Jump to content

jkirkcaldy

Member
  • Posts

    118
  • Joined

  • Last visited

Everything posted by jkirkcaldy

  1. again, there is no information about how they downloaded it etc.
  2. yeah it's a bit of a nothing post. It seems like they know enough to be dangerous but they haven't actually showed how they came to their results. What their download commands are, whether they downloaded the file from floatplane or used youtubedl for that too. They are also comparing bitrates of different codecs as if they are like for like. It would be a pretty stupid thing to lie about, especially as it could be checked so easily. It seems like they are just jumping on the bandwagon at the moment.
  3. I can't reply to this on reddit as the subreddit is in community only mode. The testing here seems extremely suspect. I suspect that the person downloading these files is encoding as they download. My test gives drastically different results to their results.
  4. IMO, GN were completely correct to call out LMG on their test results and QC process. It’s forced their hand so they are now looking at their processes. The amount of corrections in their videos was getting out of hand. Which wouldn’t have been a problem if they weren’t “labs” results. But if you are going to found a company on the basis of accurate and thorough testing then it should be exactly that. If you aren’t 100% confident in the results, then they should not be published. This should be solved by better QC and testing parameters. Personally I think they should avoid calling them “labs” results until they are confident in their testing methodology. Perhaps they could invite GN to validate their testing methodology and results in the future. I think calling out the testing of the Billet block was completely justified. From a personal pov I can see Linus’ argument against retesting. The conclusion would have been the same so there was no point. But from a tech review stand point, it was unacceptable. I believe calling out the handling of the water block after the auction wasn’t called for. The handling of this has been poor to say it kindly, I believe it was a complete fuckup, but 100% human error. There is no reason to believe that this was done maliciously, what would LMG stand to gain? This was a business dispute that should have been handled out of the public eye. i don’t believe that any of these issues are an issue that can’t be overcome. And I think that Linus may have to think about how he addresses the community going forward. Perhaps he needs a disclaimer, “these thoughts are of the individual and do not necessarily represent the ideology of LMG” or as he has said, they need to be vetted by a PR team before hand. But I believe this will be the end of WAN show as we know it. Perhaps they will move all the hot takes and opinion pieces to the floatplane pre show and keep the YouTube stream sanitised. The allegations from Madison are a completely different story and it’s good to see they are taking them seriously. But we should be prepared to hear nothing about this for the foreseeable future until the investigation has completed. LMG shouldn’t acknowledge this any more than they have done until the investigation has completed. It doesn’t do anyone any good to feed into speculation or rumour. These things take time and we should give them the time to process this properly. And now they have committed to publishing the results, there should be a public outcry if they go back on this. Though, personally, I believe they will keep their word.
  5. nextcloud for auto uploads, then you can use that for the backend storage for photoprisim. I think there are options for facial recognition too.
  6. If you are going to do anything with Plex, go with intel. It's trendy to go with ryzen these days and the core count/price is very attractive, but the gpu transcoding on the intel cpus will run rings round ryzen as well as being more power efficient. Stay away from Xeons, they are super attractive as they lure you in with high core counts and lower price points for used hardware. But it is better to spend €50-100 more on a new intel cpu as they will sip power and save you money over the course of a year. A nuc or beefed up thin client is a good goto machine for sipping power but being relatively powerful.
  7. Sounds like a permissions issue, Radarr/Sonarr can mess these up if you use them. Go to your unraid dashboard and click on tools. If you have Docker containers on yoru server click Docker safe new permissions, if not click on new permissions. Run that, it will take a while depending on how much data is on your server. Once it has completed you should be able to see all your files again.
  8. I think you should stop and think about other solutions that would be way better and far easier to implement. If you only need dual display you can set up something like parsec on the VMs, put the server anywhere. Then use something like a raspberry pi to connect to the VM. This way you still do all the heavy lifting on the VM, but the users connect to the raspberry Pi. This would be much cheaper than optical cables and getting a bifurcation card for your motherboard. 3x raspberry pis will be about $150. With parsec you need the paid version for multi display streaming, but considering a single optical thunderbolt cable is $500 and a dock is about $200, and the cost of parsec is about $10 per month, you would get a total of 18 years use (6 years of use from each VM) before it would be cheaper to use thunderbolt. That and Parsec would then let you use the VMs from anywhere in the world too. I work for a TV production company in London and we have been using Parsec Teams since the summer last year and it has worked flawlessly for us to have about 7 editors and 3-4 support staff working remotely. Means we can use our own infrastructure, the editors remote into the workstations still in the office and as far as they are concerned their workflow doesn't change apart from the fact that they are at home rather than in central London.
  9. Install the wireguard plugin on the unraid box and forward the port from your router to unraid. That's what I do to gain access to my unraid system out of the house. I also have a vpn on my router for backup if my unraid system crashes or needs to be rebooted. Just a word of warning on nextcloud, once your files are in it, you can only access them from within the nextcloud system. i.e. you can't have a smb share with your files and have them available in nextcloud.
  10. You can't use port numbers in DNS entries as far as I am aware. But if I am wrong, please send a link to where you can because I would like to see that.
  11. That doesn't really work. not unless you add the port to the domain (eg plex.example.com:32400). In which case there is literally no point to using different sub domains because plex.example.com:32400 would go to the same place as ts.example.com:32400 Also if services need specific ports to be open, these will need to be open regardless of whether you use a reverse proxy or not.
  12. A caching server is a good way to go about this, probably the only way to go about it really. As far as I know, you can't use the same files hosted on a server to serve the game to multiple client PCs. So you would still need to download the game onto each PC in the LAN party. The caching server can take some of the strain off your internet connection, so you would need to have a dedicated machine that would act as a cache. Basically you would download PUBG on your computer and the server would save the files as they came from Steam. Then in theory, anyone else that needs to download the game on your network would pull the files from your server rather than the steam servers. With the correct setup you can get some stupid speeds from the cache server. But a word of warning, I tried to do this with a Virtual machine, and you actually needed to dedicate a large amount of resources to the caching server in order to get really fast download speeds.
  13. Depending on how comfortable around PC hardware and building your own PC you may be better off building something rather than buying. I built my first Plex server for about £250. There are some advantages to buying a NAS but unless you spend a little more you will be missing out on the more advanced features. I started off with a pentium dual core and 8GB RAM (RAM was cheap when I built my machine) The idea was that I could scale up various parts as and when my needs grew. And since then I have gone from a small 4TB tower PC to a rackmount beast with 52TB (40 usable). You just don't have that scalability with a store bought NAS Also, unless you are spending money on a very good CPU, you should get the idea of transcoding 4K footage out of your mind. You can optimise the media ahead of time, but I - as well as most people I think - have found that it is generally best to have both a 4K version as well as a 1080p version. Then if you do need to transcode, you can do so from the 1080p copy and have a much better time. Also most 4K files have HDR colour toning, and as soon as you transcode this you will loose it and your file will look very washed out and flat.
  14. A reverse proxy is what you're after. I have a single static public IP address then a load of services running on different virtual machines. Here's a brief list of some of them: Plex Organizr Monitorr sonarr radarr ombi nextcloud hastebin gitlab homeassistant Bookstack wiki There are probably more but you get the idea. I have a single VM dedicated to being the reverse proxy. So all my traffic is forwarded to this internal IP address then it separates it out and forwards the traffic to where it needs to go. It can be quite a daunting task at first but it gets a bit easier as you go and there are a shed load of tutorials online.
  15. Depending on how much of an issue aesthetics are for you, you could always buy a rack-mount case full of hot-swap bays, take the rack ears off and store it on its side. They are roughly the same sort of size of a desktop case, they are usually a little longer though. Most rack-mount cases look alright from the front but are usually bare metal on the sides as you usually wouldn't see this in a rack. But I got a rack-mount case for my white-box server long before I ever got a rack to mount it in.
  16. The problem with workstation equipment is they can often have not a lot of room for expandability. The HP Z Series workstations for example only have enough space for something like 5 drives in the 840 series. So whilst they are very powerful and very quiet, you can only add a few more drives. Compared to some of the rackmount equipment where you can fit 24+ drives in a case of a similar size.
  17. you need to include a bit more information. Are you hosting this at home? What level of redundancy are you aiming for? Do you want full redundancy? This will cost way more money than you think as involves a minimum of 2 of everything. 2xinternet connections, ups, generators, minimum of two servers etc etc, What's your budget? Once we have a better understanding of what it is you are trying to do and what sort of tools/budget to work with we can help more than we can now.
  18. I use Transparent Raid from FlexRAID. From what I can gather, it works in much the same way that unraid does in terms of storage. But it runs on windows, as well as other platforms. I have 40TB usable space with 2x6tb parity drives. Although I'm considering removing one of these as I need the storage space. You can pool random disks from different vendors and different sizes. You can delete and recreate the array with no data loss. You can add a write cache using SSDs if you want. It's worked well for me for a couple of years now. Having said that in that time I haven't lost a disk so I'm not sure what the rebuild process is like. But because everything is stored on normal ntfs drives you can pull the drives out and read them on another computer.
  19. I think on the newer hyper-v you can pass through a graphics card but you need to pass through the entire card to a single VM. So you would need to have a GPU per VM.
  20. I ran an Ark server on a windows 10 Virtual machine with 4cores and 8GB RAM. This was only for a handful of friends so wasn't a huge server, but it wasn't that resource intensive. The VM was running on Windows server 2016 using Hyper-V. Hardware was a Dell R710 low powered Xeons, I think L5630 or something similar. Ran perfectly.
  21. the same reason people go to best buy or PC world and buy a desktop computer there. Most people aren't comfortable building their own systems and buying pre-built usually comes with some sort of support system as well as a single place to return the hardware if there is a fault or dealing with warranty etc.
  22. If he's an indie film maker I'm going to assume he only really works/edits as a one man team. In which case I wouldn't get a nas. I would get a DAS. If a NAS is needed, Qnap do one that I have used on a feature film before that had 2x10GB rj45 and 2xThunderbolt2 as well as a couple usb3 and 1gb ethernet. I can't remember the exact model as it was a couple of years ago now. It was about $2500 for the NAS and then the disks were extra. It had 8 x 8tb disks and in a raid array it had a throughput of around 700mbps.
  23. There is also a Hyper-V OS that you can install for free. The catch is that I think realistically you need another windows PC to manage it. But it is completely free, no limits, no trial. Just free.
×