sane

Tis said no. Take out the old chip, put in the new. You will likely have to upgrade the BIOS, so you'll need the old chip to get that done first, but you could then sell the old chip for some cash. Rough guess is it won't go incompatible till 2021, with PCIe 4 and DDR5, but obviously nothing is guaranteed. AMD have a good track record though.

Worth considering that the AMD AM4 board will support new CPUs up until 2020. That's two new generations of chips (Zen 2 & Zen 3) past the current, down to 7nm. The desktop Ryzen chips will likely hit 16 cores (maybe only 12) with the 7nm switch next year. So if you are looking for upgradability then an $85 B450 board will do you for a number of years. Intel is very unlikely to even get close, and certainly not at sane prices. PS The Threadripper 2 is a more expensive design, but will hit 32 cores.

Whatever happened to this? It's been about 6 months since the original call, but no logo has (obviously at least) appeared. Even graphic designers aren't that slow, so did this get thrown into the 'too hard' pile?

That's fine by me sir, glad you liked it. I wonder when the tame designer lime employed is going to come back with their finalised efforts? And I wonder how many will say "oh no, I preferred ...."

One last, very cut down attempt, not busy, from a free font

OK, since people were saying nice things, I gave it another half hour. Tweaked the text into an actual path, stylised it, bolted in Lev's box concept for container/disks. Also put the 'un' chain with the U at a slightly jaunty angle which looks a bit better methinks. Still not right, too busy and quickly lashed up, but might help on the ideas front. Also tried some B&W, ico, and NAS box example other looks.

If you want the marketing bull, the U & N are supposed to be reminiscent of interlocking chain, and thus subconsciously remind people of security - and can also be abstract enough to run as a icon in it's own right. Or otherwise, in practical terms, it's something to do with U and the N since otherwise they can unbalance things. As I said this was pretty much slapped together - and the RAID part isn't right. Doesn't want to be too busy, and the idea was to merge serifs to give a 'continuous flow', which could be accentuated. But it could certainly also be very different. From a logo design front the only thing that niggles is there is no reflection of the containers or VM aspects - couldn't think of any approach that wasn't naff.

OK, this is more slapped together on photohsop to illustrate some concepts rather than anything deliverable - would obviously take more on the rendered elements and tailoring text into something unique - but hey, what do you expect for free !

Sound good ! There were tutorials in the docker engine fora, but not sure exactly where the best one is now. Worth asking the question over there for pointers, I don't think they bite. Maybe for templates ...

Seems like an Asterisk/PBX docker is something that quite a few would like, but nobody has got to the idiot proof app stage. Personally I'm looking for something that can talk to various VoIP services, and to an ATA. Anyone with any pointers?

Yay! My procrastination in not upgrading from 6.1.6 wins again. Saying that, I'm not touching this till version .1 has been around a while.

Definitely something screwy on the upload of PNGs. Even with the right size, etc. it refused to work unless I reached in with an SSH session and forced it. Probably need to also mod to give a subtle white outline to the logo/text - otherwise it disappears against a dark background.

I'm running a N36L microserver, thus an AMD AthlonII Neo with 6.1.6 (don't trust betas). Nominally runs at 1.3Ghz, but in the dashboard it shows 800Mhz on idle cores (slowest state). So, either it's only in the betas, or it's not all AMD chips.

You can also use 'Unbalance', which is now a plugin, to shift subdirectories around between disks with a web GUI.

Having just done 5 > 6 I followed the "Restore saved configuration files" section and copied across the .cfg and config/shares, etc. stuff - and it all worked with no reassignment necessary. I think this also is supposed to work for 4.7? http://lime-technology.com/wiki/index.php/Upgrading_to_UnRAID_v6#Install_the_unRAID_upgrade

Yeah, that's what I was looking at. Seemed to be a toss up between the two approaches, and the unassigned devices would be needed anyway. Agree with trurl, using a cache does not require "unassigned devices" You are mistaken. Didn't mean that the cache needed 'unassigned devices', more that I did so that I could attach USB drives and backup the share contents to offline storage. Intend to have a hack at the weekend and will probably try the no-cache cache option first, then move to 'unassigned devices' if that doesn't seem to do the job. Need those dockers working to get back to the functionality I had - then move on from there.

Yeah, that's what I was looking at. Seemed to be a toss up between the two approaches, and the unassigned devices would be needed anyway.

Not against the docker part, just against the cache drive part - ever since I found out it could easily fill up and wasn't working as a true cache. Want the drive for docker and VM, and not cache. Ahh, must be looking at the wrong thread then - quite possible. Something to make it easy, rather than mucking about with command lines.

Cheers for that, though I've no idea why it'd get stopped by adblock. Now just got to work out how to have a drive for dockers that isn't a cache; and what the best route is for moving files between drives to shift from reiser to xfs (unbalance seems abandoned).

Finally updated to 6.1.6 from 5.0.4 and following the instructions I wiped, reinstalled, copying across cfg files, share directory, etc. so that I didn't lose everything. However, the "Shares" tab in the GUI has no user shares listed after this. BUT, the shares ARE listed on the dashboard, and can be seen/accessed OK. And yes, User Shares are set to yes, and I've cycled the setting/system to no avail. Anyone got any ideas? Is there some hidden posting I missed?

I'm not sure the desire for a regular and swift release schedule for UNRAID versions is compatible with an unstable OS basis in which plugins are supposed to work. You can't have the ground shifting under the plugin authors and expect swift rewrites - it just doesn't work. You need to decouple changes and enhancements to the OS and the plugins such that each can go at their own pace. Otherwise people will have to hold off and stay on 6.0 because 6.1 is incapable of running a plugin they require. I'd suggest that there needs to be a 'compatibility mode' where these changes are turned off and existing plugins can continue to run - giving time for the plugin author to change their code. Eventually, when all the plugins have been updated, at the next version, that mode can be removed/reverted to being 'compatible with 6.1'.

They did do the thing of "the data is only on one drive, so transfer speed" without considering the practical network speed or how fast a home system actually has to be.

True enough, keeping drives spun down is pretty important. I wonder if the "cache dirs" and the disk usage could be integrated (after all, cache dirs needs to maintain data on all the directories, so size is not a major addition?). I do feel that caching the data in some way would be smart - only refreshing totally when the drives were already spun up, but idle (or when commanded).

Personally I like the look of philesight a damn sight better, plus the DUC docker seems unsupported. And in any case, this is about the base OS reporting of disk usage, rather than dockers.

Can I suggest that if you are uprating the disk usage information, you go a more graphical route - say the cushionified, squarified, treemaps of SequoiaView (linux alternatives are available) or maybe the very nice pie chart of Philesight - which is opensource and GNU http://zevv.nl/play/code/philesight/