nagi

+1, currently my only method would be to create a completely separate cache pool and limit a share to that specific pool, with zero feedback as to how much space is left of the allowance. (Or unassigned but mounted & shared device of the specific desired size.) Granted, that also would allow me to e.g.: forego array protection if the data is not important at all, but it also robs me of any flexibility in reconfiguring it easily.

Hey everyone, Despite a lot of googling, did not find any result for iommu group breakdown. So I bought it, and here it is, with IOMMU Enabled in the latest stable 1.50 bios: tl;dr: good groups, except for 2 of the USB controllers being in the same one. Group 0: [1022:1482] 00:01.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 1: [1022:1483] [R] 00:01.2 PCI bridge Starship/Matisse GPP Bridge Group 2: [1022:1482] 00:02.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 3: [1022:1482] 00:03.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 4: [1022:1483] [R] 00:03.1 PCI bridge Starship/Matisse GPP Bridge Group 5: [1022:1482] 00:04.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 6: [1022:1482] 00:05.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 7: [1022:1482] 00:07.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 8: [1022:1484] [R] 00:07.1 PCI bridge Starship/Matisse Internal PCIe GPP Bridge 0 to bus[E:B] Group 9: [1022:1482] 00:08.0 Host bridge Starship/Matisse PCIe Dummy Host Bridge Group 10: [1022:1484] [R] 00:08.1 PCI bridge Starship/Matisse Internal PCIe GPP Bridge 0 to bus[E:B] Group 11: [1022:790b] 00:14.0 SMBus FCH SMBus Controller [1022:790e] 00:14.3 ISA bridge FCH LPC Bridge Group 12: [1022:1440] 00:18.0 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 0 [1022:1441] 00:18.1 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 1 [1022:1442] 00:18.2 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 2 [1022:1443] 00:18.3 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 3 [1022:1444] 00:18.4 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 4 [1022:1445] 00:18.5 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 5 [1022:1446] 00:18.6 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 6 [1022:1447] 00:18.7 Host bridge Matisse/Vermeer Data Fabric: Device 18h; Function 7 Group 13: [1022:57ad] [R] 01:00.0 PCI bridge Matisse Switch Upstream Group 14: [1022:57a3] [R] 02:02.0 PCI bridge Matisse PCIe GPP Bridge Group 15: [1022:57a3] [R] 02:04.0 PCI bridge Matisse PCIe GPP Bridge Group 16: [1022:57a3] [R] 02:06.0 PCI bridge Matisse PCIe GPP Bridge Group 17: [1022:57a3] [R] 02:07.0 PCI bridge Matisse PCIe GPP Bridge Group 18: [1022:57a4] [R] 02:08.0 PCI bridge Matisse PCIe GPP Bridge [1022:1485] [R] 07:00.0 Non-Essential Instrumentation [1300] Starship/Matisse Reserved SPP [1022:149c] 07:00.1 USB controller Matisse USB 3.0 Host Controller <= 3.2gen2 type A above type-c USB: [05e3:0608] Bus 001 Device 002 Genesys Logic, Inc. Hub USB: [1d6b:0002] Bus 001 Device 001 Linux Foundation 2.0 root hub USB: [1d6b:0003] Bus 002 Device 001 Linux Foundation 3.0 root hub [1022:149c] [R] 07:00.3 USB controller Matisse USB 3.0 Host Controller <= 2.0 controller USB: [0781:5583] Bus 003 Device 004 SanDisk Corp. Ultra Fit USB: [1a2c:0021] Bus 003 Device 003 China Resource Semico Co., Ltd Keyboard USB: [26ce:01a2] Bus 003 Device 002 ASRock LED Controller USB: [1d6b:0002] Bus 003 Device 001 Linux Foundation 2.0 root hub USB: [1d6b:0003] Bus 004 Device 001 Linux Foundation 3.0 root hub Group 19: [1022:57a4] [R] 02:09.0 PCI bridge Matisse PCIe GPP Bridge [1022:7901] [R] 08:00.0 SATA controller FCH SATA Controller [AHCI mode] Group 20: [1022:57a4] [R] 02:0a.0 PCI bridge Matisse PCIe GPP Bridge [1022:7901] [R] 09:00.0 SATA controller FCH SATA Controller [AHCI mode] Group 21: [1b4b:9123] [R] 03:00.0 SATA controller 88SE9123 PCIe SATA 6.0 Gb/s controller <= x1 #1 Group 22: [197b:0585] [R] 04:00.0 SATA controller JMB58x AHCI SATA controller <= x1 #3 Group 23: [197b:0585] [R] 05:00.0 SATA controller JMB58x AHCI SATA controller <= wifi card slot Group 24: [10ec:3000] [R] 06:00.0 Ethernet controller Killer E3000 2.5GbE Controller Group 25: [10de:128b] [R] 0a:00.0 VGA compatible controller GK208B [GeForce GT 710] <= x16 #1 [10de:0e0f] 0a:00.1 Audio device GK208 HDMI/DP Audio Controller Group 26: [1022:148a] [R] 0b:00.0 Non-Essential Instrumentation [1300] Starship/Matisse PCIe Dummy Function Group 27: [1022:1485] [R] 0c:00.0 Non-Essential Instrumentation [1300] Starship/Matisse Reserved SPP Group 28: [1022:1486] [R] 0c:00.1 Encryption controller Starship/Matisse Cryptographic Coprocessor PSPCPP Group 29: [1022:149c] [R] 0c:00.3 USB controller Matisse USB 3.0 Host Controller <= below LAN port USB3 + between ps2+typeC USB: [1d6b:0002] Bus 005 Device 001 Linux Foundation 2.0 root hub USB: [1d6b:0003] Bus 006 Device 001 Linux Foundation 3.0 root hub Group 30: [1022:1487] 0c:00.4 Audio device Starship/Matisse HD Audio Controller Used this script for the output. Note: I'm extremely new to this part of UnRaid. USB: From what I can see with my very beginner eyes, unfortunately it seems that the USB controllers are all on the same group 15 except for the one in group 23. The two USB devices you see (Keyboard + SanDisk Ultra Fit) are in the supposedly 2.0 slots. The separated group 23 is the USB 3 pair below the NIC LAN port ('USB 3.2 Gen1 USB3_34' in manual) AND the ones between the PS2 and the type-c containing group (USB3.2 Gen1 USB3_12). The board has 2x USB2, 2x USB3.2 gen1 and one type-c header. The 2.0 is probably on the same controller as the rear IO ones (and thus should not be passed through if you keep to the Unraid standard of "use 2.0") and same with the type-c header. However, since the 3.2 headers are for gen1, those probably belong to the separately grouped controller. I have not yet tested this, as I don't have the cable/adapter for it. So for all the marketing, the "Lightning Gaming Ports" actually share a group with another controller, lol. PCIe slots: I'm still very early in experimentation. I just put in a 2230 SATA controller (JMB58x as per lspci) into the empty wifi card slot and even that showed up separately. Same for the first and last X1 slots The wifi card slot is unfortunately right next to one of the x1 slots, so you'll have to use a riser for one of them if you want to use absolutely all the slots. Headless: It will post headless. Bifurcation: BIOS (well, the shipped 1.3, did not check with the newest 1.5) had 4*4 / 8+2*4 / 8+8 options for the top, CPU connected PCIe16 slot (the rest of the slots are through the chipset, 3 x1 and 2 x16, though the latter are x4+x2, plus 2 M.2 (Gen4x4), but using PCIe device in the second one will disable 2 of the 6 onboard SATA ports) Weirdness: One weird issue I was having is that the BIOS seems to be incompatible with my "AverMedia Live Streamer Cap 4K" + GT710 combo: until I plugged in an actual display, the BIOS would just reset. First time I ever saw something like that. The combo works *after* initializing with a dummy plug or a display and changing it out to the capture combo, at least for Unraid console. Test config: (just playing around) 3900X + 2x4GB DDR4 + random cards I had around. Future plans: 2x windows10 VMs, so I'll need at least 2 GPUs, and I'll also put in a 2-port intel 10Gbit NIC. Hopefully SR-IOV will work.

Just adding my own experiences, sry for the necro, maybe useful for someone else... port flapping With the same X520-DA2 combined with mikrotik S+AO00005 cable to a Mikrotik CRS305-1G-4S+IN used as a switch Jan 28 21:54:04 Storage kernel: bond0: link status definitely down for interface eth0, disabling it Jan 28 21:54:53 Storage kernel: ixgbe 0000:65:00.0 eth0: NIC Link is Up 10 Gbps, Flow Control: RX/TX Jan 28 21:54:53 Storage kernel: bond0: link status definitely up for interface eth0, 10000 Mbps full duplex Jan 28 21:54:55 Storage kernel: ixgbe 0000:65:00.0 eth0: NIC Link is Down Jan 28 21:54:55 Storage kernel: bond0: link status definitely down for interface eth0, disabling it Port flapping occured in one of the ports, the other I think just didn't log it, as it sure wasn't working, and while the switch showed link, unraid reported it as down. fw version: # ethtool -i eth0 driver: ixgbe version: 5.3.6 firmware-version: 0x80000231 expansion-rom-version: bus-info: 0000:65:00.0 supports-statistics: yes supports-test: yes supports-eeprom-access: yes supports-register-dump: yes supports-priv-flags: yes Exhibited the same problem in another PC, (old Z79 based) but not in a third (Z99). The unRaid one is a X299 based newer machine. Very weird compatibility issue.

Yeah, I know it is possible to only assign e.g.: the second to a VM, my question is more about assigning the rest of the cores of the other CCX as a seemingly separate physical CPU, as the hypervisor certainly does not need a whole 4-core CCX, and the extra cores - if utilized correctly - would certainly be welcome.

Hi, weird question: is there a way for the KVM layer to make it seem to a Windows VM as if the physically single CPU was actually 2 roughly half-sized physical CPUs? So basically take a single 8-core CPU, (take 1 off for the virtualization layer,) and pass what appears to be a separate 3 and 4 core CPU to the Windows 7 Pro VM? The question arose while I was erading up on Ryzen, because the current R7 CPUs are basically 2 separate CPUs (each called a CCX, separate cache, same system memory) on the same PCB.

Hi, I've already an unRAIDstorage server, so I'm not entirely unfamiliar with the software, but I've only used basic stuff, and have only recently upgraded to 6.2 from 5.RCsomethingIcan'tremember. However, I'm thinking about using unraid to virtualize my current office/gaming/movie/etc desktop, and I have a number of pretty technical question about it. PC in question: 4790K (has VT-d, integrated GPU will be used for unRAID) 24GB DDR3 AsROCK Z87 Extreme4 (has VT-d) Asus GTX1080 Founder's Edition Windows 7 x64 (already installed) 1. I (will shortly) have two 750GB SSDs that I want to use as mirrored system disks for a Windows 7 setup. I do not want to use the Windows built-in partition-mirroring, instead I would like to use unRAID to provide some raid protection. Is this possible? 2. Would this setup screw with the garbage collection / TRIM of the drives? I'd rather not have perpetually dropping performance. 3. Is there a way to leave the last e.g. 50GB of the SSDs out of the partitions (as extra overprovisioning)? 4. How much memory would unraid need in this setup? (only 1 VM running, everything passed through, unraid only doing raid to two SSDs) My desktop has 24GB of RAM, and I'd like to use most of it. The The PC has other disks, but AFAIK those can just be passed through to the VM. There won't be any traditional data drives for unraid to handle other than the drives that will be passed through. PS: yes, I realize I might have just jumped into the deep end. I have some experience, but it is in Windows desktops, not linux and virtualization. I did some research, but kvm-linux.org is currently 503-ing.

OK, I went to the shares page and checked sizes. It turns out it was the "system" share, with the docker.img file using the whole nine yards of that 22GB. So thanks to all who helped in resolving this "issue". I've already copied over 1.6TB of my 11... Oh boy, this is gonna take a while

Ah, that would explain it. Thanks.

Well, while I left the defaults on, I did not install or activate any docker / VM images. Does an "empty" one really consume that much space? Obviously at 16TB I'm not concerned about a mere 22GB, but I jsut want to make sure that the new array is working properly before fully transitioning from my older one.

Hi, I created a completely new new unraid array, on a completely new pendrive and system, went through preclears, started, then formatted the array with two of the three drives. My problem is that somehow 22GB space is being used on my first disk. Why? Anyone has any idea?

tl;dr: PSU problem. Hi, I've had a metric ton of problems with my unRAID box. Fortunately it turned out that it was just the PSU was aging and it could no longer support all the HDDs in the same 12V lane. I'm posting this so that others who might google for this can get one more search result. So, problems: -write errors -hangups -parity disk getting disabled -disks getting frozen -links getting reset -reiserfs i/o failures -cannot complete a parity resync without getting so many errors the sync stops and the parity disk gets disabled Oct 17 03:42:31 Storage kernel: end_request: I/O error, dev sdf, sector 77024392 (Errors) Oct 17 03:42:31 Storage kernel: md: disk0 write error (Errors) Oct 17 03:42:31 Storage kernel: handle_stripe write error: 77024328/0, count: 1 (Errors) Oct 17 03:42:31 Storage kernel: md: disk0 write error (Errors) Oct 17 03:12:47 Storage kernel: ata6.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0 (Errors) Oct 17 03:12:47 Storage kernel: ata6.00: irq_stat 0x40000001 (Drive related) Oct 17 03:12:47 Storage kernel: ata6.00: failed command: WRITE DMA EXT (Minor Issues) Oct 17 03:12:47 Storage kernel: ata6.00: cmd 35/00:00:c8:54:0e/00:04:00:00:00/e0 tag 0 dma 524288 out (Drive related) Oct 17 03:12:47 Storage kernel: res 51/10:00:c8:54:0e/00:04:00:00:00/e0 Emask 0x81 (invalid argument) (Errors) Oct 17 03:12:47 Storage kernel: ata6.00: status: { DRDY ERR } (Drive related) Oct 17 03:12:47 Storage kernel: ata6.00: error: { IDNF } (Errors) Oct 17 03:12:47 Storage kernel: ata6.00: configured for UDMA/133 (Drive related) Oct 17 03:12:47 Storage kernel: ata6: EH complete (Drive related) Oct 17 03:24:09 Storage kernel: ata4.00: exception Emask 0x12 SAct 0x0 SErr 0x400501 action 0x6 frozen (Errors) Oct 17 03:24:09 Storage kernel: ata4.00: irq_stat 0x08000000, interface fatal error (Errors) Oct 17 03:24:09 Storage kernel: ata4: SError: { RecovData UnrecovData Proto Handshk } (Errors) Oct 17 03:24:09 Storage kernel: ata4.00: failed command: READ DMA EXT (Minor Issues) Oct 17 03:24:09 Storage kernel: ata4.00: cmd 25/00:00:30:b9:f0/00:04:01:00:00/e0 tag 0 dma 524288 in (Drive related) Oct 17 03:24:09 Storage kernel: res 50/00:00:2f:b9:f0/00:00:01:00:00/e0 Emask 0x12 (ATA bus error) (Errors) Oct 17 03:24:09 Storage kernel: ata4.00: status: { DRDY } (Drive related) Oct 17 03:24:09 Storage kernel: ata4: hard resetting link (Minor Issues) Oct 17 03:24:10 Storage kernel: ata4: SATA link up 6.0 Gbps (SStatus 133 SControl 300) (Drive related) Oct 17 03:24:10 Storage kernel: ata4.00: configured for UDMA/133 (Drive related) Oct 17 03:24:10 Storage kernel: ata4: EH complete (Drive related) Oct 17 04:20:10 Storage kernel: REISERFS error (device md6): vs-13070 reiserfs_read_locked_inode: i/o failure occurred trying to find stat data of [15 17 0x0 SD] (Errors) Oct 17 04:20:10 Storage kernel: REISERFS error (device md6): vs-13070 reiserfs_read_locked_inode: i/o failure occurred trying to find stat data of [15 66 0x0 SD] (Errors) All these repeating for pages and pages and pages and pages. I thought my WD RED 3TBs were dying, but they were just dying for a bit more juice. The solution was actually simple, as it turned out: I put 3 of my 7 disks on another 12V lane, and the problems went away. Yes, the more permanent solution will be getting another PSU. (I'm looking at the Seasonic X-400, as I have another and I've had very good experience with it, and it can actually supply up to 600W, and has one 12V lane only.)

Hi guys, I just got a "Delock 89270 PCI Express Card - 2 x internal SATA 6Gb" It has a Marvel 88SE9128 chipset, PCIE 2.0x1, 2 internal SATA ports and a bundled low profile bracket. It cost me 38 €, but at least it has 3 years of warranty. It seems to be working right out of the box, without any mod/script/etc to my unRAID installation. I've only put a single 3TB WD RED on it, which was precleared on the motherboard controller. The speed seems to be in the same ballpark as with the MB controller, which is 35-45MB/s sustained. Note that I do not have an i7 CPU, only a 1st gen AMD Fusion A8-3870. Anyways, hdparam -tT gives ~145MB/sec, so the controller seems to be A-OK for the time being. (Oh, and I'm still running 5.0-rc8a, lazyness, you see.)

Just bought: Sharkoon REX8 Economy http://www.sharkoon.com/?q=en/node/2338 8 usable 5.25" drive spaces pics: http://www.sharkoon.com/sites/default/files/products/pc_cases/REX8_ECO_black_01_1.jpg http://www.sharkoon.com/sites/default/files/products/pc_cases/REX8_ECO_black_03.jpg Problems: 12cm fans can only be installed to the front with modding the case. This is because the front panel doesn't have enough space, and the rail that the drives are put on is a bit short for both using the drives and the fan brackets at the same place. If you interchange them, you are good to go, but that is a bad joke from a supposedly 8 drive case. Lucky for me, I use rubber bands to keep the disks in place (maybe the best for silencing and getting rid of resonation, but taking in and out the drives is tricky a bit) so I could The fan brackets offer too much resistance: they have a stylized 8cm fan cut out, but the screw placements are only for 12 cm fans. So you get a LOT of dead space. I plan on fixing this with a dremel... The fan and the 5.25" -> 3.5" brackets (of which you only get 4) are the same, meaning too much extra stuff on both. Plus, the quality is sub-par. Some edges seem broken. (They are not, it is just shoddy engineering.) The brackets are made of plastic. Good for getting rid of some vibration, bad for reusability and rigidness. The screws have to be driven into the material as well. (Could have been done with encased steel nuts) I had to use pliers to screw in the motherboard standoffs. (another quality issue: badly drilled holes) Although the side panel is enlarged sideways to accomodate for wire management behind the scenes, it is still hard to shut the door. The doors are a bit flimsy, very thin. No top 12cm fan placement option. There are two stability enhancing crossbars on the front which do not line up with the fans if you install them, but could be OK with a 3x5.25" drive bay, provided you get rid of the internal flaps. Good things: Has 8 usable 5.25" slots. (the top houses the USB hubs and thus uses it up) Space for 2 12cm or a 14cm fans on the side (economy only, the value has a plexi door) With these said, I only paid about $63 for it in Hungary. (Which is quite cheap here.) Note: my second (first being the Rebel9 Pro Economy) Sharkoon case. I also have a Antec P180b (which I love) from old times. I have to say that despite marketing about German precision, Sharkoon is a cheap-as-chips solution that might be passable, but can hold no candle to any serious contender.