maspiter

Thanks for the reply, will have to do some testing myself I guess when I get the chance. As per the electricity costs, my server consumes only 23 watts with drives spun down yet I still find it very much worth it to save some energy. After all it doesn't need to be available 90% of the time and it consumes under 10% of minimum online energy when in S3. Being able to put it in standby is actually one of the main reasons I chose unRAID.

Would love to hear some experiences about this too ...

If that is the case, then that would be a good idea for sure. Unfortunately haven't got the time to fix this myself nor the knowledge. I do know the SNAP plug-in uses this kind of setup so maybe some work can be taken from there?

Necessary PHP extensions (i.e. SOAP) enabled in php.ini?

Ok, I added in my changes. I also took the liberty to move the configuration section to the top and moved the $noTCP and $debug variables to the configuration section for ease of use and added a few comments about determining which value to use for $noTCP for your specific system/use case. They might be obvious but thought I'd add them anyway. You talking to me? If you're asking about the custom kernel, then sure, I am willing to share. Except I doubt it will do you any good because I only added in the firmware for my specific GPU (HD6310). I guess it should work for HD6250, HD7310 and HD7340 too (according to the method I used, https://wiki.gentoo.org/wiki/Radeon#External_resources) but these are all notebook GPUs. But, I am by no means an unRAID nor linux expert. I just happen to have just enough knowledge to follow a few guides here and there. On top of that, my bzroot is smaller than the stock one because I stripped out some stuff I don't need, mostly drivers for ethernet cards I am not using I believe. That means your ethernet will probably only work when using a Realtek (I know) card. So, as a disclaimer, using my files is at your own risk. I cannot be held responsible for data loss or any other unfortunate events that may or may not be tied to using them. etc etc ... That said, I've had no issues so far and recently even replaced a disk with a larger one flawlessly. If you're still interested send me a PM with your email or something because one file is too big to post on the forum. EDIT 1: I might have another version with all the drivers included. It is twice the size of the stock version though. You can add them to the syslinux.cfg file but better keep the stock as default. This way you can reboot and test it out without committing. Used this one for a bit but compiled another one afterwards apparently. When I have some time I'm going to redo the whole process and maybe put a more descriptive guide online if there is some interest. The wiki article is a bit basic I think. EDIT 2: I uploaded the files so others might benefit without having to contact me, don't know how long these will be available though. Unraid 5.0.5 with radeon HD6310 and some drivers removed (what I'm using now): http://www.files.com/set/53fcda459510f Unraid 5.0.5 with radeon HD6310 and stock drivers: http://www.files.com/set/53fcddb19096c I'm using an E35M1-I BTW. auto_s3_sleep.txt

Kind of late but just got around to testing that script. After excluding my disk outside the array from the HDD activity check and adjusting the idle network activity value it seems to work. (With the logging enabled I could check which idle value to use for my system.) It also has a feature to prevent sleep/shutdown with a lock file which I like. Gonna try this script out for a while now.

@Fireball3 Thanks for your script, I'm currently using it because the Dynamix script doesn't seem to take network activity into account. I have a suggestion though. Since I'm still using the free version and use a drive outside the array to host some apps, it might be a good idea to add in support to exclude a drive. I added a $nonArrayDisk variable myself and added an extra "| grep -v $nonArrayDisk" to filter it out the check_HDD_activity() but there is probably a more elegant way to do this. I also noticed a comment regarding checking for a cache drive but no commands are actually in place I think? @dopray You can try adding it to the section pre_sleep_activity or post_sleep_activity in the script on the previous page. If it is working with that script, the Dynamix script might not execute your settings in the webGUI. Have you checked this with a trivial command like creating a file i.e. and after wake-up check whether it is actually created? I had the same problem. Tested a full Slackware install on the same system and sleep worked fine so it was not a hardware problem. Tried many different things but in the end compiled my own custom kernel with added drivers for my AMD GPU. Since then, it sleeps and wakes just fine. It's a lot of work, certainly since you have to redo it each time you want to upgrade but IMO it's worth it. There's a guide in the wiki on compiling a custom kernel, one of my earlier posts also links to references I used. Hope this helps.

Been using the sleep functionality for a while, but am facing another issue. When using the network idle threshold value my server never goes to sleep even though its network activity seems to stay below the threshold (according to system stats). Anyone else having this issue? What is the best way to troubleshoot this? See here http://lime-technology.com/forum/index.php?topic=13019.msg123528#msg123528

Intel NICs are usually recommended. Make sure to buy a PCI-E card when going this route, PCI will get saturated with a single gigabit NIC. Did you get sleep working?

Tried that and a bunch of other options, didn't work for me.

I had the same problem, running an AMD zacate board with a Realtek NIC. Got it working now under 5.0.5 even with r8169 loaded. In my case, I believe it was because there was no driver for the GPU. Since you are using a Llano CPU which also has an embedded AMD GPU if I'm not mistaken, I assume it might be the same problem. I'm no linux expert at all obviously, but I had the same symptoms and error messages, even in the syslog. You can search my posts if you like. I tried compiling a kernel with error reporting enabled (or something like that) to find out what device was causing the issue, tried various utilities and sleep command line options, all to no avail. What I did (had to do) is compile my own kernel with driver support for the GPU. Of course this loads a whole bunch of other stuff because of dependencies but it has been running flawless so far. I haven't had the time to fine tune it unfortunately. Again, more info in my previous posts. Also, there is (was?) a problem with the Dynamix sleep plugin when you're running the free version (without cache) and rely on a disk outside the array ( i.e. mounted with SNAP). --> Fixed with latest version (2.1.1). Of course, again, no linux expert and would love to get it working without it ...

That seems to have done the trick. With the adjustment, I get the excluded disk form fields and applying my settings reloads the page seemingly without error. Only thing that's left now is the network idle field that doesn't keep its value. Edit: Thank you for getting to the bottom of this. Now I can start testing the script with various settings.

Sure, output in order of your commands: /dev/sda [empty line] /dev/sda|usb-Kingston_DT_Micro_1C6F654FF40FBC41591108DC-0:0 /dev/sdb|ata-WDC_WD20EFRX-68AX9N0_WD-WMC300052677 /dev/sdc|ata-WDC_WD20EARX-00PASB0_WD-WMAZA9666863 /dev/sdd|ata-SAMSUNG_HD103SJ_S246J9KB425239 /dev/sde|ata-WDC_WD3200BEVT-75ZCT1_WD-WXE708HN0888 /dev/sdb /dev/sdc /dev/sdd As I'm running the free version, there is no cache drive.

No problem, thanks for looking into it. I do have a disk outside the array (see attached screenshot) and all other values stick.

The output of the command is one empty line. This is the content of dynamix.s3.sleep.cfg: service="1" checkHDD="-a" timeout="30" interval="3" port="eth0" checkTCP="" idle="0" login="" pingIP="" stopDay="" stopHour="" dhcpRenew="" forceGb="" setWol="" preRun="" postRun="" debug="1" exclude="" excludeList="" options="-C 1 -a -m 30 -e eth0 -D 1" Since I'm running a custom kernel, I tried booting the stock unRAID kernel but get the same results, hidden input fields and error message. I did reinstall all my plugins after switching to the new kernel and copied my old plugins folder to /boot/config/plugins_old so I could compare settings or files if necessary but the dynamix subfolder is deleted. But this should not make a difference I believe. The first time I installed through the dynamix plugin control plugin, afterwards I reinstalled it manually a couple of times. I'll reinstall after deleting the _old folder and test again. EDIT: No difference unfortunately.