Nerva
-
Posts
48 -
Joined
-
Last visited
Posts posted by Nerva
-
-
I had this happen to me some years ago...
...and the support here ultimately had to have me reset all my settings in order to fix the problem, so I find it very disappointing that this bug still hasn't been fixed.
I had a new drive suffer an "issue" and some sectors (a few kb) were damaged that had files on them. Since I had file duplication, I actually just went and deleted the damaged versions and reran the scan, which keeps coming up clean. Except Scanner continues to pop up warnings telling me the now non-existent files are still damaged. When I rerun the file scan, it finds nothing wrong, yet the warning continues to pop up. This is exactly what happened previously.
This is a really stupid bug -- you guys need to fix this.
-
Forgive me if I'm using the wrong terminology here, but I'm curious about this. I'm currently using pool duplication, where every file is stored on two disks. I'm curious if there's a feature where instead of storing it on two disks, it stores every file on three disks, and then periodically Drivepool compares the three copies, and if one of them is different, it corrects the odd-one out? I think this is different from the typical parity scheme, but I'm not that well-read on those concepts.
-
-
No, disk management says it is a "basic" disk (and since it "convert to dynamic disk" is not grayed out, that means it is not currently a dynamic disk).
No, it is NTFS with 64k clusters like all the other DrivePool disks.
The fact right around the time I reinstalled the drive in the server there was a notification having to do with the pool it used to be a part of is extremely significant IMO.
I SUSPECT what might have happened a year or two ago when I pulled the drives out and deleted the pool is, FIRST I pulled the drives, and THEN I deleted the pool, and I suspect there's a bug that caused it to not delete the association with the drives when I deleted the pool, because the drives weren't connected when I deleted the pool.
-
Thanks for the suggestion. The thing, though, is that while my home server is rarely running at max load, it is also rarely running at idle, since it does a lot of low-intensity tasks 24/7, such as bit torrent and PVR.
-
I'm curious, perhaps the way to go is to buy a used mobo that was the last/best low-power architecture... I've bought used computer hardware from eBay before and had no problems.
I presume the two product lines to focus on were the Intel Atom and AMD APU -- does anyone know what the last/best mass-market versions of those were for home server applications?
-
So, eight years ago I built my first dedicated home server, based on a Norco chassis, which has 20 hot-swappable drive bays at the front, another two internal 3.5" bays, and a slim optical drive bay. I thought I could avoid the expense of SAS and just use a bunch of dirt-cheap SATA controllers, which had served me well in my previous server that was a repurposed tower PC. I went with an ASUS E35M1-M Pro motherboard with integrated AMD E-350 APU (18W TDP), combined with a SYBA 4-port SATA PCI-E card and a SYBA/IOCrest 8-port PCI-E card. Since I'm finally starting to max out the number of SATA ports I currently have, I just ordered another of the 8-port cards, which will replace the 4-port card.
Anyways, the E-350 based server has done its job pretty well for 8 years now, but obviously it was extremely slow to start with, and given its age, I was thinking I'd look around and see if there is a modern upgrade for it? What has me scratching my head, though, is that from what I can tell, the Intel Atom and equivalent AMD APU's have essentially disappeared except for laptop platforms. I see laptop Atoms that are x86-64 with better performance and lower TDP than the E-350, but they only seem to be made for mobile platforms.
I'm wondering if there are any good options these days for very low power home server motherboard/CPU's?
-
So, I used to have three pools -- one was a pair of 1TB drives, the second was a pair of 2TB drives, and the third was bunch of larger drives. The need for the second pool of 2TB drives went away, and I didn't have any immediate need for them, so I thought I deleted the pool (it certainly isn't listed anymore), I reformatted the two 2TB drives, and pulled them out of the server. Now I've decided I want to use one of the 2TB drives for extra space in third pool, so I put it back in the server, full-formatted it again to be sure it was still OK, and went to add it, but it doesn't show up under the list of available drives in DrivePool. Stablebit Scanner sees it, Windows sees it, and I can copy files onto it just fine, but DrivePool seems to have some sort of "mental" issue with it. I suspected it could be due to having been previously used in a different pool that should no longer exist (when I toggle through the pools, there's only the first 1TB one and the third one), and what's interesting is under "Feedback" it has a strange message that says there was an error measuring the 2nd pool, and the date was probably the same date as when I booted up up the server with the drive put back in -- so, indeed, it appears that DrivePool still thinks that drive is somehow associated with a pool that should no longer exist.
Any advice on how to fix this?
-
I keep having a problem with my home server -- after starting a download of a large (20+ gigs) new torrent in my "Downloads" pool (just a pair of 1TB drives mirrored to each other), the entire server is unresponsive on the network. When I check Stablebit Scanner, I see a high amount of disk activity between the mirrored drives. I use the "preallocate torrents" option, so the instant a large torrent is created, it pre-allocates the file on the pool, and I think the particular way that is done causes DrivePool to "hog" the system when it tries to immediately duplicate the preallocated space. I'm wondering if maybe the torrent software, when it does the preallocation, is actually doing a ton of discrete allocations, and this "hogs" the system.
-
Since I posted that, I ran the file system scan a few times and it found some marginal files that were recovered, so now the file system checks out fine. However, it keeps saying the disk is damaged because there's a bad sector -- shouldn't it be able to just mark that as bad, remove it from use, and clear the error flag?
-
So, I got a message saying that one of my older disks is damaged and I may have lost data. Fortunately, I use DrivePool with duplication. I had Scanner scan the disk and the file system, but it kept showing the error saying it was unable to read one cluster and I may have lost data. So I ran the windows checkdisk with the option to scan for bad sectors -- it found some problems -- but Scanner continued to say the disk was damaged etc. I reran the Scanner scan of the disk and file system, but it still showed the error. I got fed up and removed the drive from the pool and reformatted it (not using the quick format option). Scanner continued to claim the empty disk was damaged and scanned the file system. I've since re-added it to the pool again, but Scanner just won't stop saying the disk is damaged.
-
Is there a way to check what version I'm currently running? I can't seem to find it anywhere in the UI...
-
Thanks. It is Windows 7 x64.
-
I'm not sure what happened exactly, but I booted up my home server and noticed in My Computer that my pooled drives were half empty. I soon realized that much of the duplication had somehow disappeared. In theory the entire pool should have duplication, but currently (DrivePool is only 65% rechecked as I'm writing this) the legend under the pie chart show it has ~8TB duplicated, ~13TB unduplicated, and ~17TB "other", but there's no "other" shown on the pie chart itself. Also, My Computer says the pooled drive is 35.4TB with 14.4TB free. The fact My Computer is reporting that much free space (there should be less than 1TB free) is what really has me concerned...
-
I ended up moving all the drives onto other (faster PCIe) controllers, but if/when I add more drives and need to use the Supermicro PCI controllers again, I will give that a shot.
-
Scanner says it has over 6 years of use on it, so it is wayyy out of warranty and has lived a full life so to speak.
-
I guess I might as well run the drive until it dies. In the past, with my main archive pool, when one of the drives goes offline, the whole pool becomes read-only -- but is that still the case if there is full redundancy? If the bad drive fails and the DVR pool becomes read-only, then I won't be able to record shows until I remove the bad drive from the pool.
-
Well, I guess it is a good thing I am using the drives for DVR -- that's the least-critical data on my server -- it is why I use my oldest drives for today's rerun of Seinfeld and the local news report.
The problem is the DVR workload, not the use of pool duplication. I record up to 200 GB of shows every day on a 2TB disk that is extremely fragmented, and obviously the 2TB drives are not handling it well. Basically it is a write many times, read occasionally application.
-
So, I've had a home server for over 10 years, and over the years what used to be six 1TB and four 2TB drives have steadily failed until I now have only three of each, but I have of course added newer, larger drives that now store the bulk of my data.
The older, smaller drives of course use more watts/TB and take up SATA slots, but small drives still have their niche uses. I decided some months ago to take the remaining three 2TB drives, which had been used for long-term storage of "stuff I want to keep" and repurpose them for SageTV DVR storage -- using two of the drives in parallel for redundancy, and pulling the third drive out of the server to keep in reserve as a spare, so that I can expect to have a redundant 2TB DVR system for years to come.
But since the reshuffle, one of the 2TB drives has steadily reported bad sectors in Stablebit Scanner. It probably had less than a dozen bad sectors when I switched it from archive duty to DVR, but now it has over 100 -- I eventually started logging the trend:
Date #BS
2016-10-25 32
2016-10-29 372016-11-01 402016-11-03 432016-11-06 452016-11-08 472016-11-09 482016-11-14 572016-11-17 692016-11-20 722016-11-21 802016-11-24 842016-11-25 982016-11-28 1012016-11-30 114So, it is typically about 1 new bad sector per day, but occasionally there are bigger jumps.The other 2TB drive has 58 bad sectors, but it already had around 55 when I switched it over to being a DVR drive. At first I thought the other drive was just "catching up", but now it has blown past its brother and has twice the bad sectors.I'm wondering what the significance really is of this trend in bad sectors, and if anyone wants to speculate on what the underlying cause is? -
Any update on this? I had to suspend automatic scanning of drives because Scanner thinks the imaginary extra space is bad sectors.
-
Seems to me that a "Stablebit Defrag" would be a great idea for a new product -- it would complement both Scanner and Drivepool rather well -- and if they're all developed by the same company, in theory there shouldn't be any fear of bad interactions between them.
-
I've read that with files being as large as they are these days and with HDD's having NCQ, that defragmentation offers little to no benefit, and just puts more wear on the HDD. Is this true? If not, what are the best defragmentation programs these days?
I forget if it was the old Norton or Central Point, but back in DOS times, there was one that actually had a "prioritize programs and folders with full file reorder" option -- it would stick all your directory info and .exe's at the front of the drive, and literally reorder every file and every cluster on the drive from end to end, in one long session. Probably overkill, but it was cool to watch.
-
-
I have been experimenting with reshuffling what drives are connected to what controllers quite a bit the last few days -- could that be the cause of drives "showing up a bunch of times"?
I assigned my DVD drive the A drive letter and an external HDD is using the B drive letter. Eventually I intend to attach the external HDD to the router rather than the server.
At the moment the Supermicro controller has the following HDD's attached:
D - Western Digital 1TB
E - Western Digital 1TB
F - Western Digital 1TB
H - Seagate 2TB
I - Seagate 2TB
J - Seagate 2TB
K - Western Digital 3TB
I am planning another complete rearrangement of all my drives in a few days when two more 8TB drives arrive, so please let me know when you're done with comparing that list with the BitFlock info.
What I figured out from my experiments is, even the oldest drives are 4x faster than what the Supermicro can keep up with, and the oldest drives do the most work (they're used for short-term downloads and SageTV, while the larger drives are used for storing my blu-ray movie collection), so I'm going to put the oldest drives on the PCIe controllers and the new 8TB drives will end up on the Supermicro.
Scanner continues to warn about damaged drives after problem no longer exists
in General
Posted
So, isn't the HDD supposed to take those clusters out of use if they're bad? Is there a way to tell if it has done that?
Another thing that would be cool is if Scanner and Drivepool worked together on these sorts of things, where if Scanner detects a damaged file that has a duplicate, Drivepool will then remove the bad version and replace it with a new duplicate that is based on the duplicate that wasn't corrupted.