationMy system has 32GB of ram with a 32GB dedicated swap partition (passed into the kernel via resume). As I mentioned, hibernate normally works. At the time of writing, I am currently using 2.1GB of ram, and 0K of my swap, yet I still receive the error above. Sometimes however, the hibernate doesn't work and I"m not sure what's special about it. Any ideas?
Journal output:
Jul 29 13:47:18 magos sudo[88043]: strike : TTY=pts/3 ; PWD=/home/strike ; USER=root ; COMMAND=/usr/bin/systemctl hibernate
Jul 29 13:47:18 magos sudo[88043]: pam_unix(sudo:session): session opened for user root(uid=0) by strike(uid=1000)
Jul 29 13:47:18 magos sudo[88043]: pam_unix(sudo:session): session closed for user root
Jul 29 13:49:03 magos sudo[88674]: strike : TTY=pts/3 ; PWD=/home/strike ; USER=root ; COMMAND=/usr/bin/systemctl hibernate
Jul 29 13:49:03 magos sudo[88674]: pam_unix(sudo:session): session opened for user root(uid=0) by strike(uid=1000)
Jul 29 13:49:03 magos sudo[88674]: pam_unix(sudo:session): session closed for user rootHibernate will work without issue after a reboot.
Last edited by RandomRanger (2023-07-30 15:58:29)
Any sufficiently advanced magic is indistinguishable from science.
Offline
Online
Yup, it's a Samsung nvme. It looks like that post may be relevant, I would've had a hard time finding that wiki. Thanks!
Last edited by RandomRanger (2023-07-29 22:49:28)
Any sufficiently advanced magic is indistinguishable from science.
Offline
If it is this may be a systematic issue and warrant a reference in the troubleshooting articles about hibernation.
So please report on your success ![]()
Online
ISSUE RESOLUTION DEBRIEFING
My tests indicate that the APST power optimization issues cited in the wiki article linked below are likely my issue. There is a known issue relating to Kingston and Samsung nvme drive firmware is known, and while a fix is cited in the troubleshooting page for Kingston drives no such fix is mentioned for the Samsung drives. These reports are citing Linux kernel v4.10, which is from 2017. I was hesitant to believe that the issue would still persist, however, when I attempted their workaround fix, I was no long able to reproduce my issues with hibernation. Unfortunately, the fix increased power consumption by 2-3W, which is about a 40% increase in power consumption on the laptop I've been working to optimize, and thus a not a viable solution. Knowing now that the issue relates to a power optimization for the drive, it may be possible to coerce the drive into a successful hibernation by first putting a heavy load on the disk and then hibernating, however even when the device enters hibernation it does not always successfully recover. I appreciate your help in tracking this down seth.
The point of my using hibernation is to save power, so I'll likely just stop using it altogether unless I can find a fix for the firmware issue. I may be able to update the firmware and resolve the issue, but these drives were released long after 2017 and I would've imagined the firmware it released with was compatible with the fix.
TL;DR;
APST Power Optimization is broken on some Samsung NVME drives since Linux kernel v4.10 (2017).
Link to the troubleshooting section for SSD's that lead me to a workaround and described the issue:
https://wiki.archlinux.org/title/Solid_ … leshooting
My SSD is the Samsung 970 Evo Plus (2TB), model number MZ-V7S2T0
The workaround: Add this kernel parameter, per the provided wiki's instructions
nvme_core.default_ps_max_latency_us=0This workaround increases drive power consumption (+2.5W on idle) but does allow me to successfully hibernate without issues.
Last edited by RandomRanger (2023-07-30 16:09:12)
Any sufficiently advanced magic is indistinguishable from science.
Offline
If there's a better way for me to document my findings than here in this thread, I'm happy to oblige but am still learning the Arch culture and expectations.
Any sufficiently advanced magic is indistinguishable from science.
Offline
Assuming the other thread proves this to be a pattern (likely) error message ("Call to Hibernate failed: Not enough swap space for hibernation") should link to the nvme troubleshoot paragraph from https://wiki.archlinux.org/title/Power_ … leshooting
It's a wiki, everyone can edit it. Try to be brief ![]()
Online