Hibernation

[Shrek]

Is the PC safe from power cuts during Hibernation, in the sense that no corruption will occur?

 

[Bill_Bright]

Is the PC safe from power cuts during Hibernation, in the sense that no corruption will occur?

That's one of the benefits of hibernation (and hybrid mode for PCs). The environment (what's in RAM) is saved to the hibernation file so you can recover it when power comes back.

 

[qxp]

Its safer. There are three factors here - you can hope that the operating system is smart enough to flush buffers before going into hybernation and put filesystems into a consistent state, but I don't know how exactly different OSes handle it. Also, in hybernation (and even in suspend to RAM mode) the computer consumes less power, so the power supply capacitors and inductors are more likely to carry through a short drop out. If the dropout was extended then you can hope that on power on the computer will reload RAM from disk and continue as before.

There are a few things that could go wrong - the BIOS or the OS can get confused and do a fresh boot when power is restored instead of resuming from disk. I've seen this happen a few times. A worse situation is that the power can come back, the system will start resuming and then the power might be cut again just when the system is busily accessing the filesystem. You can guard against it by instructing BIOS to keep system off after power failure.

 

[Bill_Bright]

Ummm, sorry but there's some misinformation there.

you can hope that the operating system is smart enough to flush buffers before going into hybernation

in hybernation (and even in suspend to RAM mode) the computer consumes less power

First, it is important to distinguish between laptops and PCs. By default, laptops use hibernation, PCs use "hybrid" hibernation. So when inquiring about hibernation, it is important to specify whether PC or laptop.

For a laptop (for which hibernation was first created), when you close the lid, the environment is saved to the hiberfil.sys file and the system shuts downs. This is intentional so you don't run down the battery while you commute to your next destination. Then when you open the lid, you can resume right were you left off.

During this state, the computer is consuming NO power - other than keeping the RTC (real time clock) ticking so the laptop always knows the date and time.

so the power supply capacitors and inductors are more likely to carry through a short drop out.

Ummm, sorry but this makes no sense. For one, inductors do not store power so there would be nothing to "carry through". And by design (not to mention regulation), PC power supplies capacitors will bleed off charges to safe levels almost immediately upon the loss of power. This "decay" in voltage potentials ensures no dangerous (or useful) voltage remains.

In PCs, by design, hibernation works differently. In PCs, hibernation works in hybrid mode where, again, by design, the +5Vsb standby voltage required by the ATX Form Factor standard keeps the data in RAM "alive" (in a low voltage state) while the computer goes to sleep (into standby). This allows for faster wake-up speed. The low voltage state help minimize power consumption while in standby. DDR4 is good at "sipping" power during this state, DDR5 is better.

HOWEVER, should there be a power outage while the PC is asleep, the "hybrid" mode also saves the environment to the hiberfil.sys file so in that event, the environment can be restored too - it will just take a few seconds longer.

You can guard against it by instructing BIOS to keep system off after power failure.

That is the default.

 

[Wirko]

There are a few things that could go wrong - the BIOS or the OS can get confused and do a fresh boot when power is restored instead of resuming from disk. I've seen this happen a few times.

If you have a dual boot PC and put one OS into hibernation, the system shouldn't let you boot the other one. But sometimes funny things happen, just to let you know you're attempting non-standard procedures.

 

[Bill_Bright]

It is important to understand a drunk could run a stop sign, drive across 4 yards and slam into your front porch. So does that mean you should never sit on your porch or leave or enter through your front door? Of course not.

There will always be exceptions and extreme cases where something "could" go wrong. But 100s of millions of users allow their PCs to simply go to sleep (into hybrid mode) of shut the lids of their laptops (into hibernation) every single day even multiple times every day with NO PROBLEMS whatsoever. Even when there are problems, a simple reboot typically clears them making the odds of irreversible problems even less likely.

So it is important to put things into perspective and not generator or spread unnecessary FUD.

 

[qxp]

Ummm, sorry but there's some misinformation there.


First, it is important to distinguish between laptops and PCs. By default, laptops use hibernation, PCs use "hybrid" hibernation. So when inquiring about hibernation, it is important to specify whether PC or laptop.

For a laptop (for which hibernation was first created), when you close the lid, the environment is saved to the hiberfil.sys file and the system shuts downs. This is intentional so you don't run down the battery while you commute to your next destination. Then when you open the lid, you can resume right were you left off.

I think you are focusing too much on how Windows does it. I am running Linux and there are also Macs. I agree that by hibernation one usually means the suspend mode where information from RAM is written to disk in some fashion and then the power to RAM and CPU is turned off for maximum power savings. During the startup the operating system is supposed to check whether it was put into hibernation previously and restore system state and data from disk.

During this state, the computer is consuming NO power - other than keeping the RTC (real time clock) ticking so the laptop always knows the date and time.

This depends on the system. If you have wake-on-lan enabled the power supply will be active and powering the network card. If you have IPMI module active a whole small computer will be powered on and online.

Ummm, sorry but this makes no sense. For one, inductors do not store power so there would be nothing to "carry through". And by design (not to mention regulation), PC power supplies capacitors will bleed off charges to safe levels almost immediately upon the loss of power. This "decay" in voltage potentials ensures no dangerous (or useful) voltage remains.

Actually inductors do store power as the magnetic field. For large inductors it can be quite substantial. People even made electromagnetic pulse devices by having a large inductor, passing current through it and then imploding the device. https://en.wikipedia.org/wiki/Explosively_pumped_flux_compression_generator

If you have a notebook, unplug a notebook and only then unplug the power supply - you will see the power LED dim very slowly as it is using stored power. Larger power supplies - such as for a workstation - store more power and I had a couple of instances where a workstation would happily keep working when the lights in the room blinked from power outage. Of course, there was no substantial CPU load at the time. I know this was not just a brown out because my UPSes logged input voltage going to 0. The workstation was not on UPS because I just got it and was configuring it.

In PCs, by design, hibernation works differently. In PCs, hibernation works in hybrid mode where, again, by design, the +5Vsb standby voltage required by the ATX Form Factor standard keeps the data in RAM "alive" (in a low voltage state) while the computer goes to sleep (into standby). This allows for faster wake-up speed. The low voltage state help minimize power consumption while in standby. DDR4 is good at "sipping" power during this state, DDR5 is better.

HOWEVER, should there be a power outage while the PC is asleep, the "hybrid" mode also saves the environment to the hiberfil.sys file so in that event, the environment can be restored too - it will just take a few seconds longer.

On Linux at least, every system has a choice between "Saving to RAM" and "hibernation" with words that can differ for different operating system (or if you want to change them). "Saving to RAM" works by keeping the power to RAM but placing the CPU and the GPU in very low power state. "Hibernation" works by saving to disk and turning off power. Last time I used Windows it also offered both of these options.
Linux systems which were configured without a space to store RAM to will not have hibernation mode, of course.

That is the default.

On Dell workstations that I have the default choice for action after power failure is "Last state". I always change it manually to be off, except for systems that are set to always on so I can do wake on lan.

 

[eidairaman1]

Is the PC safe from power cuts during Hibernation, in the sense that no corruption will occur?

Its not suspend, it is safer, but hogs storage space for contents in ram

 

[Bill_Bright]

I think you are focusing too much on how Windows does it.

Too much? No I am not. I am using Windows to illustrate because that is what the majority of users/readers on this site are familiar with. What you don't seem to understand is, the OS does not matter!!!

These "suspend" or "hibernation" or whatever you wish to call them modes all work essentially the same way as established through consensus by the industry - not Microsoft, Apple, or Linux. And in fact, these "protocols" are set primarily by the hardware industry and again, NOT Microsoft or the others. Yes, Microsoft sits at the table and has a say, but they are just one vote among many.

This is what allows PC users to run Linux on their ASUS or Gigabyte motherboard, with their AMD or Intel CPU and their MSI or XFX graphics card, and Samsung or Crucial drive, all powered by their Seasonic or Corsair power supplies. Or, should they change their minds, they can swap in a WD drive and run Windows on that exact same hardware.

This depends on the system. If you have wake-on-lan enabled the power supply will be active and powering the network card.

No. You are wrong. You are talking about a PC where, as required by the ATX Form Factor standard, when shutdown (but still plugged in and the master power switch set to on), the PC is still in a standby mode and the required +5Vsb standby voltage is still being distributed to multiple points on the board. This is how and why "hybrid" mode is able to work in a PC - there is still voltage applied to the RAM. Note that +5Vsb voltage is regardless if running Windows, Linux or whatever OS is supported by that hardware.

You even quoted me where I was specifically said "laptop", not PC.

You also quoted me where I specifically said "it is it is important to distinguish between laptops and PCs" yet you clearly, and sadly have failed to or apparently grasp that difference.

Actually inductors do store power as the magnetic field.

Only when current is flowing through them. When power is removed from the coil, the magnetic field collapses. Please check the link in my sig to see if I might know a little about how electronics work.

Capacitors, on the other hand, can store power for long periods of time - hours and days even, depending on the design. This is NOT a problem inside laptops or on the output side of a PC power supply because those voltages are low voltage DC. So your laptop's LED still glowing for a few seconds after all power is removed is no big deal - even somewhat expected as charges decay.

Inside a PC PSU, on the primary side of the main transformer, that's a different story as deadly voltages may linger if not properly and quickly bled off when power is removed. Fortunately, in almost every jurisdiction around the world, there are laws and regulations that require such bleeder circuits to quickly and safely discharge those big filter caps to safe levels in just a couple seconds. Still, this assumes the PC PSU is not damaged (or a cheap, illegal counterfeit) so stay out of the inside of a PC PSU unless a qualified tech.

and I had a couple of instances where a workstation would happily keep working when the lights in the room blinked from power outage.

Clearly indicating you don't have an understanding of the ATX Form Factor standard requirement called Hold-up Time. That is not a criticism, just an observation. This requirement requires all ATX compliant PSUs to maintain full output power for 12ms at 100% load and 17ms at 80% load "despite a loss of input power". Why is that requirement in there? Because power grid fluctuations are common and we don't want our computers crashing every time the lights flicker. And again, this is regardless if running Windows, Linux - or no OS at all!

Now this side debate is OT. Shrek has his answer. Suspend/hibernation mode has been around now for decades. The hardware and OS industries have learned how to ensure it works perfectly every times, without causing data corruption or HW damage. Is it possible something might go wrong? Of course! Until Man can make perfection 100% of the time, there will always be exceptions. But that's why every one here has a good, robust backup plan, right?

Time to move on.

 

[qxp]

Too much? No I am not. I am using Windows to illustrate because that is what the majority of users/readers on this site are familiar with. What you don't seem to understand is, the OS does not matter!!!

These "suspend" or "hibernation" or whatever you wish to call them modes all work essentially the same way as established through consensus by the industry - not Microsoft, Apple, or Linux. And in fact, these "protocols" are set primarily by the hardware industry and again, NOT Microsoft or the others. Yes, Microsoft sits at the table and has a say, but they are just one vote among many.

This is what allows PC users to run Linux on their ASUS or Gigabyte motherboard, with their AMD or Intel CPU and their MSI or XFX graphics card, and Samsung or Crucial drive, all powered by their Seasonic or Corsair power supplies. Or, should they change their minds, they can swap in a WD drive and run Windows on that exact same hardware.

In the interest of clearing up misconceptions, it is not actually true that hibernation mode is determined by the hardware. All this is a software checkpoint written to disk and can be (and was) implemented without hardware support. Suspend to RAM does need hardware support.

You also quoted me where I specifically said "it is it is important to distinguish between laptops and PCs" yet you clearly, and sadly have failed to or apparently grasp that difference.

No, I merely have experience with a lot more systems than Windows PC or laptops.

Only when current is flowing through them. When power is removed from the coil, the magnetic field collapses. Please check the link in my sig to see if I might know a little about how electronics work.

Just to be clear magnetic fields do store energy and it is possible to extract it (thats how inductors work): https://en.wikipedia.org/wiki/Magnetic_energy
Looking at the link in your signature you surely know that - is there some language in what you are saying that I misunderstand ?

Capacitors, on the other hand, can store power for long periods of time - hours and days even, depending on the design. This is NOT a problem inside laptops or on the output side of a PC power supply because those voltages are low voltage DC. So your laptop's LED still glowing for a few seconds after all power is removed is no big deal - even somewhat expected as charges decay.

From my point of view typical capacitors store energy for longer periods than a typical inductor because we have very good insulators, but our core materials are all very lossy in comparison. But it takes a finite time for the magnetic field to collapse and during this time it can provide power.

Clearly indicating you don't have an understanding of the ATX Form Factor standard requirement called Hold-up Time. That is not a criticism, just an observation. This requirement requires all ATX compliant PSUs to maintain full output power for 12ms at 100% load and 17ms at 80% load "despite a loss of input power". Why is that requirement in there? Because power grid fluctuations are common and we don't want our computers crashing every time the lights flicker. And again, this is regardless if running Windows, Linux - or no OS at all!

Exactly. What I was saying is that with 1% load this can last a lot longer than 17ms, depending on power supply of course.

Now this side debate is OT. Shrek has his answer. Suspend/hibernation mode has been around now for decades. The hardware and OS industries have learned how to ensure it works perfectly every times, without causing data corruption or HW damage. Is it possible something might go wrong? Of course! Until Man can make perfection 100% of the time, there will always be exceptions. But that's why every one here has a good, robust backup plan, right?

Time to move on.

Robust backup plan ? Let me check if there is an existing thread for that..

 

[Bill_Bright]

In the interest of clearing up misconceptions, it is not actually true that hibernation mode is determined by the hardware.

Gee whiz dude! It seems you don't even read what others are saying when you quote them. In the interest of clearing up misconceptions, you really should before you post replies.

I NEVER said "the hardware" determines the mode. I said - twice - even emphasizing it with underlines, that it is the hardware industry. I added, yes MS sits at the table but they don't run the show. OF COURSE it is the OS (based on user settings) that is looking for (counting) periods of inactivity. But after a user defined period, the OS initiates the shutdown for itself and signals the HW to go to sleep. To suggest it happens without hardware support is just silly. The protocols, drivers, power options code, timers are all being processed by the CPU in RAM on the motherboard. Without the CPU, RAM or motherboard, the OS would be stuck on a drive (another piece of hardware) somewhere, doing nothing.

From my point of view typical capacitors store energy for longer periods than a typical inductor because we have very good insulators

Well, sorry, but your point of view is simply and totally incorrect. You might want to take some intro to electronics courses because that is NOT how inductors work, or why capacitors store energy.

Just to be clear magnetic fields do store energy and it is possible to extract it (thats how inductors work): https://en.wikipedia.org/wiki/Magnetic_energy
Looking at the link in your signature you surely know that.

Come on! You need to read AND understand your own link! It clearly says (my bold underline added),

the energy stored in an inductor when a current flows through it

This means when current stops flowing through the inductor - as happens during hibernate mode or when power is removed - the magnetic field immediately collapses. That's how inductors work.

Flowing current -> magnetic field. No current flow -> no magnetic field. 7th grade electronics.

The duration any given capacitor holds a charge is determined by many factors, not just the dielectric (insulating material) between the two plates. Also is the type of capacitor, size of the capacitor, type of dielectric, and the amount of charge. All of these factors are taken into consideration by the designers based on how long they want the charge to hold.

is there some language in what you are saying that I misunderstand ?

Ummm, English?

I said before, it is "time to move on". Apparently, you are not understanding that either. It means to stop driving this thread off topic. It is not helping the OP.

I am done here. That means I will not respond to any thing more you say.

Have a good day.

Sorry @Shrek for my part in this off-topic tangent. I trust you have the information you need. As I noted above, suspend/hibernation mode has been around now for decades. It is a very mature process. The hardware and OS industries have refined it such that any risks are minimal at worst. Not impossible sadly, but pretty close.

 

[qxp]

Well, sorry, but your point of view is simply and totally incorrect. You might want to take some intro to electronics courses because that is NOT how inductors work, or why capacitors store energy.

Come on! You need to read AND understand your own link! It clearly says (my bold underline added),

This means when current stops flowing through the inductor - as happens during hibernate mode or when power is removed - the magnetic field immediately collapses. That's how inductors work.

Flowing current -> magnetic field. No current flow -> no magnetic field. 7th grade electronics.

This is not the correct point of view. If you have an inductor, applying external electrical potential results in increasing current and building magnetic field. Removing external electric potential results in decreasing magnetic field and generated EMF.

To see that the energy in the inductor is stored in the magnetic field consider a coil of wire around empty space (air or vacuum). Its inductance can be increased by inserting a high-permeability core. Its inductance can actually be decreased by inserting a diamagnetic core. If the energy was not stored in the magnetic field why would inserting or removing a core have any effect ?

Also in the link https://en.wikipedia.org/wiki/Magnetic_energy you see the usual formula of LI^2/2 for stored inductor energy, but below that formula there is an explanation that the energy is stored in the magnetic field. In fact, this is the primary effect - the built-up magnetic field is proportional to the current (in idealized situation), and the EMF you can extract is given by an integral which in a simple situation gives I^2.

If you do a calculation taking into account radiative losses, core hysteresis and other effects the stored energy becomes more complicated and you see that it is really driven by the magnetic field.

The duration any given capacitor holds a charge is determined by many factors, not just the dielectric (insulating material) between the two plates. Also is the type of capacitor, size of the capacitor, type of dielectric, and the amount of charge. All of these factors are taken into consideration by the designers based on how long they want the charge to hold.

Ummm, English?

I said before, it is "time to move on". Apparently, you are not understanding that either. It means to stop driving this thread off topic. It is not helping the OP.

I am done here. That means I will not respond to any thing more you say.

Have a good day.

Sorry @Shrek for my part in this off-topic tangent. I trust you have the information you need. As I noted above, suspend/hibernation mode has been around now for decades. It is a very mature process. The hardware and OS industries have refined it such that any risks are minimal at worst. Not impossible sadly, but pretty close.

The problem is that the information you are providing is, at best, approximately correct.

The OP was asking "Is the PC safe from power cuts during Hibernation, in the sense that no corruption will occur?" and the answer is that depends on the OS and not perfectly. Most of the time works fine and sometimes fails is not minimal risk. I remember quite a few times when people would ask me to help restore something important after putting the laptop in hibernation, especially if they battery drains while hibernation is on. On many of these occasions there was not much I could do.

 

[eidairaman1]

Take it to private, @freeagent? I believe the question was answered in a simple manner,no need to over explain it.

 

[Harlow]

Fortunately, in almost every jurisdiction around the world, there are laws and regulations that require such bleeder circuits to quickly and safely discharge those big filter caps to safe levels in just a couple seconds.

Thank you @Bill_Bright for confirming my suspicions.

Back in 2023 I was contributing to a discussion on another forum about how long the bulk capacitor in a typical ATX PSU remains at a dangerously high voltage, after disconnecting the mains. People were quoting the discharge time as 'several years'. I think they misinterpreted a "fact" they'd found on the internet.

When I entered the discussion and politely said this was not true for electrolytic capacitors in ATX PSUs, the entire thread mysteriously disappeared. I think I upset some of the moderators, who take a very dim view of anyone discussing attempts to repair a computer PSU.

Prior to the thread being removed, this image was posted as a warning about the dangers of messing around with electricity. The original poster was slightly taken aback by the insinuation he was a total novice. Mind you, mains voltages can be lethal. If in doubt, refer to a qualified person.

 

[Bill_Bright]

I think I upset some of the moderators, who take a very dim view of anyone discussing attempts to repair a computer PSU.

And rightfully so. It is important to note anything that plugs into the wall can kill. So unless a properly trained and qualified technician, no one should be opening a PSU case. There are no "user-serviceable" parts inside.

Regarding my comment you quoted about safety laws and regulations requiring bleeder circuits, I do wish you had included the very next, and perhaps most important sentence in that paragraph which said,

Still, this assumes the PC PSU is not damaged (or a cheap, illegal counterfeit) so stay out of the inside of a PC PSU unless a qualified tech.

 

[Shrek]

Do not mess with power supplies as the main capacitors can carry a lethal charge even when the supply is not plugged in.

Teen electrocuted while taking apart unplugged computer

In an activity familiar to any kid with a passion for electronics, the life of a 16-year-old boy in Shawnee, Kan., was cut short last month while he was taking apart a computer, stripping it for parts to build another."He was always kinda tinkering with computers," Capt. Dan Tennis of the...

www.today.com

 

[Harlow]

https://www.today.com/money/teen-electrocuted-while-taking-apart-unplugged-computer-1C6365434

Just before the thread I was following on another forum disappeared, a moderator posted the link about the16-year old who tragically died "taking apart a computer" (no specific mention of disassembling the PSU though). It's the link you find when Googling "computer deaths". I spent some time trying to find more reports of death directly attributable to messing with computer power supplies, but couldn't find any. Next I researched Government statistics in my country for home deaths due to electrocution (in the same year as the 16-year old's death) and came up with eight. Although these deaths were tragedies for the families, I doubt they were all related to misguided ATX PSU repair attempts.

When I published a link to these government statistics, I suggested deaths caused by gunshot wounds and vehicle collisions were more common and the thread vanished. Maybe forums like this can be subjected to litigation in some countries for providing assistance that could lead to harm? If so, I hope people reading home car mechanic forums heed advice to bleed brake lines, when replacing disc/rotor calipers on their car. Brake failure at speed could kill more than just the guy who "mended" the braking system.

I wonder if this discussion will suddenly be terminated too? Sorry!

 

[Bill_Bright]

Although these deaths were tragedies for the families, I doubt they were all related to misguided ATX PSU repair attempts.

Assuming the reports are accurate and these tragedies did occur when the victim was attempting to repair a computer, not sure what else it could be, but misguided "PSU" repair attempts by unqualified people. The voltages leaving those PSUs are low-voltage DC voltages (+3.3VDC, +5VDC, and 12VDC) - no where near "deadly" potentials. Only the input side of the PSUs, where the mains AC voltage enters the PSU, do deadly potentials exist.

I doubt even if someone with a sensitive heart pacemaker purposely grabbed hold of the +12V rail that it would affect him/her in anyway. That said, if you have a pacemaker, please don't try it to see if I am right!

Common sense would tell anyone (with common sense) to unplug the PSU before sticking fingers and metal probes in there. But it takes training/education to be aware that storage capacitors can retain hazardous voltages for extended periods if not properly bled off when power is removed. And it takes training/education and a bit of common sense to be aware and to understand that even properly designed PSUs and the components withinb can have factory defects, or be damaged after leaving the factory such that existing bleeder circuits fail to do their jobs - with catastrophic results.