MrCrispy

There are some good deals on mini/sff pc's with Intel 7th/9th gen cpus, usually Core i5. Like the i5-7600, i5-8500 etc. Newer ones like 10th gen and newer are far more expensive. Are these good enough for general purpose use? Running a bunch of dockers for the usual purposes, and perhaps a few vm's. I'm guessing the main cpu use comes from VM since docker containers hardly use any cpu and will be idle most of the time, with disk bandwidth more of a limiting factor, correct? Also it seems like these cpu's will support hw 4k transcoding and encoding as well?

what is the new url for this link in the OP - http://lime-technology.com/forum/index.php?topic=3666.0 I tried forums.unraid.net/topic/3666 but that won't work.

When does this take effect? If I want another key, can I purchase the Pro key now and be guaranteed it will be for lifetime, even if I don't activate it for a few months?

I hope its ok to ask questions here about Tdarr usage in general? My Unraid server is dead (hw issues) so I will be using a cheap sff pc I bought and I'm new to this. My goal is to convert to hevc, having looked at Tdarr, Unmanic and FileFlows it was not clear what to pick but Tdarr seems the oldest and most configurable so I'm picking that. I have a few questions - - my hw will be 2 sff mini pcs, with an Intel Core i5-7600 and a 8500T, they are an HP EliteDesk and Dell Optiplex, both with 8GB ram. I will install a Linux distro (Debian or Ubuntu I think) on both and tdarr via docker. - from everything I've read Intel QSV is the sweet spot, cpu takes too long and Nvidia sacrifices quality. From Wikipedia it seems both 7th and 8th gen support QSV, correct? - is there any reason to use VAAPI and not QSV, I'm assuming QSV is better? - separate instance vs nodes - I don't want to pay for the premium version just now, the free version doesn't let you assign libraries to nodes. so e.g. if I have 2 different libraries I want to transcode, is it better to have to instances (one on each pc) with hdd's connected with the media needed, vs having one master node, in which case I'm guessing there will be a lot of network transfer since the media will be randomly assigned? the advantage with a single master seems to be unified stats - workers vs nodes - does node mean a separate physical machine (vm) and worker a thread on that instance? is there a guidance on how many workers/nodes can be run based on cpu? does cpu even matter or is it all dependent on the iGpu? and most imp - - what plugins are recommended, there are so many. e,g do I use Tdarr_Plugin_JB69_JBHEVCQSV_MinimalFile or Tdarr_Plugin_bsh1_Boosh_FFMPEG_QSV_HEVC or one of the many migz ones, and what tweaks do people make? - can someone give me a rough idea of the speeds/time I can expect for conversion with above hardware? - I read in some posts that Unmanic is faster/smaller file size. is this because of a lower crf/preset in their default plugin or is there something else going on? which of the 3 tools do people recommend? I'm asking here because the Unraid forums are some of the most helpful and knowledgeable, thank you in advance.

Can I use this outside unraid? I will just be running Debian for now since it only has 2 hdd's.

I remember a few years ago there were many sellers on ebay selling used enterprise DAS like NetApp DS4243, Lenovo SA120 etc for cheap prices, like $100. I dont see anything now, they cost ~$500. Is there something now thats a good deal? I want 12-24 drive storage, attached via SAS card, and reasonably quiet since its for an apartment. What are my best options?

It'd be nice if there were some screenshots of the new UI and perhaps some details on whats new between releases in the UI. Is there a thread for it? I know I've seen it before but IMO should be part of the notes.

Any plans for this? Docker compose would be a much better way to handle multiple containers and would need less effort from the devs and maintainers and it opens up other options as well. I don't know if there are technical issues preventing this or there isn't enough demand?

Docker containers should not contain any data, they are supposed to be static images and all data should be on externally mapped paths. At least that's my understanding. So if container size is growing something is wrong. Even with log files they should be getting pruned periodically.

So does shutdown/reboot perform a graceful shutdown? Does it abort current disk activity? Can it cause corruption on the file system? Does it interface with the UPS? I remember reading that there are issues with shutdown.

This may have been asked before - if I install a plugin directly (i.e not from CA) will CA still know it was installed? I'm assuming yes since it'd make sense to parse the list of installed plugins (which I believe is on the flash) but I'm sure CA keeps its own registry etc somewhere as well?

I just used this calculator, its still useful because of the stupid way HDD sizes are labeled! One tiny request - make the default unit 1TB, I think 250/500GB drives are pretty rare now and they can be added as 0.5 etc. Should be a simple change.

Is the SuperMicro X8DTE-F supported?

LXD is a containerized version of LXC containers which have been around for a while. Looks like an interesting development for larger use cases, its a basically a VM based on container principles. To me the primary advantage of Docker is its simplicity and ecosystem. Its a single process model which lets you deploy and scale apps. If you want to run an app, you use Docker. If you want to run a VM, you can now use LXD and it'll be faster than a normal VM. In most cases esp for home users the goal is to run multiple apps for which Docker is perfectly suited IMO.

There's no way to search threads on this forum, so can someone tell me if this plugin works for SuperMicro motherboards which have IPMI? I want to be able to monitor stuff and auto control fan speeds etc.

+1 as well. Better to use official tools that simplify things whenever possible.