kamhighway

Setting the spindown parameter for the ssd to never resulted in no further episodes. However, upon reboot, the same set of error messages appeared in the log. Since the rsync script does not run at bootup, I am concluding that the rsync script has nothing to do with these error messages. It seems that when UR first connects to the drive, whether upon boot or spinup, it tries to execute a command on the ssd that fails, then retries a bunch of times, and them fails. After that, as long as the ssd does not "spin down" there are no further error messages. It doesn't look like it affects the operation of the cache drive. But still, these error messages worry me.

@RobJ Thanks again. I think I'm going to make 1 change at a time until I can isolate the cause. As it seems that the issue occurs on the SSD when the server is accessed after all the drives have been spun down, the first thing I'll try is setting the SSD to never spin down. I don't know if an SSD can "spin down" anyway. Maybe that's the task that the log shows attempting to abort.

@RobJ Thank you for looking at my logs and taking the time to write your reply. At 22:00 each night I have a script that rsyncs my "can't afford to lose" shares to another unraid server. From the timestamps, it looks like the problem is triggered at the end of the rsync. The last line of the script basically uses an ipmi command to turn off the backup server. As for the timestamps at 6:50 am, 2:02 am, and 7:56 am, they look like times I was accessing the UR server after all the disks had spun down. The cache drive is a PNY 480Gb formated as BTRFS SSD. The only change I've made recently is updating to Unraid 6.1.9 from 6.1.8. I don't recall seeing these messages with 6.1.8. Is there anything in 6.1.9 that might explain these error messages? I think I will stop my backup script and see if that clears up the events happening during the 22:00 hour.

@aptalca, I assumed the guide was for your dockers since the link was posted in this support thread. I see now that I was wrong. Thanks for pointing that out. Saved me some time following a guide that was not for the docker I wanted to install.

This set of messages appears in my log file roughly twice a day. I am running Unraid v6.1.9 on a Supermicro x9scm-f with a xeon 1240 and 16 GBs of ram and a Dell H310 that has been flashed to IT mode. Mar 30 01:38:47 Media kernel: mdcmd (184): spindown 0 Mar 30 01:38:47 Media kernel: mdcmd (185): spindown 5 Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8805de88aa80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 00 0a 30 a0 00 00 20 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8805de88aa80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8805de88b080) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 06 9e 70 98 00 00 08 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8805de88b080) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8805de88ac00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 06 ce de a8 00 00 18 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8805de88ac00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8805de88be00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 06 ce de c8 00 00 10 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8805de88be00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8805de88ad80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 06 d6 bc 48 00 00 20 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8805de88ad80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0300) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 18 5d 9d 10 00 00 08 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0300) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0900) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 28 69 b4 20 00 00 10 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0900) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0f00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 1b c8 00 03 b0 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0f00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d1c80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 1f 80 00 00 d8 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d1c80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d1980) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 20 60 00 00 e0 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d1980) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d1380) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 21 48 00 00 80 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d1380) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d1080) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 21 d0 00 02 18 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d1080) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0c00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 23 f0 00 00 10 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0c00) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0000) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 24 20 00 00 88 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0000) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0d80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 24 b0 00 00 10 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0d80) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: attempting task abort! scmd(ffff8800678d0780) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: [sdh] CDB: opcode=0x28 28 00 31 34 24 d0 00 00 18 00 Mar 30 02:02:05 Media kernel: scsi target1:0:1: handle(0x000a), sas_address(0x4433221101000000), phy(1) Mar 30 02:02:05 Media kernel: scsi target1:0:1: enclosure_logical_id(0x590b11c03b4adc04), slot(2) Mar 30 02:02:05 Media kernel: sd 1:0:1:0: task abort: SUCCESS scmd(ffff8800678d0780) Here are the full log files: media-syslog-20160330-0822.zip

Thank you for your writeup. I found it quite nice to have a start to finish explanation of how everything works together. There is no doubt that things will change over time, but for those of us trying to get this implemented now, your guide is much easier to follow than the having to piece it all together from 7 pages of posts. All the best.

@cyruspy Think you might like this one: http://www.ebay.com/itm/SuperMicro-Motherboard-X9SCM-F-LGA1155-IPMI-SATA3-1u-Heatsink-6xSata-Cables-/222050604545?hash=item33b33f5601:g:rM0AAOSwIwhWT3dZ I bought one from this seller last year and it works great.

I've been using the methods I learned from speeding_ant on this thread for years: http://lime-technology.com/forum/index.php?topic=10274.0

I have the BPN-SAS-846EL2 SAS expander and find that it works well. The backplane has a sas connector on it so if your motherboard has sata ports on it, you need a reverse breakout cable to connect the backplane to 4 sata ports on the motherboard. The backplane connects the 4 sata ports to 24 drives. As long as you are not accessing more than 4 drives connected to the backplane at the same time, there is no performance hit. For this reason, I don't put my parity drive on the backplane. There is a bracket you can buy that attaches a 3.5 HD to the side of the power supply. See this link: http://www.ebay.com/itm/191819144865?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT The 846 case can take two of them. If your MB has 6 sata ports, connect 4 to the backplane and use the other two to connect parity drives mounted to the side of the power supply. About the only time the backplane becomes a bottleneck is during parity checks. This is because all the drives are being accessed at the same time. Still, it works amazingly well.

Unraid will run quite nicely on that board. I have run unraid for years on an X7SPA-HF-D510 which is an older version of your board. I upgraded to get the VM features of unraid v6 which require Vt-d -- the atom's don't have this. However, for running unraid as a NAS and for running a couple of dockers that are not cpu intensive (not plex) this board will be fine.

@hernandito Here is a screenshot of the page from the manual for our board.

@hernandito I have my board in a supermicro 4u sc846 case. Inside the case, there is one big connector that goes on top of the entire block of pins for the front panel. It makes it very easy to set up, but it means I don't need to know which pin is the power + and which is the power - pin. Looks like some of the other forum members have posted some more helpful info for you though.

I have the exact same board. You do NOT have to connect the dedicated IPMI port to your router/switch. BTW, the dedicated IPMI port is the one above the USB ports. However, I've found that the dedicated port is nice because if a VM or a docker completely goes berserk and you lose ethernet, the dedicated port is unaffected so you can still get in and power cycle the computer. There is a menu in the bios settings that is all about the settings for the IPMI functionality. I think it might be labeled "BMC settings." That's where you can assign an IP address to access your IPMI features.

@hernandito You did not say which board you have. But if its a SuperMicro, I have some experience with it. IPMI does much more than monitor temperatures. When its turned on a SuperMicro board will acquire two IP addresses. One is for the board to be able to access the internet as you expect. The second is to access a webpage where you can interact with the IPMI functionality of your board. Many SM boards come with a NIC that is dedicated to IPMI. It is usually a Realtek NIC, while the others are Intel. If you connect this port up to your switch, then you'll always be able to access the webpage at the IPMI ports IP address so you'll be able to power cycle the machine even if the computer is otherwise non-responsive. From this page, you can also monitor the boot process over the internet and even enter the bios setup screen to change settings. SM also supplies IPMIView, an app that enables you to access your board's IPMI features if you don't like the webpage. I know other MB manufacturers offer IPMI functionality, but I have no experience with them.

@StevenD That is exactly why I thought anyone who is thinking about using one of these cases in a home should be made aware of the noise level.

@MetDeath, I bought a used 4U supermicro case for my build and although it is well-built, the 5 80mm fans inside the case and the 2 40mm fans in each power supply are loud. I have a space under the stairs in the middle of my house where I want to put my server. However, the noise level is so high that if I put it there, I can hear it throughout the house. I am thinking about replacing the two supermicro power supplies with a standard ATX seasonic, and the 5 80mm fans with something quieter. It seems that SC846 series cases are not made to be quiet because system admins seem to think this is normal. If it is quieter than a hair dryer going full blast, they deem it to be quiet.

@TUMS, That makes sense. That would explain why the ddr3 memory is being dumped since the motherboards for the v3 xeons use ddr4. When the Westmere xeons were dumped, was there a corresponding dump of memory DIMMs?

@garycase I'm thinking the same thing. Only this time, I might buy the motherboard while they are still in production and relatively cheap. But that would only make sense if I was sure about what CPU Facebook is using now.

@TUMS Do you know what CPUs Facebook bought to replace these E5s?

I understand all these CPUs are coming from Facebook. What happened to the motherboards they use to be in at Facebook?

@Drider Thank you for confirming that S3 sleep works on the asrock motherboard. This is important to me.

@TUMS, I think server boards don't sleep because servers are supposed to be on 24x7. Lime Technology has given the same reason in the past as to why Unraid does not support sleep. These CPU's draw a lot of power, even at idle, so I really want my work station to go to sleep when I'm not using it. For me, this means the supermicro x9 series boards are not the best choice. I would really like to hear from someone with the ASRock board if the bios has a setting to enable s3 sleep?

@TUMS I agree with your choice of supermicro for a server. But for a workstation, I want a motherboard that can go to sleep -- which the supermicro boards cannot do. I've been looking at an Asus z9pe-d8 ws to use as a my primary workstation. I would love to know if the ASRock server boards can go to sleep?

@sonofdbn Apparently, the version labeled SROHB does not support vt-d. But, the ones labeled SR0KX do support vt-d. It looks like the information you found on cpu-world is accurate. If you want to run VMs, be sure to get the SR0KX version.

@hernandito That board will work. I have a very similar one. http://www.newegg.com/Product/Product.aspx?Item=N82E16813182349R It seems that v1 and v2 xeons can use the same socket, but the v3 versions need a different socket though they still call it a 2011 which is confusing.