Mint 17 won't suspend

[Cranky_Frankie]

I have a 64 bit Asus 4 core chip and mobo with Mint 17 Cinnamon. Last week I posted that it wouldn't suspend or hibernate, although it did when I had the 32 bit Mint installed. I was told I needed to have 2X the memory in swap file size to get it to work. I have LVM on this box so it took me a while to figurre out how to change the swap file size in LVM. I have 16 gig DDR3 in this box ( 2 x . I tried to make the swap 32gig but it says I don't have enough space to do that. Right now the swap file size is just under 16gig.

So what can I do to make the suspend work? Should I take out one DDR3 memory? That would leave me with 8gig. If that works fine but is there any other way? Do you really need 2x the physical memory in swap to make suspend work? I'd hate to have to remove the memory since I already have it.

The ironic thing is that I run free -m all the time and I'm barely using any real memory at all, so the swap file is never getting used anyway.

[ofb]

It's not a swap issue. The 2x business is old and unfounded Lore, often repeated with good intention. As long as your swap is equal to RAM, it's plenty big enough for Suspend.

The problem is Suspend and Hibernate are notorious for not working in Linux. Sorry but I don't immediately know the source of your issue. Hopefully someone good at that will chip-in soon. For myself I've stopped fighting it. So far, every laptop has worked, and every desktop has failed. I started with Linux about 2005.

EDIT: oh, and you want to post the output of 'inxi -F' so the people helping you can see what hardware they're dealing with.

[Cranky_Frankie]

The funny thing is the suspend/hibernate function *did* work when I had 32 bit Mint on this box. Since then I've upgraded to the Asus mobo, chip, and 64 bit Mint and that's when it stopped working. I don't know how to get this output any cleaner, I ran "inxi -F > hw.txt" at the terminal:

#[1;34mSystem: #[0;37m #[1;34mHost:#[0;37m Kilgore #[1;34mKernel:#[0;37m 3.13.0-24-generic x86_64 (64 bit) #[1;34mDesktop:#[0;37m Gnome #[1;34mDistro:#[0;37m Linux Mint 17 Qiana
#[1;34mMachine: #[0;37m #[1;34mMobo:#[0;37m ASUSTeK #[1;34mmodel:#[0;37m A55BM-E #[1;34mversion:#[0;37m Rev X.0x #[1;34mBios:#[0;37m American Megatrends #[1;34mversion:#[0;37m 1302 #[1;34mdate:#[0;37m 01/20/2014
#[1;34mCPU: #[0;37m #[1;34mQuad core#[0;37m AMD A10-5800K APU with Radeon HD Graphics (-MCP-)#[0;37m #[1;34mcache:#[0;37m 8192 KB #[1;34mflags:#[0;37m (lm nx sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm)#[0;37m
#[1;34m #[0;37m #[1;34mClock Speeds:#[0;37m #[1;34m1:#[0;37m 1400.00 MHz#[0;37m #[1;34m2:#[0;37m 2900.00 MHz#[0;37m #[1;34m3:#[0;37m 3800.00 MHz#[0;37m #[1;34m4:#[0;37m 1400.00 MHz#[0;37m
#[1;34mGraphics: #[0;37m #[1;34mCard:#[0;37m Advanced Micro Devices [AMD/ATI] Trinity [Radeon HD 7660D]
#[1;34m #[0;37m #[1;34mX.Org:#[0;37m 1.15.1 #[1;34mdrivers:#[0;37m ati,radeon (unloaded: fbdev,vesa) #[1;34mResolution:#[0;37m 1280x1024@60.0hz
#[1;34m #[0;37m #[1;34mGLX Renderer:#[0;37m Gallium 0.4 on AMD ARUBA #[1;34mGLX Version:#[0;37m 3.0 Mesa 10.1.0#[0;37m
#[1;34mAudio: #[0;37m #[1;34mCard:#[0;37m Advanced Micro Devices [AMD] FCH Azalia Controller #[1;34mdriver:#[0;37m snd_hda_intel #[1;34mSound:#[0;37m ALSA #[1;34mver:#[0;37m k3.13.0-24-generic
#[1;34mNetwork: #[0;37m #[1;34mCard:#[0;37m Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller #[1;34mdriver:#[0;37m r8169
#[1;34m #[0;37m #[1;34mIF:#[0;37m eth0 #[1;34mstate:#[0;37m up #[1;34mspeed:#[0;37m 10 Mbps #[1;34mduplex:#[0;37m half #[1;34mmac:#[0;37m 40:16:7e:77:1d:2d
#[1;34mDrives: #[0;37m #[1;34mHDD Total Size:#[0;37m 640.1GB (11.0% used) #[1;34m1:#[0;37m #[1;34mid:#[0;37m /dev/sda #[1;34mmodel:#[0;37m WDC_WD6400AAKS #[1;34msize:#[0;37m 640.1GB
#[1;34mPartition:#[0;37m #[1;34mID:#[0;37m / #[1;34msize:#[0;37m 571G #[1;34mused:#[0;37m 66G (13%) #[1;34mfs:#[0;37m ext4 #[1;34mID:#[0;37m /boot #[1;34msize:#[0;37m 237M #[1;34mused:#[0;37m 48M (22%) #[1;34mfs:#[0;37m ext2
#[1;34m #[0;37m #[1;34mID:#[0;37m swap-1 #[1;34msize:#[0;37m 16.31GB #[1;34mused:#[0;37m 0.00GB (0%) #[1;34mfs:#[0;37m swap
#[1;34mRAID: #[0;37m No RAID devices detected - /proc/mdstat and md_mod kernel raid module present
#[1;34mSensors: #[0;37m #[1;34mSystem Temperatures: cpu:#[0;37m 0.0C #[1;34mmobo:#[0;37m N/A #[1;34mgpu:#[0;37m 0.0
#[1;34m #[0;37m #[1;34mFan Speeds (in rpm): cpu:#[0;37m N/A
#[1;34mInfo: #[0;37m #[1;34mProcesses:#[0;37m 183 #[1;34mUptime:#[0;37m 14 min #[1;34mMemory:#[0;37m 1001.9/15173.0MB#[0;37m #[1;34mClient:#[0;37m Shell #[1;34minxi:#[0;37m 1.8.4#[0;37m #[0;37m
#[0m

[ofb]

Wow, neat. Next time just highlight the terminal output with your mouse. Then you can paste it with a middle click.

Fab glitch though. Mine does that too. I'll have to figure out why the char sets don't play well.


EDIT - here's a quick cleanup

Code: Select all

System:    Host:  Kilgore  Kernel:  3.13.0-24-generic x86_64 (64 bit)  Desktop:  Gnome  Distro:  Linux Mint 17 Qiana
Machine:    Mobo:  ASUSTeK  model:  A55BM-E  version:  Rev X.0x  Bios:  American Megatrends  version:  1302  date:  01/20/2014
CPU:    Quad core  AMD A10-5800K APU with Radeon HD Graphics (-MCP-)   cache:  8192 KB  flags:  (lm nx sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm)  
Clock Speeds:   1:  1400.00 MHz   2:  2900.00 MHz   3:  3800.00 MHz   4:  1400.00 MHz 
Graphics:    Card:  Advanced Micro Devices [AMD/ATI] Trinity [Radeon HD 7660D] 
X.Org:  1.15.1  drivers:  ati,radeon (unloaded: fbdev,vesa)  Resolution:  1280x1024@60.0hz 
GLX Renderer:  Gallium 0.4 on AMD ARUBA  GLX Version:  3.0 Mesa 10.1.0 
Audio:    Card:  Advanced Micro Devices [AMD] FCH Azalia Controller  driver:  snd_hda_intel  Sound:  ALSA  ver:  k3.13.0-24-generic
Network:    Card:  Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller  driver:  r8169 
IF:  eth0  state:  up  speed:  10 Mbps  duplex:  half  mac:  40:16:7e:77:1d:2d
Drives:    HDD Total Size:  640.1GB (11.0% used)  1:   id:  /dev/sda  model:  WDC_WD6400AAKS  size:  640.1GB 
Partition:   ID:  /  size:  571G  used:  66G (13%)  fs:  ext4  ID:  /boot  size:  237M  used:  48M (22%)  fs:  ext2 
ID:  swap-1  size:  16.31GB  used:  0.00GB (0%)  fs:  swap 
RAID:   No RAID devices detected - /proc/mdstat and md_mod kernel raid module present
Sensors:    System Temperatures: cpu:  0.0C  mobo:  N/A  gpu:  0.0 
Fan Speeds (in rpm): cpu:  N/A 
Info:    Processes:  183  Uptime:  14 min  Memory:  1001.9/15173.0MB   Client:  Shell  inxi:  1.8.4  

[all41]

ofb pretty much nailed it.
All these systems enter suspend but I have only known 3 desktops to recover properly- EVER.
2 of them are still here. I have several computers around here right now and the ones with Mint work as follows:

1 desktop w/Mint 16 suspends and recovers normally
1 desktop w/Mint 16 suspends but will not recover
1 desktop w/Mint 17 suspends but will not recover, but worked normally when it had Vista
1 desktop w/Mint 17 suspends and recovers
1 desktop w/Mint 17 suspends but will not recover (this PC), however it does suspend and recover booting to the original Win7 hard drive.
2 laptops work perfectly, older Dell Mint 17, newer Compaq dual boot Mint 16 & W8.1
These all use Mate de
Plus among friends, neighbors, and relatives this pattern pretty much repeats.
I will follow this post. I would love to know how to remedy this.

[ofb]

While we're at it please show us the outputs for these four -

free

cat /etc/fstab

sudo blkid

sudo fdisk -l


EDIT: all41, have you ever had the swap stop working after failed tests of Hibernate and Suspend? I did once, but haven't been able to replicate. Only time I've been faced with a broken swap so I'm still pretty curious how that could happen.

[all41]

all41, have you ever had the swap stop working after failed tests of Hibernate and Suspend? I did once, but haven't been able to replicate. Only time I've been faced with a broken swap so I'm still pretty curious how that could happen

Not that I am aware of. How do you know if that happens? Gparted has always shown my swap partitions. Like right now 4Kb out of 2GB swap is being repoted as used. The system monitor
resource graph rarely shows any swap usage. Actually I have never thought about swap being used for suspend operations

[ofb]

Just check the output of 'free'. If swap is a row of zeros, then it's not active.

And yeah, with al the RAM we have now, it's easy to not notice it's broken.

Um, hit archwiki for a decent explanation of suspend and RAM. Or it was last time I read it. If it's no good say so.
https://wiki.archlinux.org/index.php/Su ... _hibernate

[all41]

Good link--good stuff. After weekend I will install Mint 17 on the system that is now suspending normally with Mint 16.
If it breaks the suspend mode there it will be good info to post.
Thanks for the input! All the best your way

[ofb]

Hey, while you're curious, add these two experiments. On that good 16 box, see if you can H&S the live-ISOs for 16 and 17. Load the proprietary drivers if any detected, and put an unsaved text file on the desktop. I'm curious if the live session behavior matches the installed. It'll be useful to know either way.

H&S is full of mysteries. I've had results like everything looks like it works but i can't click that text file. Last item to look for, if all else is good, is what happens with wifi and wired connections. That can get odd, even on 'working' boxes. Have an Asus laptop that doesn't do wifi until you close the lid briefly. However after you do that Suspend, you have no wired connection till you reboot.

[all41]

Will do

[Cranky_Frankie]

There is one key piece in this thread, the one where it was noted when Mate was the desktop suspend worked. As part of the upgrade to a new chip/mobo/64 bit I switched from Mate to Cinammon. This is one variable I hadn't even considered. I mean, the kernel is the same, why would the interface matter? But apparantly it does. I think I will try going back to Mate and see what happens. If that fixes it I'll be sure to post here.

[rbmorse]

With 64-bit Linux, in order for suspend to work the swap file has to be at least as big as installed RAM plus a little for overhead (32 Mb?). So, on a machine with 16 Gb RAM installed, a 16 Gb swap won't do it. 17 should be just fine.

My machine has 32Gb RAM, my swapfile is 34 Gb and suspend/hibernation work fine.

[all41]

Hey, while you're curious, add these two experiments. On that good 16 box, see if you can H&S the live-ISOs for 16 and 17. Load the proprietary drivers if any detected, and put an unsaved text file on the desktop. I'm curious if the live session behavior matches the installed. It'll be useful to know either way.

ofb,

This is the results:

Mint 16 64bit booted from hard drive--created a text file on desktop and left it unsaved and the file was usable after both S&H
Mint 16 64bit booted live dvd. No proprietary drivers needed--created unsaved desktop text file and it was usable after Suspend. Hibernation not offered as a shutdown menu option.

Mint 17 64bit booted live dvd. Unsaved test file was usable after both S&H. In the suspend mode the power led turned orange and the monitor went into standy. Hibernation was strange-it took a couple minutes and the fans shut down but the power led never went off, changed color or into a blinking state. The diplay went dark but the monitor did not enter standby.
Repeated these steps a few times to make sure my observations were accurate.
Perhaps because there is no persistance?

Ehernet connection good after all tests. This does not have wifi but I have a dongle that I can try.

[ofb]

With 64-bit Linux, in order for suspend to work the swap file has to be at least as big as installed RAM plus a little for overhead (32 Mb?)

That's interesting to hear. What's it based on?

[ofb]

Hibernation not offered as a shutdown menu option.
...
Perhaps because there is no persistance?

Ah, right. 'not offered' just reminded me that it's not offered when there is no swap.

IIRC the live ISO /used/ to use the swap partition on the hard drive if detected. I suppose one should run 'free' in the live session to initally check if a swap is in use. I haven't paid any attention to this with Mint ISO's yet. Sorry. Should have warned you.

Also yes about persistence. Full Hibernation would mean power Off. The session might get written to the hdd swap, but the 're'booted live-session would not remember that it had a session there to restore. 'D'oh.'


EDIT -- incidentally, WinXP H&S has worked flawlessly on the desktops that can't do it in Linux. Have you noticed the same? I'm curious about this because the devs are smart. There must be a reason they don't simply do what XP does, but what on earth could it be?

[rbmorse]

With 64-bit Linux, in order for suspend to work the swap file has to be at least as big as installed RAM plus a little for overhead (32 Mb?)

That's interesting to hear. What's it based on?

A post I found on Ubuntuforums while searching for help with the same problem a while back. Didn't bookmark it.

[all41]

incidentally, WinXP H&S has worked flawlessly on the desktops that can't do it in Linux.

Crap! I think my hd just gave up the ghost. Ater all the back and forth live etc I got an error something like 'error attempt to read ouside of disk 001' ?
But then it finally booted to the installed Mint 16, but after rebooting a couple times it hung on the bios screen. Unplugged it from sata and the computer booted to the second
hard drive which contains Win7. Bios will not load with the 320Gb Linux drive even when connected by itself. I think the HD is just bricked. I am thinking I may have booted live during a hibernation period or something. ?
So I hung a 160Gb to the sata bus and it works. I had previously installed 17 it. Not a screamer and five years old. With all the systems around here storage is in short supply.

[all41]

With 64-bit Linux, in order for suspend to work the swap file has to be at least as big as installed RAM

There is at least one exception.
As you can read I have been experimenting and comparing H&S with installed vs live boot modes, and had a meltdown of sorts.
Up and running with a replacement hd and a freshly installed Mint 17 Mate V2 and this machine handles both H&S flawlessly.
This would probably be classed as older and somewhat resource limited by todays standards. It also H&S fine using Mint 16.
This is my inxi output and as you can see there is 6Gb ram, but only 4Gb Swap:

uno@uno-HP-Compaq-dc5750-Small-Form-Factor ~ $ inxi -F
System: Host: uno-HP-Compaq-dc5750-Small-Form-Factor Kernel: 3.13.0-24-generic x86_64 (64 bit)
Desktop: N/A Distro: Linux Mint 17 Qiana
Machine: System: HP-Pavilion product: AY022AA-ABA p6330f
Mobo: MSI model: IONA version: 1.0 Bios: American Megatrends version: 5.11 date: 01/28/2010
CPU: Dual core Intel Core i3 CPU 530 (-HT-MCP-) cache: 4096 KB flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx)
Clock Speeds: 1: 1200.00 MHz 2: 1200.00 MHz 3: 1467.00 MHz 4: 1200.00 MHz
Graphics: Card: Intel Core Processor Integrated Graphics Controller
X.Org: 1.15.1 drivers: intel (unloaded: fbdev,vesa) Resolution: 1680x1050@60.0hz
GLX Renderer: Mesa DRI Intel Ironlake Desktop GLX Version: 2.1 Mesa 10.1.0
Audio: Card: Intel 5 Series/3400 Series Chipset High Definition Audio driver: snd_hda_intel
Sound: Advanced Linux Sound Architecture ver: k3.13.0-24-generic
Network: Card: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller driver: r8169
IF: eth1 state: down mac: 40:61:86:92:1a:74
Drives: HDD Total Size: 160.0GB (2.9% used) 1: id: /dev/sda model: WDC_WD1600AVBS size: 160.0GB
Partition: ID: / size: 11G used: 4.3G (41%) fs: ext4 ID: /home size: 132G used: 144M (1%) fs: ext4
ID: swap-1 size: 4.00GB used: 0.00GB (0%) fs: swap
RAID: No RAID devices detected - /proc/mdstat and md_mod kernel raid module present
Sensors: None detected - is lm-sensors installed and configured?
Info: Processes: 263 Uptime: 7 min Memory: 344.5/5832.3MB Client: Shell inxi: 1.8.4
uno@uno-HP-Compaq-dc5750-Small-Form-Factor ~ $

[ofb]

This is my inxi output and as you can see there is 6Gb ram, but only 4Gb Swap:

Yup. That'll be because it'll only matter when your RAM is filled with more than can be placed in the Swap.

But it gets ever more interesting. Back to that archwiki page, specifically here:
https://wiki.archlinux.org/index.php/Su ... Ffile_size

2/5ths? So to check the box I'm on right now,

n@glam ~ $ cat /sys/power/image_size
826269696
n@glam ~ $ free
total used free shared buffers cached
Mem: 2048708 1574100 474608 23196 133880 709256
-/+ buffers/cache: 730964 1317744
Swap: 2093052 0 2093052

And yes, 2/5 of 2048708 is 819483.2.

Curiouser and curiouser. Note this is a LVM LUKS full disk encryption install, which means the swap size is decided by the devs who wrote the installer.

Re HDD failure. Bummer. Start a thread for it if you like and I'll join you there to compare troubleshooting ideas.