00b5

Is it common to use a config folder in the cache drive? /mnt/cache/Config/xbmc/userdata Do most dockers store locally needed data like this in "Config", and if so, does it need to be made a cache only share before you try to use it? Hell, while someone is answering that question, i put my image in /mnt/cache/docker.img. Does it (by default) not move things in the root of cache?

Do we have a current (ie, I skipped all the docker stuff until b12) docker setup guide? I think I just need to setup an "image" on a drive somewhere (i'd like to do my cache so that all the spinning drives can stay spun down when they can) and then start using it? In fact, the "docker" tab in the GUI might do it all for me, I'm just not sure. I'm not trying to start a discussion about "lack of documentation", its just that its all over the place during the beta cycle, so if someone can point me to the right/current info, I will roll from there.

Anyone still running XEN VMs in these later betas? I'm still on 5a (i think) so I was waiting to move to a docker supported version, and that waiting got longer and longer since the vm just "works". Can I upgrade to beta8, and then start working on moving my apps in XEN over to docker versions, while keeping them going on the VM until I'm ready? If so (and since i don't have to convert my cache filesystem any longer) it seems like it should be easy to move over now?

Do you mean the headless XBMC web interface, or the docker web interface to get it working in the first place? Now that I see a headless XBMC install in docker, I think i'll strongly consider trying to move to docker....though i JUST got sickrage/couchpotato almost working well with torrents and my seedbox.

Did you modify the stuff after the : yourself? (i made it red/bold) You shouldn't be, since inside the container, the app is looking for whatever that folder name is. So if you change something like :/Data to :/a/path/i/like, then the app inside doesn't understand it correctly, unless you go back in and update it yourself. If you can even do that, now you are stuck having to constantly do that every time. Otherwise, it does look like they start and then quickly exit (one has 1, one has 0, you can probably look those up as reasons for exiting). I'd take out all the "extra" options you might have, and start as simple and easy as possible, just to get them running (and keep running)

When you say switch, do you mean that its all changing away from XEN to KVM, or you just mean when they also ADD KVM support, that you will work on similar tools/images for KVM ?

I wish I could take credit for this, but it wasn't me that figured this out. I believe this resolution came from fiscalcon, here... or, possibly from jumperalex, here... I used the example from jumperalex in my setup, and it seems to be working fine for me also. Totally fixed my automounting, thanks!! I had "noperm" in mine (no idea why, that was weeks ago i set it up) and updating it to x-systemd.automount and rebooting, boom, fixed

Are you doing any passthrough with that asrock 8 core atom? I can't find if its supported or not. Seems like a great low power option, lots of cores to share for vms, etc, but if I can't do passthrough, i'm not sure I'm ready to make the leap to that setup just yet.

You should be able to run most/all of that in the one VM (assuming you are going to run V6 beta now). I'm not a cpu expert here, but I don't see why that CPU wouldn't be plenty to do what you want so far. I'm pretty sure you can run the SSDs in a RAID1 for the cache drive, but it does seem a little overkill. It is a good idea, but you could acheive the same thing by just running an rsync and backing up the misc stuff on the cache drive to the array, should make rebuilding it a little easier. Again, nothing wrong with it, just a little extreme I think. I think your E3 will be plenty, the socket 1150 is the newest intel right now, so it should be around for quite a while. No harm getting a cpu with built in video, you never know what you might retask it for down the road, but the iGPU won't be needed if you are using IMPI. 2x8GB ram is good as well, there are other threads around here about how using 4 ECC dimms isn't the best idea, but doing 2x8 now should be enough, and allows you to add 2x8 or 2x16 if you want to replace them down the road. Adding drives to unRAID causes them to get wiped, so definitely get the data off that drive before adding it to unRAID (doesn't look like it should be an issue with your build). Just run the drives off the onboard ports. When you run out of onboard ports, you'll need to look into a separate controller. Not sure about the PS, but a 550 isn't ZOMG huge, but it might not be in its sweet spot when your server is sitting idle most of the time, or in a lower power state. Not sure about the UPS/etc, but otherwise, good luck with the build, looks fine!

What about a spare power supply for a 2nd unRAID box or just to have a spare PS laying around? Surely I can get something other than an $80 PS as a backup/spare for testing purposes. I suppose in the long run, I might build a 2nd box, and it would have no more than 6 hdds, green/reds at the max (probably more like 3 if I ever did it). Its also nice to just have a spare to power up old MB's, etc. Any suggestions?

The basic answer to your question was that unRAID doesn't need graphics to run, but it is helpful for first time setup, espically if you have never set up unRAID before. Now, whether you system will continue to run, or even post properly without graphics all depends on the MB. My MB will not boot/post without a gpu, I tried. So i just got a hold of a pci video card so i wasn't wasting my pci-e x16 slot. Just wanted to add this in for anyone else down the road doing a search, etc.

its harder to find non server mbs that support ECC. You'll have to really flex your googleFu. But if you are ok with an actual server board, start here: Newegg Intel Server MBs with ECC and 6+SATA

And this one has 12, just depends on what you want: http://www.newegg.com/Product/Product.aspx?Item=N82E16811147165 What size hdds do you have? You said you outgrew your space, maybe its better for you to upgrade to all 4TB hdds first if you haven't already? Anyway, if you are buying a case for something like 10 slots for hdds, and you can fit a rackmount case, I think its the way to go.

What makes a good quality unmanaged switch? I just bought one of these, as far as I could tell most of them are the same until you get into ones that you can manage?

Buy at least one 4TB for your parity drive (or even a 5 if they come out) that way, when you want to upgrade a year from now with some bigger drives, you don't have to buy +1 because your pairty is too small. I have all 2TB drives, so if I want to start adding 4TB drives, I have to get +1 just to start adding them.

That card is PCI, and only SATA 1. I don't think you want that one? That MB doesn't have a PCI slot. You need an PCI-E card, and since that is all you'll ever put in there, might as well get one with as many ports as you'll ever need. Something like this: http://www.newegg.com/Product/Product.aspx?Item=N82E16816124050 Not sure if there are cheaper ones out there, also not sure of that chipsets support in unRAID (didn't look it up)

I am even more confused! You talk about a pre-clear when adding a parity drive, but that only happens when adding a data drive. I do not understand the workflow you have been using. The preclear does avoid the long downtime issue, but there is obviously something we do not understand about your workflow that is causing i it to happen for you. Sorry.... Nothing like trying to dig yourself out of a hole only to realize you are heading in the wrong direction. I made the comments I am trying to articulate here: Re: unRAID Server Release 6.0-beta3-x86_64 Available « Reply #286 on: February 07, 2014, 09:47:42 AM » I had added a 4TB parity drive (which I did pre_clear before using - Don't you do this for parity drives as well as data to ensure the drive is solid?), and pre_cleared the 3TB drive which had been parity before re-adding it as a data drive. At that time I only had my 6.0 server, so pre_cleared the 3TB disk, stopped the array, added the drive and tried to start the array, but it wanted to clear the drive again, which took 6 hours to complete. Because the array was in a starting state (I guess) none of my shares were visible. Once the GUI clear finished I had to click Start on the array again, which happened quickly and I could format the new disk. This was what I had been trying to avoid this time, but am having other issues. As mentioned, I wanted to go through the clear process last night while everyone was sleeping, however I couldn't stop the array as it was stuck unmounting user shares. I had forgotten to shutdown my VM, so I did that but nothing changed. I finally ended up rebooting the server and when it came back up I stopped the parity check, and tried to stop the array, but it still kept complaining about unmounting user shares, which was strange. I did this twice more with the same result, and left it doing a parity check which is still going. I don't understand why immediately after a reboot (and I had not restarted the ArchVM or anything) that it failed to dismount the disks. Did you get your question(s) answered yet? The formatting in your post seems to be wonky, i'm not sure what you are still asking. Assuming you are not pre-clearing a drive just to test it/burn it in: Parity drives do not need to be pre-cleared. Upon replacing your Parity drive (or adding one, if none existed) your unRaid array will start to calculate parity. This WILL take a while, but if i recall, you can still use the array while it is doing its math, it will just be slow Data drives can be pre-cleared (or not). If you do NOT pre-clear, then unRAID will read/fill the entire drive to make sure that it is fully blank, and (i think) set it all to zeros so it doesn't have to do anything extra with parity. This will take quite a while (your previous listing of 6 hours sounds good) and your array cannot start until it finishes. The pre-clear script, however, will leave a special signature on the drive, which unRAID will see, and not try to do anything special to the disk. Your array should then start quickly, allow you to format the drive with its filesystem, and off you go. This is often preferred, since your system only need to be down to install the hdd (either to start preclearing it, or just power it up if you used another system to do the preclear). Pre-clear often takes 24+ hours on 2TB+ hdds depending on I/O speed. You also asked about why you couldn't stop your array right after starting it. I would insure that you are not running any plugins, etc, and that there are no systems trying to use those shares. Any chance you have some other system automatically doing something to those shares when they are available? There are links around here to track down what is still using your shares, which files are open, etc, etc. I don't have it handy, but its worth looking into if you are trying to solve this issue. I haven't seen anyone say they ran a preclear on a v5 system, then installed the drive into a v6 system and it worked without issue. I'm sure someone has done it, and I have no reason to doubt it, but since I've never seen someone say they did it, I can't tell you with 100% it works. (again, I assume it does, don't have enough spare unRAID boxes to test it though!)

Man, all these pc-q25 builds make me jealous, I want to build a new setup with one of these, and maybe find a supermicro itx board, a xeon, and run some nice new 4tb hdds and V6 with VMs, all in a tiny little box. I'll stick with my tower till i'm forced to upgrade for now. Maybe between now and "then" I can upgrade some 2tbs to 4tbs Nice build, nice and neat.

Glad you were able to move from one pc running unRAID to a new pc/case/etc running the same version & plugins without any hiccups. I recently moved everything to a new case, and was as concerned about getting all the cabling right/etc, and I wasn't even upgrading! Looks good, now you need to find somewhere to tuck it away, and forget all about it

You can also just ebay something like this: http://www.ebay.com/itm/Intel-Pro-1000-XT-Server-Adapter-PCI-X-Gigabit-Network-NIC-Card-A51580-015-/190987176497?pt=US_Internal_Network_Cards&hash=item2c77b90a31 Its pci-x, but it fits in a PCI slot. Just make sure you have space for the rest of the card to hang off the back. (like thats not where your sata ports are, or something else). Here is more info on using a PCI-X card in a regular pci slot (this guy actually is putting in a dual port GB card): http://thehomeserverblog.com/esxi/intel-pro1000-dual-gigabit-nic-pci-x-card-in-pci-slot/ I just did this to replace my realtec card, doesn't seem any faster, but the copy speeds do seem more consistent (they don't jump all around like they used to).

That's the beauty of the eSATA port which is standard in my Thermaltake case - just plug in the SATA docking cradle and away you go. That is a very nice case aswell. Oh decisions decisions I am after being gifted an i7-3930K Processor (SROKY). Lucky or what!!!. Will this be ok to use instead of an Intel Xeon E3-1240v3 for my proposed build? I'm sure it is probably overkill but its free and seems like a nice piece of kit. Grats on the free swag. Just verify that the cpu will work with vt-d for you (some K series don't support it, or they do, but most mbs won't let K series do it). Also that i7 doesn't support ecc memory, so if that was swaying you one way or the other in regards to a server MB and ecc memory, you'll have to switch that up. Not sure where you left that, but do some looking into full support of that CPU for virtualiztion before commiting to buying parts for it.

I just did a swap of my e8400 to a q9650 (both 3ghz core 2 based, but the new one is a quad). Just worked, cept i broke the cpu cooler, it was the plastic ones with the pushpins. Bought a $10 one with a backplate and screws, put some new thermal paste, much better. Not that the system runs any faster or anything, but i also added 4gb of ram (had to find some DDR2 for $30 used) but its all in prep of trying to run 6 with some VMs

Don't upgrade to 3TB, go straight to 4TB. Or at the least, get a 4Tb for parity, and then save a little with 3GB replacements if you really want. You can get a pci-e 1x sata card that will support 2 hdds, that might suffice for you at the moment. I didn't look at your MB layout or anything, just shooting from the hip on this one. There is a standard goto monoprice card that is suggested in the FAQs.

The 212 evo says: •Versatile all-in-one mounting solution supporting the latest Intel® LGA 1366 / 1155 / 1150 and AMD FM2 / FM1 / AM3+. So, there is a chance that the backplate you installed would just work with another cooler. Depends on how specific/proprietary cooler master was with the design. I'm guessing the back of your MB plate in your case isn't open, eh?

Use the intel ark site to make sure the CPU you pick supports all the VT stuff you want (if you want full passthrough, etc) http://ark.intel.com/ I was looking at getting (new) a haswell E3-1220v3 ($200) and a server mb that is known to support the VT features you want with the appropiate amount of expansion slots (do you want to pass through a bunch of video cards to VMs, so you need 2 or 3 pci-e 16x slots? etc). I was thinking of a supermicro board, newegg had some nice combos available. I'd also go the route of getting half the max ram your board can do, so if you have 4 slots, and it supports 32gb, get 2 8gb in there now, so that if you ever want to max it out, its easiest. Don't buy 4 4gb, or something like that, since you are just wasting slots/money if you upgrade stuff down the road. If it only has 2 slots, then get as much as you think you'll ever want up front, for the same basic reasons.