garycase

Actually, I kinda liked what I found when I checked out what toby said Found this definition for Australian slang: "Gaz - Used to describe someone of unparalleled skill ..." :)

Definitely confirms that a bit of airflow control makes a NICE difference. With that change, this is a near-perfect case for a mini-ITX system ... and with drive capacities of 8TB (and larger) there's not much reason to use anything else !! With a Plus system in one of these cases you can have 56TB of fault-tolerant capacity, plus a fault-tolerant BTRFS cache pool !! And with the nifty AsRock Avoton board with 12 SATA ports, you can do this with very low power consumption, and no need for an add-on SATA controller

Yes, I had the same thought !! Every once in a while you hear stories of large amounts of cash and/or jewelry found in items donated to Goodwill. I never could understand how that could ever happen ... but after this experience it's a lot easier to understand

I didn't know that either (about Syslinux editing) => it is indeed a really nice feature.

Trying to find the quote from "Gaz" you referred to, just to see what he said. Is that someone on a different forum? The only user close to that here is "Gazz" ... and that user doesn't have any posts relative to this.

I should have noted that there's no problem running virtual machines without vt-d ... as long as you have hardware virtualization support (vt-x), which the Avoton DOES have, VM's will run just fine => they simply won't be able to use passthrough. ... Just something to be aware of Other than that, the C2750 is an AMAZING performer => PassMark of nearly 4000 with a mere 20w TDP !!

The 8-core Avoton has enough "horsepower" (PassMark 3929) to handle a fair number of Dockers, but it does not support vt-d, which is important if you need to do device pass-through. I'd think this is a pretty severe limitation for VMs, so I'd recommend using a board that supports E3 12xx series Xeons instead ... this will also let you use CPU's with a good bit more "horsepower".

Glad I told you about the missing earrings !!! ... I told my wife I'd done that => and after cooling off a bit (YOU DID WHAT !! was the initial reaction) she just spent a LOT of hours the last 3 days going through EVERYTHING in the room she thought she had "stashed" the earrings in for safekeeping and ... ... Lo & Behold, she FOUND her earrings in a small bag she'd put in a zipper pocket in a very heavy Eddie Bauer down parka that I've had for years, but not worn in at least a decade => purchased when we lived in Colorado ... we now live in San Antonio, where it's NEVER cold enough for that. It was stored in an extra closet in our 2nd spare bedroom (NOT where she had initially thought). I suspect we may have moved some of the older stuff from the main guest bedroom to that room after she'd put them in that pocket, but before she thought about retrieving them. In any event, they're now in our safe -- which I bought shortly after she had lost the earrings so she'd have a safe place to keep her jewelry without having to remember where that was

Don't know if Daniel is using the filters (I'd think so, but don't know for sure) ... but he noted earlier that he's numbered his disks starting at the top => so disk1 is on top; disk2 next; disk3 next; etc. (only 3 disks currently installed). As expected, the uppermost disk (disk1) is the warmest.

Excellent results => and I agree that all you really needed was the skirting, which you did a very nice job of. I suspect my next mini-ITX server will be in this case ... with a bit of airflow skirting that I'll blatantly copy from you

It's probably somewhere close to my wife's 2-carat diamond earrings that she put somewhere "safe" since we were going to be gone for 3 weeks ==> that was about 3 years ago !!

... by the way, when you get the Noctua fans installed, it'd be interesting to see what happens to the temps. You wouldn't want to wipe out your array with yet-another pre-clear run; but running a parity check would be essentially the same test ... runs all the drives at once for many hours.

EXCELLENT results => that skirting REALLY did it's job. If you hadn't already ordered them, I wouldn't even bother with the higher speed fans ... although it certainly won't hurt anything to improve things even more. I agree there's NO reason to mod the drive cage !!

Good picture of exactly how the cage is configured. The biggest flaw is the complete blockage of slot #6 on the fan side => is that the plastic piece that you said could easily be turned into "swiss cheese" ?? ... if so, that's probably a good idea -- I'd remove it to do the drilling, however Other than that, I don't see anywhere else that you need to modify the cage - skirting the cage to avoid any alternative paths for the airflow will provide FAR more airflow over the drives than any mods you might do to the cage. If you want more, I'd do as I suggested earlier -- get a fan with a higher airflow rating -- rather than modifying the cage itself.

The pictures of the side of the drive cage on Newegg certainly look like there's ample airflow -- the unvented areas seem to be where the drives are mounted (thus holes wouldn't be useful ... and in fact the solid contact is better for thermal dissipation. But clearly actually having the case in hand makes it easier to see exactly where any additional airflow vents would be handy [e.g. didn't realize there were blocks in sections you can't see => e.g. the removable section for a graphics card.]

Looking at the detailed pictures of the drive cages on Newegg's site, I don't think you need more holes in the sides. If you've got the cage skirted so the only path for airflow from the fans is through the existing holes in the cages, you should be just fine. If you want to improve even more on the cooling capacity, replace the fans with higher rpm units. A 1500rpm unit would provide about 25% more airflow than the stock fans => and a high quality unit like the Noctua NF-F12 will run at 1500rpm with the same 22dba noise level the stock fans have at 1200rpm.

Nicely done. Anything that forces the airflow to be directly through the drive cage and eliminates "leakage" is going to help ... as you've found out. Doing this with aluminum instead of cardboard wouldn't make any difference => you're not conducting heat (where aluminum WOULD help) ... you're just blocking air leakage. Basically, given the fixed factors of the ambient temperature and the conductivity of the drive mounts, there is ONE remaining factor that determines how well your drives are cooled -- AIRFLOW. Anything you do to improve the amount of air flowing over the drives will help keep them cooler. There are two fundamental ways to increase the airflow: (1) Use higher CFM fans (2) Control the path of the air (as you've just done)

Icy Dock has used two different configurations that I know of in their 5-in-3 cages (possibly more). One required soldering the fan wires; one did not. The fan is removable => so just shut down; pop the fan mount off the rear of the cage; and have a look to see whether or not yours requires any soldering. If not, it's trivial to just pop in a new 80mm fan. If it does require soldering, it's VERY simple -- you simply need to cute the two existing wires and then solder the new fan's wires in place.

It uses a standard 80mm fan ... and the mount is detachable, so it's very easy to replace.

I didn't say "no reason" ... I said "no compelling reason" A couple of reasons to move away from Reiser ... => Reiser is no longer an evolving file system, since the primary developer is in prison for life (he killed his wife in 2006 and was convicted in 2008) => It has a maximum partition size of 16TB ... so it won't work with disks > 16TB Not a problem NOW, but clearly disk sizes are likely to move beyond that limit in a few years (they're already at 8TB, with 10TB and 12TB disks already announced) => Writes on Reiser disks that are very full (> 95%) can be very slow (actually the writes are still quick; but the file system can "stall" for a good while until it "decides" where to put the data). XFS doesn't have this problem. But Reiser is a very stable and reliable file system ... clearly it's working quite nicely with v5 (and has in fact been used since UnRAID's inception). There's really no reason to convert existing disks ... especially if they're already full and you're not writing new data to them.

There's no conversion tool => but the simple fact is there's also no compelling need to switch existing disks to XFS. If you WANT to do that, then there are several approaches ... => Add a new disk in XFS format; copy all the data from an RFS disk; then reformat the RFS disk ... and repeat this basic process until all are converted => Ensure all your backups are up-to-date; then reformat all of your array disks to XFS; and then copy all your data back to the server from your backups => Build an entirely new server with XFS disks; then copy everything from your old server to the new one etc.

Perhaps. Remember to post your mods when you are done. Just in Case! Anything's possible, but I really doubt that you'll have any significant problem with temps. The Lian-Li PC-Q25B has similar airflow, and the temps with 3 drives vs. the temps with 7 drives when I fully populated the case were not appreciably different (perhaps 1 degree higher on the upper drives).

I also like to see my drives in the 30's, but the reality is that the low 40's are not "bad" in any way ... especially on drives that have been active for many hours. These drives are rated to operate up to 60 C, so you're still WAY below the threshold. In short: I wouldn't be at all concerned with the temps you're seeing. I suspect they'll all operate in the 30's in normal operation (when all of the drives aren't spinning at once). If you want to drop the temps a bit, use fans with higher airflow ratings. ... but that's really not needed with the temps you're seeing.

If you sustain temps at that level, there's no reason to "fiddle" with the airflow I'm not surprised that the bottom drive is the coolest [heat does, after all, rise ] ... although it also hasn't been running very long, so it may run somewhat warmer after another hour or so.

How you do it isn't nearly as important as the simple fact that you're doing it (Backing Up) I got an ad from Seagate for their external drives which makes a very appropriate statement r.e. the importance of backing up ...