jsebright

Members
  • Posts

    68
  • Joined

  • Last visited

Everything posted by jsebright

  1. Don't know if things have changed, but when I first started out with Unraid - or when I did some additional playing around - I found that the tool didn't set the keyboard, which caused me issues when using the GUI mode (a temporary thing). Some of us get used to the " and @ keys being in certain places... Still not sure that was ever resolved. Totally get that Unraid is a GUI, and should start off as such with the installer. The options setting in the imager are important in helping new (and old) users avoid the command line, in the same way the browser interface sorts out (nearly?) all the configuration. I also understand that there are very good imaging tools available, so providing a default image for users that want it and understand that it doesn't do the nice things an installer does will satisfy users that prefer their own imager.
  2. I run Unraid headless, and use the default (non-gui) option. Yes 2,4,8,10 should be isolated. I've just checked that the IsolCPUs setting matches the flash setting and it does, so have rebooted. Flash settings are OK. I go to settings / CPU pinning and the banner appears as soon as I do that. Isolated CPUs match the flash settings. I try removing the all cpu pinning in the settings page, check that they've been removed from the flash settings (they have) and reboot. When I go into the cpu pinning settings I no longer get the banner. Might go back to this. I decide to have another go at setting 2,4,8,10. I apply the changes and get the banner. Check the Flash settings and they match up so I reboot again. I go into the dashboard again and don't get the banner - Hooray! I can change docker pinning and still don't get the banner - Hooray! So, all cleared now, but no idea why that was such a problem. I presume it started when I changed motherboard & cpu (going from 6 cores to 12), and I guess something was failing to check. Could it be the orders were different (2,4,8,10 vs 2,8,4,10 ?) - it certainly appeared to be pinning the correct cores, so it was just the checking that was getting it wrong. Anyway, the solution for me (and perhaps for anyone else who gets this) was to clear all pinnning, reboot, then re-apply pinning. Thanks for your help in this.
  3. Thanks for your reply. I think I'll leave it as is, though I might see if I can restrict it to a single disk and move other shares to other drives. Although perhaps I'm trying to be too neat with it. Beats the pain of trying to copy the files to another disk (which seemed problematic when I tried it briefly).
  4. I really like what it's doing, having de-duped backups of a number of systems is great. Thanks @binhex for building and maintaining this docker. Now, a few questions... I am noticing that when a backup is running (full image backup in this case) that urbackup maxes out one cpu core. It alternates every minute or so which one it is, but it 's unusual in doing this. Also, I initially allocated a single disk for backups, outside of the array. Then I thought I might run out of space, so set it to be inside the array instead (avoiding cache). Due to the number of links (I believe) if I run "Compute All" on the shares then the size of the urbackup share can't be calculated. Would I be better off putting this back onto a dedicated disk instead, and is there a good way to move the backups or do I have to start over again?
  5. Thanks both for your posts. It's odd then. The boot settings show the same "isolcpus" as the CPU pinning page. (as did /boot/syslinux/syslinux.cfg) A few days ago I swapped motherboards (& processor), and had gone in to reset the pinnings, saved & rebooted, then got this message. I've just rebooted and had the same problem. Looking at the Flash config I see some old & invalid stub settings in the GUI version which I have removed, and have rebooted again. Gone into that page and it's flashed up the banner again. What else can I check?
  6. As soon as I visit Settings > CPU Pinning I get the banner with the title "CPU Isolation: A reboot is required to apply changes" The banner persists even though I make no changes. It shows itself when logging into the dashboard on another browser. It clears when the server is rebooted. I'm not expecting this banner to show unless I make changes to the "CPU Isolation" section. I should be able to make and apply VM or docker pinning changes without a reboot. Don't think diagnostics should be necessary but including them anyway. Note that I've recently made some hardware changes, but spotted this both before and after the change. tower-diagnostics-20200421-0955.zip
  7. That's the issue, my switches only support Static Link aggregation. Thanks Netgear. I'm connecting 2 ports from my Unraid box to a Netgear GS116Ev2 where the ports can be set for Static LAG. That connects via 2 ports to a Netgear GS908E with LAG set up at both ends. Then into my windows box which I can configure for static LAG. I've been trying to improve the speed (just for the sake of it) without spending too much. Looks like the best way might be to put a card in each machine and use a direct cable run (avoiding the switches), but I've got to use the existing cables as putting in a run of fibre is not feasible. There are some 2.5G cards that have recently come on the market (Amazon didn't have any recently, but they now do) - I might get a pair of those to try. Thanks for replying.
  8. Also interested to know if Static LAG is possible. Most discussions around this sort of thing seem to result in a recommendation to use 10GBe - I'm keen to avoid the cost of this (seems more expensive in the UK).
  9. Sorry, can't remember. Assuming you mean Windows drivers in the guest VM, then I would have updated to the latest ones ASAP. I keep changing things around to suit other family members, and am moving at least one (what will be a heavily used gaming based machine) back to physical. Keep the virtual for remote desktopping into and an occasional, low use KVM machine.
  10. Would love to see a special deal for a limited licence which allows for just a single drive (possibly a parity as well) - I would like to have a second unraid box just for a few dockers and VMs - PFSense mainly.
  11. The test preclear was a success, but at the top of the Log file there is a line (just under Preclear Disk Version: 1.0.6) that says: Disk /dev/sdd is a SSD, disabling head stress test. The disk is not an SSD, but is connected to a cheap PCIE SATA board. I think it's this: [1095:3132] 06:00.0 RAID bus controller: Silicon Image, Inc. SiI 3132 Serial ATA Raid II Controller (rev 01) I'm going to assume the SSD identification won't cause any problems and start another PreClear run. @gfjardim - Thanks for the updates and the excellent support.
  12. I've just had a similar error to Simon above. A new "shucked" WD 6tb drive. Unassigned devices shows: Near the end of the preclear log I see these lines: Nov 24 14:25:56 preclear_disk_WD-WX21D690LD9T_12518: syslog: sd 3:0:0:0: [sdc] Attached SCSI disk Nov 24 14:25:56 preclear_disk_WD-WX21D690LD9T_12518: syslog: WDC_WD60EDAZ-11U78B0_WD-WX21D690LD9T (sdc) 512 11721045168 Nov 24 14:25:56 preclear_disk_WD-WX21D690LD9T_12518: syslog: excluded disks=sda sda2 sdb sdc sdf Nov 24 14:25:56 preclear_disk_WD-WX21D690LD9T_12518: syslog: Command: /usr/local/emhttp/plugins/preclear.disk/script/preclear_disk.sh --notify 1 --frequency 4 --cycles 1 --no-prompt /dev/sdc Nov 24 14:25:56 preclear_disk_WD-WX21D690LD9T_12518: syslog: Disk /dev/sdc is a SSD, disabling head stress test. Note that the disk is not an SSD. Only thing I can think of is that during the zeroing run I added a USB drive to the server (an old drive in the donor USB case). That seemed to pause the zeroing whilst it formatted, but is working as a UD attached disk. Preclear logs attached. What do I do now. I was going to use it as a replacement parity disk, but wanted to give it a try out. Is it unreasonable to just go ahead and swap it with the existing parity? Thanks, Jon. TOWER-preclear.disk-20191124-1934.zip
  13. That worked for me. Thanks. Been having issues for a few days, trying multiple installs with issues when the driver installs. Right pain in the wotsit. Also using Seabios - that's what I was on when I tried the Q35, not going to change either if it works.
  14. I'm not currently running PiHole, but I am running an adblocker on my router. Have just tried switching that off along with suspending adblock and ublock in the browser, but it makes no difference. Have tried a few browsers - the only one to show anything is Edge (and IE) on Windows which show "Undefined" so something is happening at the client end. If you look at "System Stats" you will see network activity (as long as you're not in a VM on the Unraid server!), so it's likely it's just the display widgets which are broken client side.
  15. I think this is broken at the moment. Just tried uninstalling and reinstalling the container, but when running the test the speeds don't change and the IP Address is blank. In W10 on Edge, the speed counters and IP show "undefined". When I click "Start" there's definitely network traffic though.
  16. Agree. I'm fairly new to unraid and remote access is great, but I'm just evaluating whether I want to run it and use it as a desktop machine hosting several VMs, and I'm stuck because I can't set the keyboard. It would be great if it was just a setting to change the layout of the connected keyboard, but also would be helpful if it was part of the usb installer to set it right from the beginning. [I've installed nerdpack, kbd, and typed "loadkeys uk " but it doesn't appear to change anything]