Nem

Members
  • Posts

    177
  • Joined

  • Last visited

Posts posted by Nem

  1. Planning on building a new server. Would like to get a quick sanity check on components in terms of compatibility (with each other, and with unraid in general). I'll be migrating a few drives over from my current system (7 drives in total)

     

    Link: https://ca.pcpartpicker.com/user/nem2k/saved/C7yTHx

     

    Use case:

    • # of concurrent users for all of the below: 2
    • Emby (iGPU transcoding)
    • Media management (Radarr, Sonarr etc)
    • Web server
    • MySQL database for some side projects
    • File server for backups
    • Nextcloud
    • Wireguard
    • Will eventually be installing a GPU (3070) and passing it through to a VM for some machine learning things once in a while
    • Might run a 2nd instance of pihole here incase my primary goes down

     

    Questions:

    1. Will the cooler fit inside the 4U case?
    2. Is the PSU overkill?
    3. Is the RAM compatible? (dont see it on the QVL for the motherboard)
    4. In terms of PCIe, eventually I can see myself filling the slots with GPU, HBA, 10Gb NIC. The GPU will likely be a dual slot card, rendering one of the PCIe inaccessible. How do people usually get around a problem like this?
    5. Was originally aiming for a 2378G, but those things are impossible to find. How much of a real world performance drop will the 2176G be?
    6. Instead of the Xeon/ECC/C246 combo, I was also considering a more consumer-level setup based on a 10th gen i7. Any major benefits one way or the other?
    7. Any other potential pitfalls to watch out for in terms of unraid compatibility?
  2. I've been eyeballing some of the new Xeon E-23xxG CPUs for a new server build, for example, the E-2378G (https://www.intel.ca/content/www/ca/en/products/sku/212262/intel-xeon-e2378g-processor-16m-cache-2-80-ghz/specifications.html)

     

    I see that it was released Q3'21, but have not been able to find a single retailer to purchase the CPU from in Canada

     

    Is this because of current supply chain issues causing low supply? Or is it because server-grade components don't reach consumer outlets until a while after their launch date? (I'm not sure how enterprise and consumer markets differ in terms of rollout schedules)

     

    In essence, when might I realistically be able to get my hands on one of these? If it's supply chain issues I can just keep searching and hoping for stock, but if its because consumer rollout is typically delayed then I might just look at the 22xx series instead...

  3. @ljm42 tried a couple of tests...

     

    > ping rebindtest.unraid.net
    
    Pinging rebindtest.unraid.net [192.168.42.42] with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 192.168.42.42:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

     

    > ping <personalhash>.unraid.net
    
    Pinging <personalhash>.unraid.net [192.168.187.10] with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 192.168.187.10:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

     

    In the second case, it is trying to resolve <personalhash>.unraid.net to my old unraid ip (192.168.187.10), rather than the new one (192.168.1.10)

  4. 9 minutes ago, ljm42 said:

    try https://192.168.1.10:4343

     

    that seems to work, thanks

    I think I'm confusing myself with my own setup. why does :4343 work, but not :8181?

     

    also, how does the redirect to xx.unraid.net:4343 fit into the picture? what is the purpose of that? that is the page that :8181 redirects me to, but now just leads to a connection time out (it used to work in my old apartment)

  5. 6 hours ago, Squid said:

    use https://192.168.1.10:8181

     

    Also, not that there's anything wrong, but usually in the case of a conflict between 80 / 443 you would tend to adjust the host port of the container (leaving the container port alone) to solve the conflict.

     

    https://192.168.1.10:8181 doesn't work. It doesn't even seem to redirect me to xx.unraid.net:4343. I just immediately get a ERR_SSL_PROTOCOL_ERROR

  6. I used to run my unraid machine at 192.168.187.10. Under management access, I had the HTTP and HTTPS ports set to 8181 and 4343, because 80 and 443 were being used by a docker container

    Therefore, on my local network I would access the webGUI at 192.168.187.10:8181

    I just moved to a new apartment and set up a new network under 192.168.1.x. I went into the webGUI and set the static ip to 192.168.1.10. I seem to be able to access my shares and docker containers just fine, but I'm unable to reach the webGUI at 192.168.1.10:8181. It redirects me to xx.unraid.net:4343, where I get `ERR_CONNECTION_TIMED_OUT` ... is there something else I need to change somewhere? How much of this is due to incorrect settings, vs having a new public IP as a result of being in a new apartment and new ISP?

  7. I've been using nothing but WD Reds (WDxxEFRX) in my server, which I believe are CMR drives, but it looks like theyre not being made anymore and are becoming harder to find. The newer alternatives (WDxxEFAX) are SMR drives, which I understand are not as good for NAS performance?

     

    A few related questions:

    1) Is it ok to mix drives from different companies? For example, if I wanted to add some Seagate IronWolf drives to my current WD array?

    2) What is the difference between IronWolf/NAS/Pro models? And are they CMR/SMR?

    3) How are the IronWolf drives compared to WD Red? Any other companies that are recommended?

     

    The new drive will be used as a parity drive (for now). As a parity drive, it would have to be at least as large as my largest array drive, but I'm not sure if a 6GB Seagate is exactly the same size as a 6GB WD Red

  8. 1 hour ago, bobbintb said:

    Yeah, same for me. It's asking to log in but I can't find a default login anywhere. Tried the usuals.

     

    when you first install v3, it should launch a wizard to set up the server and to create an admin account.  I dont believe theres a default admin account

     

    I had a similar problem when I installed v3 - think there were some settings or cache that was retained from my previous v3-preview container. Try clearing your cache or reinstalling the container (manually delete the appdata config directory)

  9. 1 minute ago, digiblur said:

    Has this container been updated to allow you to pick a better encryption method? Last time I used it my client app complained about several weak security options in the app and I didn't see an easy way to change them.

     

    I changed my ciphers within the webgui itself. under advanced VPN config, I just added `cipher AES-256-CBC` to the server and client config directives

  10. I just updated my container (lsio-preview ombi 3) and the page is only half loading (menu shows, but nothing else) and devtools is showing the following error:

     

    ERROR Error: Uncaught (in promise): Error: No NgModule metadata found for 'undefined'.
    Error: No NgModule metadata found for 'undefined'.
        at t.resolve (vendor.js:1)
        at t.getNgModuleMetadata (vendor.js:1)
        at t._loadModules (vendor.js:1)
        at t._compileModuleAndComponents (vendor.js:1)
        at t.compileModuleAsync (vendor.js:1)
        at t.compileModuleAsync (vendor.js:1)
        at e.project (vendor.js:1)
        at e._tryNext (vendor.js:1)
        at e._next (vendor.js:1)
        at e.next (vendor.js:1)
        at t.resolve (vendor.js:1)
        at t.getNgModuleMetadata (vendor.js:1)
        at t._loadModules (vendor.js:1)
        at t._compileModuleAndComponents (vendor.js:1)
        at t.compileModuleAsync (vendor.js:1)
        at t.compileModuleAsync (vendor.js:1)
        at e.project (vendor.js:1)
        at e._tryNext (vendor.js:1)
        at e._next (vendor.js:1)
        at e.next (vendor.js:1)
        at O (vendor.js:1)
        at O (vendor.js:1)
        at vendor.js:1
        at t.invokeTask (vendor.js:1)
        at Object.onInvokeTask (vendor.js:1)
        at t.invokeTask (vendor.js:1)
        at e.runTask (vendor.js:1)
        at m (vendor.js:1)
        at <anonymous>
    It @ vendor.js:1


    vendor.js:1 Uncaught (in promise) Error: No NgModule metadata found for 'undefined'.
        at t.resolve (vendor.js:1)
        at t.getNgModuleMetadata (vendor.js:1)
        at t._loadModules (vendor.js:1)
        at t._compileModuleAndComponents (vendor.js:1)
        at t.compileModuleAsync (vendor.js:1)
        at t.compileModuleAsync (vendor.js:1)
        at e.project (vendor.js:1)
        at e._tryNext (vendor.js:1)
        at e._next (vendor.js:1)
        at e.next (vendor.js:1)

  11. I just upgraded to 6.4.1 and on my main dashboard I'm seeing warnings regarding a couple of drives and their SMART status

     

    The warning is UDMA CRC error count: 1 and CRC error count (no UDMA): 1

     

    Can anyone explain what these errors mean, whether theyre serious, and what can be done about them?

     

     

    dashboard.png

    details.png

  12. the docker instructions says to remove the default admin account and assign admin privs to another user

     

    im wondering what a typical user setup should be:

    1) 1 admin level user with a strong password and 1 regular user. both accounts used for daily vpn logins or

    2) 1 new local admin account, and 2 regular user accounts. only the regular priv accounts are used for vpn connections

     

    is it normal to recreate an admin account after removing the default one, and use only that for setup and admin duties? Im trying to understand what the security reasons are for removing the default admin account, and whether creating another local admin account defeats that purpose

  13. are there any plugins/docker apps that can take information about the unraid machine (current cpu usage, ram usage, disk usage etc) and pipe it out to json in real time so it can be used elsewhere?

     

    ultimately I want to create a dashboard with a live view of how my machine is currently performing without having to log in to the unraid webgui

  14. given that the preclear plugin has not been updated for 6.4, I need to turn to WD data lifeguard to preclear a new disk I purchased to replace my existing parity. I dont really want to risk my server not booting

     

    Im looking at the options in the program: Quick test, extended test, erase

     

    Which option would be equivalent to running the preclear plugin for 2-3 passes? Would it be running erase 2-3 times?

  15. 13 hours ago, binhex said:

     

    ok so you def do have dns resolution issues there, i see airvpn send a dhcp option to set the dns client to use their dns, i wonder if they also block dns resolution using other name servers, hmmm. 

     

    ok try this, create a new env variable with key name NAME_SERVERS and set the value to '10.4.0.1', this is airvpn's name server.

     

    @binhex so I took a look at deluge again this morning and it was downloading again without me having changed anything. So the problem seems a bit random, but as of right now everythings seems OK

     

    if I make the change you suggested I cant connect at all. I keep seeing the following:

    2018-01-26 17:51:50,560 DEBG 'start-script' stdout output:
    Fri Jan 26 17:51:50 2018 RESOLVE: Cannot resolve host address: europe.vpn.airdns.org:443 (Name or service not known)
    Fri Jan 26 17:51:50 2018 RESOLVE: Cannot resolve host address: europe.vpn.airdns.org:443 (Name or service not known)
    Fri Jan 26 17:51:50 2018 Could not determine IPv4/IPv6 protocol

    plus, the solution seems a bit inflexible. what if their DNS servers change? Im wondering if theres a more flexible solution that wouldn't require manual intervention when something changes on their end