Jump to content

mgutt

Moderators
 View Profile  See their activity

  • Posts

    11,355

  • Joined

  • Last visited

  • Days Won

    124

 Content Type 

Everything posted by mgutt

  1. mgutt
    Mit welchem Befehl hast du den Server in den Ruhezustand geschickt? Hast du im BIOS und über das WebTerminal geprüft, dass WoL aktiv ist? Hier sind ein paar Tipps wie man WoL testen kann: https://wiki.unraid.net/Setup_Sleep_(S3)_and_Wake_on_Lan_(WOL) Wegen dem automatischen Herunter- und Hochfahren kämen mir diese Lösungen in den Sinn: Skript + BIOS Power On By RTC Man lässt jede Nacht um 2 Uhr ein Skript starten, das alle 5 Minuten prüft ob alle Festplatten schlafen und wenn ja, wird der Server heruntergefahren. Im BIOS stellt man dann bei Power On By RTC einfach ein, dass er den Rechner täglich um 7 Uhr einschalten soll. Skript + smarte Steckdose + BIOS Power On @ Power Loss Wie zuvor lässt man das Skript den Server herunterfahren. Im BIOS muss bei Power Loss die Option "Power On" gewählt werden. Dh sobald das Board wieder Strom bekommt, fährt es automatisch hoch. Das erledigen dann zwei simple Regeln der smarten Steckdose: A) wenn der aktuelle Verbrauch unter 3W liegt (oder was auch immer der Server im abgeschalteten Zustand verbraucht), Steckdose abschalten und B) um 7 Uhr Steckdose einschalten
  2. mgutt
    Hast du mal in der Fritz!Box geschaut ob die VMs, denen du eine feste IP zugeordnet hast, wirklich die Mac-Adressen der VM verwenden und nicht die vom Unraid Server?
  3. mgutt
    This should not help. The script sorts by file modification time and not the year in the filename (that was the idea I mentioned in my last post) All files, renamed by this bulk renamer tool, will probably receive a newer file modification time and so pushed to the top of the preloading list.
  4. mgutt
    This should be possible. Maybe I could also add an option which allows sorting by the year which is inside the parent folder name like "2004" in "/Lost (2004)/S01E01.mkv". But by that recently added content which has an old airing date, won't be cached.
  5. mgutt
    Schau in den App Store vom Panasonic. Jellyfin, Emby oder Plex verfügbar? Dann würde ich eines davon nehmen. Für DLNA geht auch Plex, da Plex zusätzlich alle Filme nach Genre, Titel usw in getrennten DLNA-Ordnern verfügbar machen kann. Ich sage dir aber gleich, dass DLNA fast immer wegen Codecs nervt. Musst du einfach mal ausprobieren. Plex, Emby und Jellyfin sind alle gleich "mächtig". Plex gilt am komfortabelsten. Wenn DLNA doof ist und die anderen nicht erwünscht, bleibt noch Kodi auf einem Android Stick / Box deiner Wahl. Viele packen per Sideload Kodi auf einen Amazon Fire Stick. In dem Fall wird gar keine Software auf dem Server installiert, sondern Kodi greift auf die Filme über SMB zu. Ich persönlich mag Kodi gar nicht, weil die Einrichtung nervt und es irre aufwendig ist das so eingerichtet zu bekommen, dass man mal eben eine Serie auf dem Smartphone weiter schauen kann.
  6. mgutt
    This error is not important, but of course I will try to fix this. Preloading is done only to available / free RAM. You can't see or measure the preloaded data. If you re-execute the script you should get a notification with the amount of files that were cached / skipped. If all files were skipped, you can be sure that the preloading works. Which HDDs are you using? This completely depends on the spinup speed. It seems your HDDs are a little bit slow in spinup compared to mine. The slowest HDDs I know of (in spin up) are Ironwolfs (those hated WD SMR drives are freaking fast): I made this test by spinning down the HDD and copying a 500MB file through Teracopy (Windows) to the HDD. Teracopy finally shows the transfer time which is a good value to compare the latency. Important: You need to wait 1 to 2 minutes after a spindown or the HDD is still spinning on your next test = spins up faster than usual.
  7. mgutt
    Overflow is only valid for Shared Paths. A Direct Path (/mnt/cache) does not overflow to the array. This means, if the cache becomes full and you use /mnt/cache, the Plex Container will produce errors, because it is out of storage. That is the reason why I suggest in my guide to set "Min. free space" in your Global Share Settings. Example: 1TB SSD cache /mnt/cache is your SSD /mnt/disk1 is your array /mnt/user/Movies uses disk1 and uses the cache setting "yes" Scenario 1 Plex appdata is set to /mnt/user/... You upload more than 1TB of movies to your server. After occupying 1TB of your SSD, further movie files are written to the array. As the SSD is full, Plex writes to the array Scenario 2 Plex appdata is set to /mnt/cache... You upload more than 1TB of movies to your server. After occupying 1TB of your SSD, further movie files are written to the array. As the SSD is full, Plex produces errors Scenario 3 You set 100GB Min. Free Space Plex appdata is set to /mnt/cache... You upload more than 1TB of movies to your server. After occupying 900GB of your SSD, further movie files are written to the array. As there are still 100GB left, Plex works as usual
  8. mgutt
    This is the WebGUI of my backup NAS: As you can see it claims all disks are sleeping. But that's not true: smartctl -n standby -i /dev/sdb smartctl -n standby -i /dev/sdc smartctl -n standby -i /dev/sdd smartctl -n standby -i /dev/sdf smartctl -n standby -i /dev/sdg smartctl -n standby -i /dev/sdh They return all: Power mode is: ACTIVE or IDLE "/dev/sde" (the SSD) is the only sleeping disk: smartctl -n standby -i /dev/sde smartctl 7.1 2019-12-30 r5022 [x86_64-linux-4.19.107-Unraid] (local build) Copyright (C) 2002-19, Bruce Allen, Christian Franke, www.smartmontools.org Device is in STANDBY mode, exit(2) Only for testing I executed this command: mdcmd spindown 1 Now disk 1 sleeps: smartctl -n standby -i /dev/sdh smartctl 7.1 2019-12-30 r5022 [x86_64-linux-4.19.107-Unraid] (local build) Copyright (C) 2002-19, Bruce Allen, Christian Franke, www.smartmontools.org Device is in STANDBY mode, exit(2) Spin down delay is set to 30 minutes: Fun fact: It worked for 2 weeks without problems until I: - opened this page: http://tower/Tools/SysDevs - then this page: http://tower/Tools/HardwareProfile - clicked on "Show Details", which spun up the disks - this took a little bit long, so I opened http://tower/Main - again http://tower/Tools/HardwareProfile - again clicked on "Show Details" and now I was able to see hardware informations - as it did not contain the brand of my RAM I executed "dmidecode -t memory", which did not display the brand name, but I found now the RAM part number - I closed the WebGUI and that's it It seems that clicking on "Show Details" killed the Cronjob to check the status of the disks?! Update, 2020-01-25 Now I'm having the same problem on a complete different Unraid server, too. A manual execution of the smartctl command (that is used by Unraid to check the standby status), returns the correct value: "sdg" is in STANDBY mode root@Thoth:~# smartctl --nocheck standby -A /dev/sdg smartctl 7.1 2019-12-30 r5022 [x86_64-linux-4.19.107-Unraid] (local build) Copyright (C) 2002-19, Bruce Allen, Christian Franke, www.smartmontools.org Device is in STANDBY mode, exit(2) "sdl" is spinning: root@Thoth:~# smartctl --nocheck standby -A /dev/sdl smartctl 7.1 2019-12-30 r5022 [x86_64-linux-4.19.107-Unraid] (local build) Copyright (C) 2002-19, Bruce Allen, Christian Franke, www.smartmontools.org === START OF READ SMART DATA SECTION === SMART Attributes Data Structure revision number: 16 Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000b 100 100 016 Pre-fail Always - 0 2 Throughput_Performance 0x0005 132 132 054 Pre-fail Offline - 96 3 Spin_Up_Time 0x0007 159 159 024 Pre-fail Always - 412 (Average 414) 4 Start_Stop_Count 0x0012 100 100 000 Old_age Always - 3195 5 Reallocated_Sector_Ct 0x0033 100 100 005 Pre-fail Always - 0 7 Seek_Error_Rate 0x000b 100 100 067 Pre-fail Always - 0 8 Seek_Time_Performance 0x0005 128 128 020 Pre-fail Offline - 18 9 Power_On_Hours 0x0012 100 100 000 Old_age Always - 6711 10 Spin_Retry_Count 0x0013 100 100 060 Pre-fail Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 86 22 Helium_Level 0x0023 100 100 025 Pre-fail Always - 100 192 Power-Off_Retract_Count 0x0032 095 095 000 Old_age Always - 6414 193 Load_Cycle_Count 0x0012 095 095 000 Old_age Always - 6414 194 Temperature_Celsius 0x0002 157 157 000 Old_age Always - 38 (Min/Max 24/55) 196 Reallocated_Event_Count 0x0032 100 100 000 Old_age Always - 0 197 Current_Pending_Sector 0x0022 100 100 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0008 100 100 000 Old_age Offline - 0 199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 0 But the dashboard doesn't show it AND sdl does not spindown anymore altough inactive for more than 30 minutes (my spindown setting): So I started monitoring smartctl executions by this: while true; do pid=$(pgrep 'smartctl' | head -1); if [[ -n "$pid" ]]; then ps -p "$pid" -o args && strace -v -t -p "$pid"; fi; done I tested it by executing smartctl with a different terminal and it catches it directly. Then I waited a long time and several errors were returned: Broken Pipe errors (happen often) strace: Process 8668 attached 13:53:01 ioctl(3, SG_IO, {'S', SG_DXFER_NONE, cmd[16]=[85, 06, 2c, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, e5, 00], mx_sb_len=32, iovec_count=0, dxfer_len=0, timeout=60000, flags=0, status=02, masked_status=01, sb[22]=[72, 01, 00, 1d, 00, 00, 00, 0e, 09, 0c, 00, 00, 00, ff, 00, 00, 00, 00, 00, 00, 00, 50], host_status=0, driver_status=0x8, resid=0, duration=0, info=0x1}) = 0 13:53:01 ioctl(3, SG_IO, {'S', SG_DXFER_FROM_DEV, cmd[16]=[85, 08, 0e, 00, 00, 00, 01, 00, 00, 00, 00, 00, 00, 00, ec, 00], mx_sb_len=32, iovec_count=0, dxfer_len=512, timeout=60000, flags=0, data[512]=[40, 00, ff, 3f, 37, c8, 10, 00, 00, 00, 00, 00, 3f, 00, 00, 00, 00, 00, 00, 00, 39, 31, 35, 30, 42, 41, 30, 38, 31, 32, 39, 31, ...], status=00, masked_status=00, sb[0]=[], host_status=0, driver_status=0, resid=0, duration=1, info=0}) = 0 13:53:01 brk(0x584000) = 0x584000 13:53:01 brk(0x5a7000) = 0x5a7000 13:53:01 brk(0x5c8000) = 0x5c8000 13:53:01 ioctl(3, SG_IO, {'S', SG_DXFER_FROM_DEV, cmd[16]=[85, 08, 0e, 00, d0, 00, 01, 00, 00, 00, 4f, 00, c2, 00, b0, 00], mx_sb_len=32, iovec_count=0, dxfer_len=512, timeout=60000, flags=0, data[512]=[04, 00, 05, 32, 00, 64, 64, 00, 00, 00, 00, 00, 00, 00, 09, 32, 00, 64, 64, 1a, 22, 00, 00, 00, 00, 00, 0c, 32, 00, 64, 64, 7f, ...], status=00, masked_status=00, sb[0]=[], host_status=0, driver_status=0, resid=0, duration=8, info=0}) = 0 13:53:01 ioctl(3, SG_IO, {'S', SG_DXFER_FROM_DEV, cmd[16]=[85, 08, 0e, 00, d1, 00, 01, 00, 01, 00, 4f, 00, c2, 00, b0, 00], mx_sb_len=32, iovec_count=0, dxfer_len=512, timeout=60000, flags=0, data[512]=[04, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, ...], status=00, masked_status=00, sb[0]=[], host_status=0, driver_status=0, resid=0, duration=0, info=0}) = 0 13:53:01 write(1, "=== START OF READ SMART DATA SEC"..., 41) = 41 13:53:01 ioctl(3, SG_IO, {'S', SG_DXFER_NONE, cmd[16]=[85, 06, 2c, 00, da, 00, 00, 00, 00, 00, 4f, 00, c2, 00, b0, 00], mx_sb_len=32, iovec_count=0, dxfer_len=0, timeout=60000, flags=0, status=02, masked_status=01, sb[22]=[72, 01, 00, 1d, 00, 00, 00, 0e, 09, 0c, 00, 00, 00, 00, 00, 00, 00, 4f, 00, c2, 00, 50], host_status=0, driver_status=0x8, resid=0, duration=8, info=0x1}) = 0 13:53:01 write(1, "SMART overall-health self-assess"..., 57) = 57 13:53:01 write(1, "\n", 1) = -1 EPIPE (Broken pipe) 13:53:01 --- SIGPIPE {si_signo=SIGPIPE, si_code=SI_USER, si_pid=8668, si_uid=0} --- 13:53:01 close(3) = 0 13:53:01 exit_group(0) = ? 13:53:01 +++ exited with 0 +++ strace: Process 20585 attached 14:09:01 brk(0x5c7000) = 0x5c7000 14:09:01 ioctl(3, SG_IO, {'S', SG_DXFER_FROM_DEV, cmd[16]=[85, 08, 0e, 00, d0, 00, 01, 00, 00, 00, 4f, 00, c2, 00, b0, 00], mx_sb_len=32, iovec_count=0, dxfer_len=512, timeout=60000, flags=0, data[512]=[10, 00, 01, 0b, 00, 64, 64, 00, 00, 00, 00, 00, 00, 00, 02, 05, 00, 84, 84, 60, 00, 00, 00, 00, 00, 00, 03, 07, 00, a0, a0, 9b, ...], status=00, masked_status=00, sb[0]=[], host_status=0, driver_status=0, resid=0, duration=5, info=0}) = 0 14:09:01 ioctl(3, SG_IO, {'S', SG_DXFER_FROM_DEV, cmd[16]=[85, 08, 0e, 00, d1, 00, 01, 00, 01, 00, 4f, 00, c2, 00, b0, 00], mx_sb_len=32, iovec_count=0, dxfer_len=512, timeout=60000, flags=0, data[512]=[10, 00, 01, 10, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 02, 36, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 03, 18, 00, 00, 00, 00, ...], status=00, masked_status=00, sb[0]=[], host_status=0, driver_status=0, resid=0, duration=3, info=0}) = 0 14:09:01 write(1, "=== START OF READ SMART DATA SEC"..., 41) = 41 14:09:01 ioctl(3, SG_IO, {'S', SG_DXFER_NONE, cmd[16]=[85, 06, 2c, 00, da, 00, 00, 00, 00, 00, 4f, 00, c2, 00, b0, 00], mx_sb_len=32, iovec_count=0, dxfer_len=0, timeout=60000, flags=0, status=02, masked_status=01, sb[22]=[72, 01, 00, 1d, 00, 00, 00, 0e, 09, 0c, 00, 00, 00, 00, 00, 00, 00, 4f, 00, c2, 00, 50], host_status=0, driver_status=0x8, resid=0, duration=1, info=0x1}) = 0 14:09:01 write(1, "SMART overall-health self-assess"..., 57) = 57 14:09:01 write(1, "\n", 1) = -1 EPIPE (Broken pipe) 14:09:01 --- SIGPIPE {si_signo=SIGPIPE, si_code=SI_USER, si_pid=20585, si_uid=0} --- 14:09:01 close(3) = 0 14:09:01 exit_group(0) = ? 14:09:01 +++ exited with 0 +++ Unable to detect device t (happens every now and then) strace: Process 3649 attached strace: [ Process PID=3649 runs in x32 mode. ] strace: [ Process PID=3649 runs in 64 bit mode. ] 13:57:01 read(3, "8:0x61b6)\n \"ST(250|320|500|64"..., 4096) = 4096 13:57:01 read(3, " \"-v 12,raw48,Device_Power_Cyc"..., 4096) = 4096 13:57:01 read(3, " 9,minutes\"\n },\n { \"Maxtor Dia"..., 4096) = 4096 13:57:01 read(3, " // it might need"..., 4096) = 4096 13:57:01 read(3, "3[015]\",\n \"\", \"\", \"\"\n },\n {"..., 4096) = 4096 13:57:01 read(3, " \"(Hitachi )?HDT7210((16|25)SL"..., 4096) = 4096 13:57:01 read(3, " },\n { \"Toshiba 2.5\\\" HDD MK..4"..., 4096) = 4096 13:57:01 read(3, "04ACA500/FP1A\n \"TOSHIBA MD04A"..., 4096) = 4096 13:57:01 read(3, " \"\", \"\", \"\"\n },\n { \"Seagate"..., 4096) = 4096 13:57:01 read(3, "\"\", \"\"\n },\n { \"Seagate Barracu"..., 4096) = 4096 13:57:01 read(3, "T3000DM001\",\n \"\", \"\",\n \"-v"..., 4096) = 4096 13:57:01 read(3, "Constellation ES (SATA 6Gb/s)\", "..., 4096) = 4096 13:57:01 read(3, " \"\", \"\", \"\"\n },\n { \"Seagate M"..., 4096) = 4096 13:57:01 read(3, "\"-v 187,raw48,Uncorrectable_ECC_"..., 4096) = 4096 13:57:01 read(3, "bly explained by the WD firmware"..., 4096) = 4096 13:57:01 read(3, "viar Green\", // tested with WDC "..., 4096) = 4096 13:57:01 read(3, "/82.00A82,\n // WDC WD80EFAX"..., 4096) = 4096 13:57:01 read(3, "SB ID entries\n ////////////////"..., 4096) = 4096 13:57:01 read(3, " \"USB: Samsung; \",\n \"0x04e8:0"..., 4096) = 4096 13:57:01 read(3, "at\"\n },\n // Micron\n { \"USB: M"..., 4096) = 4096 13:57:01 read(3, " { \"USB: Maxtor OneTouch 4; \",\n"..., 4096) = 4096 13:57:01 read(3, " \"\", // 0x0114\n \"\", // 0x0"..., 4096) = 4096 13:57:01 read(3, " usbjmicron\"\n },\n { \"USB: Verb"..., 4096) = 3817 13:57:01 read(3, "", 4096) = 0 13:57:01 close(3) = 0 13:57:01 lstat("/dev/", {st_mode=S_IFDIR|0755, st_size=3920, ...}) = 0 13:57:01 write(1, "/dev/: Unable to detect device t"..., 36) = 36 13:57:01 write(1, "Please specify device type with "..., 47) = 47 13:57:01 write(1, "\nUse smartctl -h to get a usage "..., 41) = 41 13:57:01 exit_group(1) = ? 13:57:01 +++ exited with 1 +++ And those spindown / standby checks appeared (shortened) 14:40:01 openat(AT_FDCWD, "/dev/sdb", O_RDONLY|O_NONBLOCK) = 3 14:40:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 14:53:44 openat(AT_FDCWD, "/dev/sdg", O_RDONLY|O_NONBLOCK) = 4 14:53:44 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 14:55:01 openat(AT_FDCWD, "/dev/sdj", O_RDONLY|O_NONBLOCK) = 3 14:55:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 14:59:01 openat(AT_FDCWD, "/dev/sdd", O_RDONLY|O_NONBLOCK) = 3 14:59:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:02:01 openat(AT_FDCWD, "/dev/sdk", O_RDONLY|O_NONBLOCK) = 3 15:02:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:02:01 openat(AT_FDCWD, "/dev/sdh", O_RDONLY|O_NONBLOCK) = 3 15:02:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:11:01 openat(AT_FDCWD, "/dev/sdg", O_RDONLY|O_NONBLOCK) = 3 15:11:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:11:01 openat(AT_FDCWD, "/dev/sdi", O_RDONLY|O_NONBLOCK) = 3 15:11:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:11:01 openat(AT_FDCWD, "/dev/sdb", O_RDONLY|O_NONBLOCK) = 3 15:11:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:11:01 openat(AT_FDCWD, "/dev/sdf", O_RDONLY|O_NONBLOCK) = 3 15:11:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:30:01 openat(AT_FDCWD, "/dev/sdd", O_RDONLY|O_NONBLOCK) = 3 15:30:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 15:33:01 openat(AT_FDCWD, "/dev/sdk", O_RDONLY|O_NONBLOCK) = 3 15:33:01 write(1, "Device is in STANDBY mode, exit("..., 35) = 35 In the last hour it did not check sdc, sdl and sde. Why? And why are these checks done in a random order and with a different amount of disks?
  9. mgutt
    You are right, at least we should test it.
  10. mgutt
    @T0a Consider using powertop or my script: Reason: https://www.reddit.com/r/homelab/comments/cvs8ue/my_7_watt_homelab/ Thanks to @Miidnight
  11. mgutt
    Check command exists: https://stackoverflow.com/a/33297950/318765 Check library exists: https://serverfault.com/a/513938/44086
  12. mgutt
    No, because we still aren't able to control which data is moved to the swap. And finally I think its isolated from the host system (where Plex is reading the data from). This is the explanation when Linux writes to Swap: https://tldp.org/LDP/tlk/mm/memory.html 1.) This refers only to used RAM. But the script does not use RAM. It uses only the Buffer Cache (available or "free" RAM). And those data is never written to the Swap. 2.) The only solution to really reserve the RAM by preloading a file would be vmtouch. But I don't know how to force Linux to move this reserved data to the Swap 3.) Someone mentioned, that cgroups are a solution to have isolated settings for a specfic process. This means it seems to be possible to set for example "100MB" RAM for the process "vmtouch" and after it exceeds this value it should write to swap (theoretically). But I was not able to find the right commands to test this scenario. 4.) I was able to install vmtouch, but it does not work very well. Sometimes it does not output the cached data and sometimes the locking daemon process crashes. I would say its not reliable under Slackware which is a huge problem for this project as its the only tool (I know of) which is able to lock data in the RAM.
  13. mgutt
    @diggforbeer @dada051 I still was not able to find a solution to move files, that are cached in the RAM to the swap partition/file. This website claims the Swap is used if the RAM is full, but I was not able to provoke writing directly to the Swap. As long this is not possible, we can not use a SSD
  14. mgutt
    @Morrtin Hast du überhaupt die iGPU für Plex freigeschaltet? Normal sollte die nämlich 4K schaffen. Natürlich vorausgesetzt du hast den Plex Pass und HW Transcoding aktiviert (sieht man dann übrigens im Plex Dashboard am "(hw)" hinter dem Video-Codec)
  15. mgutt
    Genau das geht nicht, weil Unraid keine Treiber für die iGPU mitbringt. Oder hat sich da was geändert?
  16. mgutt
    AMDs waren schon immer Stromschlucker. Die einzige Ausnahme bilden die aktuellen mobilen Prozessoren. Nur B550 kann man mit BIOS Mod sparsam machen. Ansonsten warten alle auf einen sparsamen X570 "Nachfolger". Messgerät? Meine Empfehlung wäre ein ELV Basic oder falls eine Fritz Box vorhanden ist, dann eine DECT 200. Bloß kein zu günstiges. Die messen meist nur Blödsinn (Scheinleistung, statt der korrekten Wirkleistung). Bei deinem Vergleich musst du aber die Festplatten vom Server abstecken und ohne hochfahren, denn die würden bei beiden gleich viel verbrauchen. Ich tippe mal auf 15W der Intel und 40W der AMD. Allerdings nur, wenn du den Intel auch optimiert hast (was Windows bereits ist): https://forums.unraid.net/topic/98070-reduce-power-consumption-with-powertop/
  17. mgutt
    Also dein PC ging nie in den Standby? Zwischen Standby und Leerlauf liegen bei dir denke ich mal 30W. Ich mache meinen auch nie aus, aber der ist halt im Standby, wenn ich den Rechner verlasse. Mit Unraid wäre die Situation eine andere. In dem Fall würde der PC inkl GPU durchgehend Strom verbrauchen. Egal ob die VM läuft oder nicht. Also @ChrisV, ist der Stromverbrauch egal, dann wäre das evtl eine Option. Falls du aber denkst, dass zwei Geräte weniger verbrauchen als eines, ist diese Annahme meiner Ansicht nach falsch. Außer dein aktueller Server ist ein Stromfresser, aber dann hat der schlicht die falschen Komponenten verbaut. Jemand bei Reddit hat behauptet, dass das geht, als ich sagte, dass es nicht ginge. Meine Frage nach dem "wie" hat er nicht beantwortet. Im Forum gibt es niemanden, der es geschafft hat: https://forums.unraid.net/topic/91379-ryzen-3400g-vega-11-gpu-passthrough/ https://forums.unraid.net/topic/83849-ryzen-igpu-passthrough/ https://forums.unraid.net/topic/86713-amd-3400g-passtrough-integrated-gpu/ https://forums.unraid.net/topic/72524-ryzen-apu-igpu-passthrough/ Du kannst die iGPU übrigens nicht mal für Unraid nutzen. Also die iGPU Plex zuordnen, kann man auch vergessen. Wenn man also einer VM eine dGPU zuordnen möchte, muss man Unraid erstmal GPU-less konfigurieren. Viele machen das meine ich mit einer GT710, mit der Unraid eingerichtet wird und die sie dann am Ende wieder abstecken. Ist auf jeden Fall kein 08/15 Setup. Mach dich also auf ein paar Hürden gefasst. Danach wird es aber laufen. Allerdings werden nicht alle Spiele funktionieren, da manche wegen DRM oder Anti-Cheat den Betrieb in einer VM unterbinden. Auch das gilt es also zu berücksichtigen. In meinen Augen ist so ein Setup nur für jemanden, der sich keine Hardware kaufen will und nur einen Rechner zu Hause hat und dem der Stromverbrauch egal ist.
  18. mgutt
    This is a usual setup, used by many people. So yes, this would work. But consider: - the GPU will draw power, even if the VM is shutdown - it could be possible that some games do not work in a VM because of DRM or Anti-Cheat - 4 cores are ok for a VM, but if other people in your household upload files to the server, this could have a negative impact to your VM (as Unraid uses the cores, too)
  19. mgutt
    Live-Transcoding should use the preloaded file part as well. This is independent from the resolution or bitrate as the source file is everytime the same. You could verify it as follows: Execute the following to clean the RAM: sync; echo 1 > /proc/sys/vm/drop_caches Spindown disks and wait 1 minute to be sure that the HDDs does not rotate anymore. Start a movie through your mobile client without Wifi (to force transcoding) and measure the delay. After that clean the RAM again: sync; echo 1 > /proc/sys/vm/drop_caches Execute the preload script. Spindown disks, wait 1 minute. Start a movie that was preloaded. This time the movie should start faster which proofs that even transcoding uses the preloaded data.
  20. mgutt
    Hmm need to check this error 😕 EDIT: Ok, seems to be no problem. It's only because I interrupt the loop when enough videos have been preloaded. Will try to fix this someday.
  21. mgutt
    As long as your default spin down time in the Unraid settings.
  22. mgutt
    dirty_writeback is the only risky setting, but it's not raised by powertop, it's reduced. The default is 30 seconds. Powertop sets it to 15 seconds. This means Linux starts 15 seconds earlier to empty the RAM to disk. This can save energy as all devices can reach the sleep state 15 seconds earlier. Note: A UPS is absolutely recommend or all data of the last 15 seconds will be lost. Unraid default: Even 30 seconds. Example: You move important data to your server (so it's "deleted" from your client). If now a power outage occurs in the 30 seconds since the first file was uploaded, everything is lost. I think this point needs more attention through Limetech. vm.dirty_ratio should be reduced to 1% and vm.dirty_writeback_centisecs should be reduced to 0.5 seconds (disabling has a huge negative impact on performance) if no UPS is connected. I raised my dirty_ratio to 50% and left the 30 seconds because I'm using a UPS and it allows to upload a huge amount of files even to HDDs with maximum 10G speeds (as it's written to RAM and emptying the RAM happens later)
  23. mgutt
    No, because by default Linux does not switch off unused pcie, usb, sata, iGPU, etc. This script sets only all devices from (permanently) ON to AUTO (reduce energy consumption / switch off, if not used). This is different compared to Windows or Apple where energy saving is the default. Source: https://www.kernel.org/doc/html/v4.14/driver-api/pm/devices.html#sys-devices-power-control-files "The setting can be adjusted by user space by writing either “on” or “auto” to the device’s power/control sysfs file. Writing “auto” calls pm_runtime_allow(), setting the flag and allowing the device to be runtime power-managed by its driver. Writing “on” calls pm_runtime_forbid(), clearing the flag, returning the device to full power if it was in a low-power state, and preventing the device from being runtime power-managed. User space can check the current value of the runtime_auto flag by reading that file." This has nothing to do with underclocking currently used devices. It only reduces power of a device that is not used (which is controlled by its driver). This is the same as CPU C-States.
  24. mgutt
    Maybe it's possible to monitor the traffic of the disks or use the Plex API to spindown all unused drives after several minutes instead of waiting for the usual time. But of course this would cause delay if someone starts s second movie from a different disk. What do you think? After they spin-up the usual timeout should keep them in this state or what do you mean?
  25. mgutt
    Wait for the next release. It comes in a few days and supports complete path names and not only sharenames so you can set a disk path.
×
×
  • Create New...