System not waking up from power save mode.

You have a problem with Salix? Post here and we'll do what we can to help.
Post Reply
pnbalaji
Posts: 99
Joined: 23. Apr 2013, 21:02

System not waking up from power save mode.

Post by pnbalaji »

Hi,

I have been using Salix 14.2 on my desktop for quite some time and I am experiencing a weird issue.

I usually do not shutdown my desktop as my whole family is using the desktop every now and then. However, the monitor goes blank sometimes (looks like into power save mode) and nothing works at that time. I tried Ctrl-Alt-backspace, Ctrl-Alt-F1 thru Ctrl-Alt-F10 to see if I can get into a different console. I assigned a shortcut (Super T) for xfce4-terminal and tried that so that I can see if "sudo reboot" works or not. However, that doesn't work either. The only thing that works is turn off the desktop manually by pressing the power button for about 10 seconds and then reboot it. This week it happened twice and I am worried that this might damage the HDD of the desktop. It appears that the system is going into a power save mode, but not waking up after. I looked at the dmesg output and I could see a line saying "System wakeup is disabled by ACPI".

How do I get the last boot log in Salix? I was using a different OS and experienced the same issue. From the last boot log, I could see that there was a line saying "Forward time jump detected". I am guessing that the same thing is happening now also. When I was using the different OS in the past, the command "journalctl -ab-1 > /tmp/lastboot.log" was used to capture the log of last boot. However, the journalctl command is not working in Salix.

Can someone help?

Thanks,
Balaji.
pnbalaji
Posts: 99
Joined: 23. Apr 2013, 21:02

Re: System not waking up from power save mode.

Post by pnbalaji »

Any updates on this? I really need some direction to fix this issue.

Thanks,
Balaji.
DidierSpaier
Posts: 518
Joined: 20. Jun 2016, 20:15

Re: System not waking up from power save mode.

Post by DidierSpaier »

To investigate, open a terminal in full screen mode, then type one of these commands (an interesting message can be either in /var/log/messages or in /var/log/syslog):

Code: Select all

sudo tac /var/log/messages|most
sudo tac /var/log/syslog|most
While in most you can use the up, down, page up and page down arrow keys to navigate. The display will be in reverse chronological order.
westms
Posts: 298
Joined: 17. Mar 2013, 18:51

Re: System not waking up from power save mode.

Post by westms »

Hi,
pnbalaji wrote:However, the monitor goes blank sometimes (looks like into power save mode)[...]
There are several reasons for this: Bios setup, X server, XScreenSaver and Xfce Power Manager.
pnbalaji wrote:[...]and nothing works at that time.
Again, there are several causes.
pnbalaji wrote:I tried Ctrl-Alt-backspace, Ctrl-Alt-F1 thru Ctrl-Alt-F10 to see if I can get into a different console. I assigned a shortcut (Super T) for xfce4-terminal and tried that so that I can see if "sudo reboot" works or not. However, that doesn't work either.
Generally speaking, there are only three simulated consoles that can be reached via Ctrl-Alt-F[123]. The GUI is then accessible via Alt-F7.

If a crash or a deep sleep mode, e.g. hibernation, is assumed, then it is no longer possible to reach a program using a key combination.

Btw: The X server Ctrl-Alt-backspace key combination is disabled by default.
pnbalaji wrote:The only thing that works is turn off the desktop manually by pressing the power button for about 10 seconds and then reboot it. This week it happened twice and I am worried that this might damage the HDD of the desktop.
This procedure does not cause any mechanical or electrical disadvantages for the hard disk. Only data can be lost if the filesystem was not synchronized.
pnbalaji wrote:It appears that the system is going into a power save mode, but not waking up after. I looked at the dmesg output and I could see a line saying "System wakeup is disabled by ACPI".
First, it must be checked in the BIOS setup whether ACPI is managed by the BIOS or the operating system. The operating system mode is required. The settings in the screensaver and in the power manager programs must also be checked afterwards. If the problem persists, the problem can be solved by disabling the screensaver as well as by shutting down the power management for the screen in the power manager.


If timeouts are desired to conserve hardware or save energy, but no useable behavior can be generated using screensaver and power manager, then an attempt can be made to have the X server do the work on its own. For this a file, e.g. 50-serverflags.conf, located in the directory /etc/X11/xorg.conf.d with the following contents can be created (excerpt from: /etc/X11xorg.conf-vesa):

Code: Select all

Section "ServerFlags"

# Set the basic blanking screen saver timeout.

#    Option      "BlankTime"    "10"

# Set the DPMS timeouts.  These are set here because they are global
# rather than screen-specific.  These settings alone don't enable DPMS.
# It is enabled per-screen (or per-monitor), and even then only when
# the driver supports it.

#    Option      "StandbyTime"  "20"
#    Option      "SuspendTime"  "30"
#    Option      "OffTime"      "60"

EndSection
For the desired option line, the respective comment mark is removed and the intended timeout value is set. The values given above are only examples. One can start with zero values to see if any effect can be achieved.

File ownership is root:root and file permissions are 0644.

After this implementation, the X server must be restarted.


There are other approaches to circling the problem. Not always a satisfactory solution can be found.


Another cause may be too small swap space. What is the sizes of main memory and swap partition?

Another evil reason for the behavior described is the dying of the hardware. Often the power supply has arrived its end of life.
pnbalaji wrote:How do I get the last boot log in Salix?
man lastlog

But is /var/log/lastlog the same as on systemd-contaminated operating systems?

Add:
pnbalaji wrote:"Forward time jump detected"
Can have multiple causes.

Serious hardware problems can bring the real-time clock to a halt, as well as an empty backup battery can be the cause. Also, a wrong procedure when changing the backup battery can cause unpredictable changes to the BIOS configuration data, which again can draw any malfunctions. Could the battery be empty or was replaced recently?

If VMware (or similar) is used, under certain conditions, the host system time can be synchronized at startup, overwriting VM's hardware clock. Later, during booting, the ntpd sets the time correctly. Is this Salix installation running on bare metal? The other used different OS also?
Last edited by westms on 15. Dec 2017, 18:15, edited 1 time in total.
DidierSpaier
Posts: 518
Joined: 20. Jun 2016, 20:15

Re: System not waking up from power save mode.

Post by DidierSpaier »

westms wrote:
pnbalaji wrote:How do I get the last boot log in Salix?
man lastlog

But is /var/log/lastlog the same as on systemd-contaminated operating systems?
I don't know but in Slackware lastlog does this:
NAME
lastlog - reports the most recent login of all users or of a given user
So as I posted the logs to look as are /var/log/syslog and /var/log/messages.
pnbalaji
Posts: 99
Joined: 23. Apr 2013, 21:02

Re: System not waking up from power save mode.

Post by pnbalaji »

Hi,

Sorry for the delayed reply. It appears that I missed to subscribe to this post and I was not aware of the replies to this post. I experienced the issue again today and I captured the messages and syslog. In /var/log/messages, I am seeing the below text.

Code: Select all

Dec 21 17:48:54 darkstar kernel: [    0.181939] pci 0000:00:1e.0: PCI bridge to [bus 11] (subtractive decode)
Dec 21 17:48:54 darkstar kernel: [    0.181359] pci 0000:00:1e.0: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.181229] pci 0000:00:1d.7: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.181041] pci 0000:00:1d.2: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.180868] pci 0000:00:1d.1: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.180698] pci 0000:00:1d.0: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.180343] pci 0000:00:1a.7: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.180157] pci 0000:00:1a.2: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.179981] pci 0000:00:1a.1: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.179809] pci 0000:00:1a.0: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.179639] pci 0000:00:19.0: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.178710] DMAR: Disabling IOMMU for graphics on this chipset
Is there a way to attach the log files to this post?

lastlog command doesn't seem to help. The output of the lastlog for my userid is below.

Code: Select all

bash-4.3$ lastlog -u balashaki
Username         Port     From             Latest
balashaki        pts/2    192.168.1.2      Fri Dec 15 20:13:36 -0600 2017
bash-4.3$ 
I am not sure why the last login was shown as Dec 15 even though I login into my system every day.

FYI, my desktop has 8GB of RAM and my swap size is about 6.5GB. So, I don't think it is the memory or insufficient swap space issue.

Screenshots of my BIOS are below. The first screenshot shows the Processor Power Management as GV1/GV3 only and I am not sure what it really means.
Image

The second picture shows ACPI Standby mode as S3 and it is disabled and I can't change it. These are the only two power related options I could see in my BIOS.
Image

Thanks,
Balaji.
westms
Posts: 298
Joined: 17. Mar 2013, 18:51

Re: System not waking up from power save mode.

Post by westms »

pnbalaji wrote:FYI, my desktop has 8GB of RAM and my swap size is about 6.5GB. So, I don't think it is the memory or insufficient swap space issue.
Well, this thinking is not enough. For the amount of 2-8 GiB RAM, the same amount of swap space is needed. If hibernation is required, two times RAM is required as swap space. So in your case 8 GiB swap is necessary, with hibernation 16 GiB.

Without using the hibernation state, this will not be your problem, because the swap space will only be used up a little earlier. With hibernation, all of the current main memory usage, as well as all already swapped out memory pages, must fit into the swap space. Why is hibernation not used? Have there ever been any difficulties in the past?
pnbalaji wrote:Screenshots of my Bios are below. The first screenshot shows the Processor Power Management as GV1/GV3 only and I am not sure what it really means.
It can be seen that C-States are turned off and Enhanced-Intel-SpeedStep (GV3) is turned on. Why off?

Both can be turned on, as can be seen on the far right in the "Help Message" column. So switch over for testing purposes next. After that, a cold boot is recommended. GV1 could possibly be Intel SpeedStep without extension.
pnbalaji wrote:The second picture shows ACPI Standby mode as S3 and it is disabled[...]
No, it is enabled, but grayed out. State S3 means Suspend-to-RAM. Operating system and applications with the associated data and open files remain in the voltage-supplied RAM. Everything else is more or less switched off. If then the external power supply (wall plug, power grid) fails and in a notebook the battery is empty, then everything is lost.
pnbalaji wrote:[...]and I can't change it.
Changeability may depend on the state of C-States approval.

On my computer, ACPI is allowed in the Bios for use in the operating system, but not used by the operating system, because of my refusal because of problems. I am on the safe side with it. With your computer, it could be the other way around, which can be quite problematic.

Code: Select all

Dec 21 17:48:54 darkstar kernel: [    0.181939] pci 0000:00:1e.0: PCI bridge to [bus 11] (subtractive decode)
Dec 21 17:48:54 darkstar kernel: [    0.181359] pci 0000:00:1e.0: System wakeup disabled by ACPI
Dec 21 17:48:54 darkstar kernel: [    0.181229] pci 0000:00:1d.7: System wakeup disabled by ACPI
ACPI (C-States) is turned off in Bios, but the OS tries to put the computer in a sleep state and complains about it here? Is this logging to be assigned to a time immediately before switching C-State or immediately after booting?

For a test to isolate the problem, I had already written that ACPI must be turned off in XScreenSaver and Xfce Power Manager. Either that, or allow ACPI in the Bios for testing.

The interesting question is, why did something change suddenly? Something must have been changed, or a defect in the hardware had occurred.
Post Reply