[SOLVED] 5.0 wake not working


Lleot

Recommended Posts

I recently decided to upgrade my 4.7 unraid server to 5.0.

Followed the instructions, re-added my users, did the permission fixer etc all seems fine.

Previously sleep/wake/wol worked perfectly.

 

Now I send the server to sleep using

echo -n mem >/sys/power/state

 

It goes to sleep fine.

Now I send WOL packet and the power comes back on.

 

But nothing else works! I connected my TV to the server and the video doesn't come back.

I can't connect to it using telnet/smb/internet.

I press the off button on the server and it doesn't turn off the server. Had to do a hard reset.

 

I tested turning it on via the power button and by removing the keyboard and they are both the same so I don't think the problem is with wol. It seems to go to sleep very quickly - not sure if it is a problem with the sleep or the wake. I wrote a script that would do cp /var/log/syslog /boot after calling the sleep command to see if I could capture the syslog, but it didn't seem to work. Any better way of capturing the syslog just before sleep/after wake?

 

I really need sleep to work. Do I have to downgrade back to 4.7? Can I move between versions without redoing the permissions fixer + removing all my users?

Link to comment

I have the same problem. Here is the hardware:

Motherboard: Gigabyte GA-D510UD

NIC: Realtek RTL8111D

No software plug-ins or customization with a 16GB flash and one 2TB HD.

 

The sleep and wakeup has been working for years until I updated to the final UnRaid release 5.0. Sleep was working with 4.7 and all the 5.0 RC up to RC12aAiO. The computer will go to sleep and wakeup without any problems or need of rebooting at all.

 

I installed the final V5.0 a week ago. I put to sleep using the command "echo -n mem >/sys/power/state" and send a magic packet to wakeup. The server wakeup but it is frozen, it will not even respond to a ping and the server show on my Asus router as offline. The only thing left is to turn it OFF by holding the power switch for 4 seconds. I tried many things like starting without shares or any customizations. No success, same problems. I reverted to RC12aAiO and everything is working fine.

 

As suggested, I made a backup of my flash drive. Made it easy to go back to last known good software version.

 

Link to comment

I wrote a script that would do cp /var/log/syslog /boot after calling the sleep command to see if I could capture the syslog, but it didn't seem to work. Any better way of capturing the syslog just before sleep/after wake?

http://lime-technology.com/forum/index.php?topic=29389.msg263858#msg263858

 

 

What does this show?

cat /sys/power/state

 

Thanks for reply.

 

root@Tower:~# cat /sys/power/state
freeze standby mem

 

I attached syslog, but I'm gonna try safe mode to see if it makes any difference.

 

syslog_2013-09-22_22.02.55.txt

Link to comment

The sleep and wakeup has been working for years until I updated to the final UnRaid release 5.0. Sleep was working with 4.7 and all the 5.0 RC up to RC12aAiO. The computer will go to sleep and wakeup without any problems or need of rebooting at all.

I suspect that RC14 would work for you also.  The RC's up through RC12a used the Linux kernel version 3.4.* (or earlier).  RC13 was the first to use Linux 3.9.*, but that's when a ReiserFS problem was discovered, so RC14 reverted to 3.4.* (see Kernel and File Versions).  Then Tom patched the kernel and RC15 and all versions since have returned to Linux 3.9.*.

 

I attached syslog ...

 

Took a look, noticed lines like the following:

Sep 22 21:59:24 Tower kernel: Atom PSE erratum detected, BIOS microcode update recommended    (multiple occurrences)

Sep 22 21:59:24 Tower kernel: Performance Events: PEBS fmt0+, LBR disabled due to erratumAtom events, Intel PMU driver.

Sep 22 21:59:24 Tower kernel: pci 0000:00:03.2: System wakeup disabled by ACPI    (numerous occurrences)

I do not know if these items are significant or even relevant to you, and they may not be, but they have the potential to be the cause of your issues with Sleep/Wake.  It looks like someone has determined that your BIOS microcode may be buggy with respect to this Atom CPU, so they have decided to disable certain features of it, possibly features necessary to Waking.  This may not have been discovered when the earlier kernels were released, which could explain why earlier versions worked for you (lots of speculation here, sorry).  Certainly the many statements of "System wakeup disabled" are highly suspicious!

 

What would help is if you could provide for comparison syslogs from previous versions, where Sleep/Wake worked correctly.  Do you still have some, and can attach them?

Link to comment

I suspect that RC14 would work for you also.  The RC's up through RC12a used the Linux kernel version 3.4.* (or earlier).  RC13 was the first to use Linux 3.9.*, but that's when a ReiserFS problem was discovered, so RC14 reverted to 3.4.* (see Kernel and File Versions).  Then Tom patched the kernel and RC15 and all versions since have returned to Linux 3.9.*.

 

I'm on RC16 and have a similar situation and would like to give this a try, what are the steps to downgrading? Also running simple features, will the downgrade process cause any problems?

 

thanks.

Link to comment

I have this same issue, and I also have "System wakeup disabled by ACPI" even though I have S3 sleep support.

 

Sleeping the unRAID server works fine, but waking up results in an unresponsive server, with no SMB shares mounted, and even a keyboard plugged directly into the box still results in no signal to the monitor. I basically have to hard reset the server at this point.

 

I'll try rolling back to RC 14... are there any issues I need to be aware of in downgrading?

Link to comment

I'll try rolling back to RC 14... are there any issues I need to be aware of in downgrading?

Only the whole list of changes in the Release Notes, from RC14 all the way to 5.0-final.  For one example, around RC14 a bug in reiserfs was discovered that could lead to data loss.  But if such things don't bother you, then you'll be OK.  :)

 

 

Link to comment

The sleep and wakeup has been working for years until I updated to the final UnRaid release 5.0. Sleep was working with 4.7 and all the 5.0 RC up to RC12aAiO. The computer will go to sleep and wakeup without any problems or need of rebooting at all.

I suspect that RC14 would work for you also.  The RC's up through RC12a used the Linux kernel version 3.4.* (or earlier).  RC13 was the first to use Linux 3.9.*, but that's when a ReiserFS problem was discovered, so RC14 reverted to 3.4.* (see Kernel and File Versions).  Then Tom patched the kernel and RC15 and all versions since have returned to Linux 3.9.*.

 

I attached syslog ...

 

Took a look, noticed lines like the following:

Sep 22 21:59:24 Tower kernel: Atom PSE erratum detected, BIOS microcode update recommended    (multiple occurrences)

Sep 22 21:59:24 Tower kernel: Performance Events: PEBS fmt0+, LBR disabled due to erratumAtom events, Intel PMU driver.

Sep 22 21:59:24 Tower kernel: pci 0000:00:03.2: System wakeup disabled by ACPI    (numerous occurrences)

I do not know if these items are significant or even relevant to you, and they may not be, but they have the potential to be the cause of your issues with Sleep/Wake.  It looks like someone has determined that your BIOS microcode may be buggy with respect to this Atom CPU, so they have decided to disable certain features of it, possibly features necessary to Waking.  This may not have been discovered when the earlier kernels were released, which could explain why earlier versions worked for you (lots of speculation here, sorry).  Certainly the many statements of "System wakeup disabled" are highly suspicious!

 

What would help is if you could provide for comparison syslogs from previous versions, where Sleep/Wake worked correctly.  Do you still have some, and can attach them?

 

Spent an evening trying to update the microcode but I'm pretty stuck. Looks like there's a module missing which means it doesn't work in unRaid. modprobe microcode doesn't work, according to this link:

http://www.linuxquestions.org/questions/linux-from-scratch-13/updating-cpu-microcode-in-lfs-4175433777/

The kernel configuration should have the following to build the microcode update driver as a module (see them in "Processor type and features"). Recompile the kernel if it is not configured to do the microcode update.

 

Not sure what to try next. Will probably have to request to add microcode updating to the unraid kernel for the next version and cross fingers.

 

Don't have a syslog for 4.7 but do have one for 4.5.6, not sure if that helps.

 

 

 

syslog2.txt

Link to comment

Do you know if this is a motherboard compatibility issue or a NIC issue?

 

I'm debating on whether or not to go buy a network card that is known to "work" so I don't have to troubleshoot this anymore.

 

I doubt it's a network issue. Try connecting a keyboard to the machine and using that to make it wake.

 

For me I have been too busy to deal with downgrading so I just turning my server off instead of sending it to sleep.

 

Edit: I found out theres something in linux called pm_trace which can trace the sleep https://wiki.ubuntu.com/DebuggingKernelSuspend but it seems not to be included with the unRaid kernel. If I had more time I would rebuild with it and test it out (http://lime-technology.com/wiki/index.php/Building_a_custom_kernel).

Link to comment
  • 4 weeks later...

hello,

I have the same problem. Here is the hardware:

Motherboard: Gigabyte GA-D525TUD

NIC driver :  r8168

No software plug-ins or customization with a 4GB flash and four 2TB HD.

 

i've tried many versions of RC 5.x but it 's always bad with the same result : i must stay 4 seconds on power button to reset the server.

perhaps i would be better to change the motherboard, which one would be the best ?

 

Link to comment
  • 2 weeks later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.