Why doesn't TW configure resume from swap by default?
Devs, I put my computer to sleep yesterday only to find it would not wake from sleep forcing a hard reset. Checking the config, there was no resume=/dev/by/uuid/..... configured. Leap 15.4 configured this automatically. Is there a reason why TW doesn't configure resume? It was quite a shock to have this laptop, that has always woken from sleep without issue suddenly die trying to wake. Where/how did resume not get set? -- David C. Rankin, J.D.,P.E.
On 2024-08-31 23:48, David C. Rankin wrote:
Devs,
I put my computer to sleep yesterday only to find it would not wake from sleep forcing a hard reset. Checking the config, there was no resume=/dev/by/uuid/..... configured.
It is not needed for sleep. It is only needed for hibernation.
Leap 15.4 configured this automatically. Is there a reason why TW doesn't configure resume?
It was quite a shock to have this laptop, that has always woken from sleep without issue suddenly die trying to wake. Where/how did resume not get set?
Maybe you did not tell the installation to create a partition for hibernation. -- Cheers / Saludos, Carlos E. R. (from 15.5 x86_64 at Telcontar)
In data domenica 1 settembre 2024 14:26:28 CEST, Carlos E. R. ha scritto:
On 2024-08-31 23:48, David C. Rankin wrote:
Devs,
I put my computer to sleep yesterday only to find it would not wake
from sleep forcing a hard reset. Checking the config, there was no resume=/dev/by/uuid/..... configured.
It is not needed for sleep.
It is only needed for hibernation.
Leap 15.4 configured this automatically. Is there a reason why TW
doesn't configure resume?
It was quite a shock to have this laptop, that has always woken from
sleep without issue suddenly die trying to wake. Where/how did resume not get set?
Maybe you did not tell the installation to create a partition for hibernation.
-- Cheers / Saludos,
Carlos E. R. (from 15.5 x86_64 at Telcontar)
As I am facing a similar problem with a PC on TW: where in the install process it asks me to set up a "partition for hibernation"? I departed until now by the idea that hibernation was writing to SWAP. Was I wrong?
On 2024-09-01 14:32, Stakanov via openSUSE Users wrote:
In data domenica 1 settembre 2024 14:26:28 CEST, Carlos E. R. ha scritto:
On 2024-08-31 23:48, David C. Rankin wrote:
Maybe you did not tell the installation to create a partition for hibernation.
As I am facing a similar problem with a PC on TW: where in the install process it asks me to set up a "partition for hibernation"? I departed until now by the idea that hibernation was writing to SWAP. Was I wrong?
As I do not use TW, the screens can be different over there from what I get on Leap. There is a toggle somewhere in the partition setup, when creating swap, to make it big enough for hibernation. -- Cheers / Saludos, Carlos E. R. (from 15.5 x86_64 at Telcontar)
In data domenica 1 settembre 2024 14:50:36 CEST, Carlos E. R. ha scritto:
On 2024-09-01 14:32, Stakanov via openSUSE Users wrote:
In data domenica 1 settembre 2024 14:26:28 CEST, Carlos E. R. ha scritto:
On 2024-08-31 23:48, David C. Rankin wrote:
Maybe you did not tell the installation to create a partition for hibernation.
As I am facing a similar problem with a PC on TW: where in the install process
it asks me to set up a "partition for hibernation"? I departed
until now by the idea that hibernation was writing to SWAP. Was I wrong?
As I do not use TW, the screens can be different over there from what I get on Leap.
There is a toggle somewhere in the partition setup, when creating swap, to make it big enough for hibernation.
-- Cheers / Saludos,
Carlos E. R. (from 15.5 x86_64 at Telcontar)
ah, yeah, that boils down to "swap plus". Only to discover then that when you grow in RAM all the providence was void..... will have to set up a different SSD for SWAP I guess.
On 9/1/24 7:26 AM, Carlos E. R. wrote:
It is not needed for sleep.
It is only needed for hibernation.
Without it set, when I closed the lid and open it again - my box hardlocks. Never happened with 15.4 - always just worked. How to diagnose "sleep" issues from "works on 15.4" to "broken on TW" (same laptop) -- David C. Rankin, J.D.,P.E.
On 2024-09-03 02:21, David C. Rankin wrote:
On 9/1/24 7:26 AM, Carlos E. R. wrote:
It is not needed for sleep.
It is only needed for hibernation.
Without it set, when I closed the lid and open it again - my box hardlocks. Never happened with 15.4 - always just worked.
How to diagnose "sleep" issues from "works on 15.4" to "broken on TW" (same laptop)
A swap partition is totally irrelevant for sleep. In sleep, the CPU stops, but the RAM is kept powered. Several chips are kept powered, many others not. Machine should resume instantly. In hibernation, the memory contents are written to hard disk, and then everything is powered off. It takes time to resume. Both modes are different. You have to know which one you are interested in. Which one activates when you close the lid, is configurable. There is a wiki page with things to do for diagnosis, but I don't remember which one; maybe tomorrow I can find it in one of my bugzillas. It is is 3 AM here, so perhaps you can find it. Anyway, I know little about TW, I don't like living at the edge. -- Cheers / Saludos, Carlos E. R. (from 15.5 x86_64 at Telcontar)
On 9/2/24 8:27 PM, Carlos E. R. wrote:
There is a wiki page with things to do for diagnosis, but I don't remember which one; maybe tomorrow I can find it in one of my bugzillas. It is is 3 AM here, so perhaps you can find it. Anyway, I know little about TW, I don't like living at the edge.
Thanks Carlos, You shouldn't worry about TW, it is just Leap that has rolling updates. You can make it just like leap by watching the list for any gotchas, and just not updating until after they are resolved. With my cynicism, I give TW and 2-thumbs Up. The rolling part is a bit clunky compared with Arch, but that will improve as the handling of the myriad of different possible repos is fine tuned. Arch is a lot simpler, You have 2 (actually 3) repos, and than the AUR (Arch user repo - e.g. buildservice). The primary 2 (3) are "core", "extra" and ("testing", e.g. staging). So from that standpoint, it is much easier for Arch to make the rolling part seamless -- but TW is not bad at all. TW has hidden (not provided) all the /etc/systemd/*.conf files. The defaults are in /usr/lib/systemd, but the user config in /etc/systemd is up to you. I used the defaults on 15.4 only adding the HandlePowerKey = 'suspend' So my sleep is what they call "suspend" on 15.4 which is also the default for HandleLidSwitch=suspend. (which is the same for the /usr/lib/systemd/logind.conf) So my TW sleep/suspend is configured exactly like 15.4, it just doesn't work. -- David C. Rankin, J.D.,P.E.
On 9/2/24 8:27 PM, Carlos E. R. wrote:
There is a wiki page with things to do for diagnosis, but I don't remember which one; maybe tomorrow I can find it in one of my bugzillas. It is is 3 AM here, so perhaps you can find it. Anyway, I know little about TW, I don't like living at the edge.
Wiki page is 2020 for suspend to RAM and really has nothing useful. The other page is pm-utils which doesn't exist anymore in TW. The Arch wiki had good info: https://wiki.archlinux.org/title/Power_management/Suspend_and_hibernate Here is the bottom line from testing (with systemd defaults): Tumbleweed Sleep: # echo s2idle > /sys/power/mem_sleep # cat /sys/power/mem_sleep [s2idle] deep # systemctl suspend (s2idle reached but system fan not stopped, power remains on) # echo deep > /sys/power/mem_sleep # cat /sys/power/mem_sleep s2idle [deep] # systemctl suspend (deep is s3 suspend, and suspend works) The odd part is just days ago when I closed the lid -- all hell broke loose trying to come out of sleep resulting in a hard-reset and praying the SSD wasn't corrupted. Tonight, invoking s2idel or s3 suspend from the command line both worked as advertised. The only surprise is s2idle does NOT result in the system being powered off with the little power-button like blinking to indicate you are in sleep mode. In s2idle the RAM and disk are frozen effectively stopping CPU processing, but otherwise the system looks like it is ON. Suspend [deep] in /sys/power/mem_sleep is your normal s3 suspend mode and it does power down the system and the power-button blinks normally. In either case, wireless networking is reestablished as soon as the laptop is brought out of sleep -- very nice. I don't know why the lid closing caused suspend to fail to resume -- but after adding the resume= line (which has nothing to do with suspend to RAM), it all seems to work fine. Dunno. -- David C. Rankin, J.D.,P.E.
participants (3)
-
Carlos E. R. -
David C. Rankin -
Stakanov