garycase

The disadvantage of a cache drive has always been that it's not fault-tolerant. v6 resolves that by providing the ability to use a protected cache pool => if you choose not to do that, then you still have the same disadvantage as before ... i.e. files on the cache are NOT protected. That's ALL that the symbol means, so I don't really see where it's an issue. However, perhaps a different symbol for cache-only shares vs. cached shares could be implemented. I personally don't see the need ... and I would NOT make it a green ball, which has a completely different meaning with regard to shares (i.e. "everything's protected") ... but perhaps it could be a green triangle [With a tool-tip that says "Unprotected cache-only share"]

I'd agree IF it was used for two different states of the SAME share ... but that's not the case. It ALWAYS means exactly the same thing ... "Some or all files are on unprotected storage". For a cache-only share when the cache drive isn't a protected pool, that will always be the case ... but a user should certainly know if he's configured cache-only shares => and this simply reminds him that those aren't protected. For other shares that are cache-enabled, it simply notes when you have files that haven't yet been moved to the protected array. But the meaning of it is completely consistent. As I noted earlier, if a user doesn't realize that files on the cache aren't protected, this simply reminds them of that. And if that's a concern, simply adding another cache drive and creating a pool will result in the symbol ALWAYS being the little green circle you like

I don't see it as a "shared" symbol => it's a symbol with a very specific meaning -- i.e. "Some or all files are on unprotected storage." That's it ... it means the same thing whether the reason they're on unprotected storage is because that's where you wanted them (e.g. a cache-only share) or they're simply cached until the mover runs and moves them to the protected array. If someone is concerned about the "unprotected storage", then they should learn a bit more about their system => they may then decide to not cache their shares and/or to add a 2nd cache drive and set up a cache pool so the cache is fault tolerant.

Bingo. That was the main reason to make the change, IIRC. It's not a bad idea, actually. It lets every flash drive be a backup for the keys on the others. I still like having a more organized backup that's specific to individual flash drives, but I can certainly understand the attraction of this approach.

Well I'll be a monkey's uncle.... Going to reboot and try it out now! And..... It works!! Nice feature for those of us with multiple keys! Personally, the OCD in me doesn't want keys on a flash drive they're not associated with I just keep a folder on my PC called "UnRAID Keys", with subfolders named "GUID xxxxyyyyzzzztttt" (i.e. the actual GUID) ... and store the relevant keys in each of the subfolders. As with all of my documents, that folder is backed up in numerous places, so the odds of ever losing my keys are exceptionally low. And each of my UnRAID flash drives has a small Dymo label on it with the GUID (in very small print).

or you could put a green circle sticker over top of the triangle on your monitor! (just don't scroll ) :) ... I think a cache pool is "slightly" better By the way, I think the cache pool is one of the nicer improvements in v6. The primary reason I've not used cache drives in my servers is simply that I want files I copy to my fault-tolerant server to BE fault-tolerant when I copy them there. With a cache pool I can now use a cache to speed up writes and that will still be true

Agree ==> HOWEVER, the yellow triangle is NOT an "error" condition. It simply means "Some or all files are on unprotected storage." In the case of a cache-only share being stored on an unprotected cache drive that is the normal condition ... NOT an error. The user should certainly KNOW if he's storing data on a cache-only share ... and this indicator simply reminds him that the share is unprotected. Don't like looking at it? ... update the cache to a pool so it's protected

Agree !! It's nice to have the triangle that notes there are some unprotected files, but the user should certainly know if he's configured a cache-only share (and the cache isn't a protected pool). These indicators at one point weren't working right -- but they seem to be working perfectly now. Tested it with a cache drive and it changes between the triangle and green circle depending on whether there's data in the cache. I presume if I had a cache pool it would just always stay green. But you can definitely have too many states ... then they become kinda useless. I think just glancing at what's not green works fine.

Makes sense -- I'll do it right now.

If you want to remove the jumper there's no problem doing so, and you'll never need to replace it as long as you're using v5 or later. HOWEVER ... once you remove it you need to reformat the drive and all data will be lost => so be sure there's no data on the drive before you do this. In addition, you'll need to recompute parity (either with a new parity sync or just running a parity check (which will correct a LOT of "errors").

Always nice to confirm pass-through works well, as there are indeed occasional board/CPU combos that don't seem to "play nicely" with this feature. That's a very nice video card for that little ITX system. How close a fit is it for the Q25B ??

Yes, that's exactly right. As itimpi noted, you may then want to confirm the settings on your shares are the way you want them, but if you have a "plain vanilla" setup, it should be as simple as you've noted

Rather than trying to get too fancy with icon shapes and colors, I'd follow a couple of basic rules ... => Either use different shapes for each color (to resolve the color-blind problem) ... or, if you do have a common shape [e.g. green balls vs. red balls for disks that are good/disabled] be sure there's a differentiating inner symbol (i.e. an exclamation point in a red ball) => Most importantly, be sure every symbol has a fully descriptive hover text. In it's simplest form, you really only need two symbols => one for "all's okay"; and one for "you should read the hover text because something needs attention"

Agree. Seagate reports several values that other manufacturers simply don't show -- these are among those (along with high fly writes). You can ignore the raw values (ALL drives have errors that are corrected by ECC) ==> the only thing that matters is the "Value" ... if these are good, the drive's fine.

+1 for TeraCopy Used it for years, and really like both the interface and the features. I always use the automatic verification -- almost doubles the time to copy, but ensure the copies are good. And unlike Windows Explorer, if it encounters errors, you'll see exactly which files had problems AND the copy doesn't abort -- the rest of the data is still copied.

Haven't used the SilenX Red, so can't say ... 'nor have I used the Cougar units, but at $30 for 4 fans it's certainly intriguing ... hard to believe it'll be the same quality as some of the others you're looking at; but it just may be => every once in a while you do indeed find some nifty bargains. You didn't mention whether or not you'd try adding some baffles to block off the airflow deflection on the sides of the drive cage (like Daniel did). That certainly made a BIG difference in his case ... I'd think it's worth trying it => you may even find you don't need to replace the fans

The Gentle Typhoons are indeed very nice fans ... haven't done any wholesale replacements; but did buy one for a system that had the fan fail, and I've been very impressed at both the airflow and the noise level. Did you block off the area around the drive cage to prevent force the air path across the drives? [The mod Daniel did] From Daniel's experience, it really seems like that's all you need to do to get excellent airflow and good cooling. FWIW, running hot (your old 7200 rpm drive) is a very good sign that a drive is failing. Low 40's isn't bad; but if it starts running close to 50 I'd replace the drive.

PM me a link. Your changes sound interesting and useful. I'll review and incorporate the improvements. Joe L. Joe => Did you ever have a chance to look at these modifications and, if so, are they going to be incorporated? I noticed on the first post in this thread that there haven't been any new releases of cache_dirs, so I assume this hasn't been done yet. Just curious if it's in the works, as it does sound like a useful set of changes for x64 systems.

They may in fact be SMR drives (Seagate has released both 5TB and 8TB SMR units ... with a 10TB "coming soon") => but some very extensive testing of the 8TB units has shown they work quite well in UnRAID, so the technology by itself isn't the issue. But there have been enough problems reported with these 5TB externals that I'd nevertheless avoid them. With regards to the 8TB units, you may want to read this: http://lime-technology.com/forum/index.php?topic=39526.0 ... and (until Daniel adds his testing overview to that post, which he plans to do), you can read about the extensive testing that was done on these here (skim the thread and look for posts by danioj): http://lime-technology.com/forum/index.php?topic=36749.0

VERY nice -- I run VERY few add-ons, but this will definitely be one of them

High fly writes can be ignored ... they're not a problem. HOWEVER ... both pending sectors and offline uncorrectable are parameters you want to be ZERO. I'd replace the drive; reformat it; and use it for backups or some other off-line use.

Agree ... as I noted before, a backup is essentially just insurance against data loss. And as with any other insurance, you weigh the risk of loss against whether or not you're willing to accept the loss. Personally, I've spent thousands of hours and thousands of dollars compiling the data and media I have stored on my server -- and while I still have most of the original DVDs and could re-rip, re-render, re-index, etc. to reconstruct my collection, I have NO desire to do so. Bottom line: For my 45TB or so of data, a thousand dollars or so to have it all backed up is inexpensive insurance I'm more than willing to spend. And I didn't really even spend that much, as many of my backup disks are simply lower-capacity disks (1-2TB) that I repurposed as backups when I upgraded my servers to larger disk sizes. I agree that with dual parity a system is notably more reliable, and that can certainly be taken into account if you're evaluating the risk of loss and considering whether or not to "buy insurance" (i.e. backup your data). But higher reliability still doesn't provide protection against theft, fire, or other catastrophic system losses. My view is pretty simple: If it's worth building a fault-tolerant server for your data, it's almost certainly worth backing that data up. Period.

Definitely tempting the fates. Backups are just insurance ... and, like insurance, it's okay to not bother with losses you can afford to take. I suspect you have insurance on your home ... most of us wouldn't want to absorb that loss if we had a fire or major storm damage ... and on your cars (although that's less clear ... collision insurance isn't always worth maintaining on older cars); and perhaps on some of our more valuable jewelry. Etc. A lot of folks think nothing about buying those policies ... but don't consider just how valuable their non-tangible assets are -- i.e. your data. I always suggest one simple rule: Assume your server was stolen tonight and never recovered. If there's anything on it that you'd be heartbroken about losing, you should have it backed up. If you wouldn't care about the stuff on it, then it's apparently a risk you're willing to take. The cost of backups, with drives often under $30/TB, is actually very nominal compared to the cost (not just in $$, but also your time and effort) that you likely put into acquiring the data ... and in many cases may be irreplaceable. By the way, dual parity is still NOT a replacement for backups [it would, of course, reduce the risk of loss due to failed drives]

Perhaps there's more to it then it seems => but it sure seems like LimeTech could provide a tiny little Windows utility that could be downloaded and run against a flash drive to show the GUID

Nice find. Bookmarked that so I don't have to "re-learn" every time I use it (which is rare).