planetwilson

Members
  • Posts

    235
  • Joined

  • Last visited

Posts posted by planetwilson

  1. Actually I think I answered my own question, I had it isloated with VFIO and that made it run at high idle power. By not doing that, installing Radeontop, the power draw came right down, I assume because drivers got loaded? Anyway I can still use it in my VMs and power usage is much lower so I'm happy.

  2. Not sure if this is the right place. I recently upgraded my server to 13700K, Z690 motherboard an 64GB 3200 RAm. I am getting 100W idle which feels quite high. I have done a lot of the tuning mentioned here but have narrowed it down to my GPU, a Radeon RX6800. If I remove that then it idles around 40W.

    Anyone know how I can reduce the idle consumption of the card - it should be less than 10W at idle from reading about.

  3. On 5/3/2023 at 11:03 AM, Stixbnr32 said:

    I have a i7 13700 with RX 7900XT in it.

     

    I use the iGPU for transcoding and 7900xt in a VM.  After some tuning I now get similar performance as my bare metal 5800x3d.  Was ALOT of mucking around/tuning.

     

    I have assigned all P cores other than 0-1HT but would be nice to pin Unraid to E-Cores and use those 2 P cores in the VM seems like a waste of processing power.

     

    Now wondering if its worth pinning some e-cores to VM but not sure if they will be used correctly or not.  To be honest maybe ill just leave as is until the build is stable/working.

     

    Now to sell my 5800x3d RIG with RTX 3070 :P.

     

    Using the latest 6.12rc4 build with the 6.1 Linux Kernel was key to getting it all going.

     

    I am in no means an "intel" guy I always buy AMD but the iGPU transcoding means i could remove my T600 Quadro and sell my RTX 3070.  I have tested AV1 transcodes on plex too and it works fine with iGPU so future proofed there.  So big win for my first intel build since Gen 1 Ryzen.  If i was to build a new baremetal system it be a 7800x3d all day.

     

    So does that mean the VM has 14 cores - albeit 7 x 2 HT ones? If they are presented as individual cores to the VM, what happens if it tries to use a HT pair at the same time, th performance will tank on one won't it as only the host knows they are a HT pair? I've never quite understood this with hyperthreading...

     

  4. I am looking to upgrade my system. I currently have an i7 9700 with 32GB of DDR4 2666 memory. My system is used for gaming and media server mostly. I have 6 cores assigned to the gaming VM (which has an RX6800 passed through) and then unraid and docker just does its stuff although I don't specifically limit to cores 1 and 2.

     

    I would like a bit more CPU grunt in the gaming side of things and am currently thinking of trying to get a 13700 with Z690/Z790 motherboard. Plan being to assign 8 P cores to the VM and the 8 E cores to unraid (I am assuming the E cores still can do iGPU transcoding?)

     

    Is it worth going with DDR5? I could spread the cost of this by re-using my existing memory but I think that locks me into the choices of a DDR4 motherboard then from what I can tell.

  5. Currently have an Intel i7 9700 (8 core), gigabyte Z390 mobo, 32GB DDR4 ram paired with a Radeon RX6800 GPU gaming VM which I have configured to use 16GB ram and 6 out of the 8 cores (not dedicated). Server is used for file shares, Plex (occasional intel transcode) and Arr containers and the gaming VM of course.

     

    I was thinking of maybe upgrading to an Intel 13500, Z790 mobo and 32GB of DDR5 memory with the same graphics card. I like the idea of having the efficiency cores do the Unraid stuff with the 6 performance cores being dedicated to the gaming VM. Apparently the DDR5 can make a fair bit of difference too.

     

    Would I really see much of a difference? Given it is going to cost me approx £500 to upgrade minus whatever I get selling the old bits.

  6. My Plex has started erroring with this:-


     

    2023-05-20 10:03:47,233 DEBG fd 8 closed, stopped monitoring <POutputDispatcher at 23098522090320 for <Subprocess at 23098509334544 with name plexmediaserver in state STARTING> (stdout)>
    2023-05-20 10:03:47,233 DEBG fd 10 closed, stopped monitoring <POutputDispatcher at 23098519824720 for <Subprocess at 23098509334544 with name plexmediaserver in state STARTING> (stderr)>
    2023-05-20 10:03:47,233 WARN exited: plexmediaserver (exit status 255; not expected)
    2023-05-20 10:03:47,233 DEBG received SIGCHLD indicating a child quit
    2023-05-20 10:03:50,239 INFO spawned: 'plexmediaserver' with pid 156
    2023-05-20 10:03:50,471 DEBG 'plexmediaserver' stdout output:
    Error: Unable to set up server: sqlite3_statement_backend::prepare: no such table: metadata_agent_providers for SQL: update metadata_agent_providers set online = 0 (N4soci10soci_errorE)
    2023-05-20 10:03:50,473 DEBG fd 8 closed, stopped monitoring <POutputDispatcher at 23098509157776 for <Subprocess at 23098509334544 with name plexmediaserver in state STARTING> (stdout)>
    2023-05-20 10:03:50,474 DEBG fd 10 closed, stopped monitoring <POutputDispatcher at 23098507118928 for <Subprocess at 23098509334544 with name plexmediaserver in state STARTING> (stderr)>
    2023-05-20 10:03:50,474 WARN exited: plexmediaserver (exit status 255; not expected)
    2023-05-20 10:03:50,474 DEBG received SIGCHLD indicating a child quit
    2023-05-20 10:03:51,474 INFO gave up: plexmediaserver entered FATAL state, too many start retries too quickly

     

    I have tried replacing it with the database/blob backup it takes itself all the way back to two weeks but same error. Plex was working fine the day before yesterday so I am not quite sure what is going on if those backups don't work either.

     

    UPDATE: I have replaced my DB files from a week ago and pulled the older version :1.32.0.6950-1-01 and I am back up and running. No idea what has caused it but will wait a couple of weeks before attempting to go latest again.

  7. Not sure if the same bug on Safari on a Mac, (I've never touched text sizes) I cleared down cookies and local storage but still get this:-

     

    [Error] Did not parse stylesheet at 'http://192.168.44.10/plugins/community.applications/skins/Narrow/css.php?v=1678545693' because non CSS MIME types are not allowed in strict mode.

     

     

    Content loads but no styles.

  8. I’ve read a few threads and it seems some people don’t have the issue and some do. Is there a fail safe way of avoiding the reset bug now? I used to have big problems a few years back with an RX580. I’m thinking of switching from an RTX 2070 to something like RX6800. 

  9. I am trying to get a Cloudflare/argo tunnels setup with Unraid itself. It work fine for my dockers etc btw, this is just the unpaid web interface I am struggling with. All incoming requests go to swag and I have nginx setup correctly for the web socket bits. It kind of works but web sockets isn't working and I get CORS errors because https://myservername.mydomain.com isn't in the list of ALLOWED_ORIGINS if I do a "unpaid-api report -vv" on the console. If I change the TLD under the mgmt access page then I can get the http version in there but not https. Is there a way to just manually edit that list of allowed origins? Is it in a config file anywhere?

    • Upvote 1
  10. 17 hours ago, Whaler_99 said:

    Doesn't work at least on my system. I can excluded an entire selection, but that isn't what I want. I want to exclude my Plex cache and transcode folders only.

     

    You can put multiple values in with commas, mine looks like this:-

     

    /mnt/user/appdata/binhex-plexpass/Plex Media Server/cache,/mnt/user/appdata/binhex-plexpass/Plex Media Server/Metadata,/mnt/user/appdata/binhex-plexpass/Plex Media Server/Media

     

  11. I am getting an error when it backs up syslog which causes the whole job to fail.

     

    [22.01.2023 08:43:43] Verifying Backup sonarr4K
    [22.01.2023 08:43:43] Backing Up: swag
    [22.01.2023 08:43:47] Verifying Backup swag
    [22.01.2023 08:43:48] Backing Up: syslog-192.168.44.10.log
    [22.01.2023 08:43:48] Verifying Backup syslog-192.168.44.10.log
    syslog-192.168.44.10.log: Size differs
    [22.01.2023 08:43:48] tar verify failed!
    [22.01.2023 08:43:48] Backing Up: syslog-192.168.44.10.log.1
    [22.01.2023 08:43:48] Backing Up: tailscale
    [22.01.2023 08:43:48] Backing Up: tautulli

     

    Syslog will update though during the backup process won't it?

    If it helps, I have unpaid acting as a syslog server posting its own syslog to itself, that is so the syslog data is retained post reboot. Perhaps the logic is only expecting a single syslog file?

  12. 1 hour ago, planetwilson said:

    What was the workaround please? This has happened to me as well (has been working fine for years, I guess OPenSSL or similar updated and I don't have any options for stronger intermediate certs)

    Also my cert is SHA-256 so I don't understand how it is too weak?

  13. On 1/15/2023 at 6:16 PM, nraygun said:

    Crapola!

    I did find a workaround for now that allows the weaker cypher. I'll check with the VPN provider.


    @binhex Did the recent Deluge container update upgrade OpenSSL to 3.0?

     

    Thanks for the quick response!

    What was the workaround please? This has happened to me as well (has been working fine for years, I guess OPenSSL or similar updated and I don't have any options for stronger intermediate certs)

  14. I too am trying to get this to work. I have an argo tunnel setup through Cloudflare and that all looks to be working. I have a single ingress rule sending everything to my swag docker and from there to various dockers. I have tried in the nginx config for any particular app to place the websocket specific http upgrade code but it hasn't made any difference. I am not sure if it is the argo tunnel or nginx which is the issue currently.

     

     

  15. On 3/10/2021 at 6:55 PM, ConnerVT said:

     

    Two possibilities here:

    1.  Battery is worn out, or component failure in the UPS;

    2. You are not plugged into a backup battery outlet in the UPS.  Many units have outlets that aren't backup battery powered.  Just offer surge protection.  You're not plugged into one of those, are you?

     

    OMG...I have been running for several years it would seem plugged into the surge only socket....I only came to this post as I just had a power cut and got an immediate parity check which didn't feel right.

     

    • Like 1
    • Haha 1