beefjerky

I use standard spinning hard drives for my drivepool. They can certainly saturate my network while moving larger files, as they're very fast for sequential transfers. However, for syncing and moving a bunch of small files, hard drives suck and they can slow to a crawl. An SSD would greatly help with those transfers, however, it doesn't look like it's really an option. I did hear back from support via a help ticket I put in, and they confirmed that the SSD optimizer would be of no help for my situation with transfers to/from a Veracrypt drive. Also, the size of the Veracrypt container precludes affordable SSD storage. This is especially the case since I really need that file to be duplicated. Having the container on a single SSD and duplicated to a HDD also wouldn't help since DrivePool handles the duplication in real time, so it would still be limited to HDD speeds. So, I'd need two very expensive SSDs. Regardless of whether Primocache can handle network transfers on a server, I'll avoid it. The data in the container is too important to risk corruption from a power outage or crash. So, for now, I'll just live with it. When 3+TB SSDs inevitably become cheaper, I'll reassess.

Hmm, I ended up contacting support for Primocache, and they gave me a different answer. They said it would work to accelerate transfers on client computers if run on the server. That said, they did tell me that regardless of whether using an SSD or RAM as cache, there was no protection at all from crashes or power outages. So, if the server suddenly loses power or crashes, that data is permanently gone. I was not comfortable with that. The SSD cache option for DrivePool makes it clear that no data would be lost in those scenarios. However, since no one from the development team has responded, I'm still not sure if it would work to do what I want. And, I can't really think of an easy way to test it either...

So, I have a question about the SSD Optimizer and a very specific use case. On my drivepool, I have a duplicated 1.5TB Veracrypt container. This container contains personal data that I don't want to leave open to just anyone. I regularly update the contents of this container via Beyond Compare running on my laptop. So, here are my questions: 1) From reading the description of the SSD Optimizer, it only applies to newly written files or overwritten files. If my understanding is correct, this will do nothing for optimizing writes inside of that Veracrypt container. Is my understanding of this correct? 2) I see some here mentioning PrimoCache, which is a more low-level caching software. However, I've seen a post or two here that says it doesn't work over the network, but need some clarification. It seems that it will do nothing if running on a client computer accessing a network drive on the server; this makes sense. However, if running on the server, will PrimoCache accelerate writes to the server from a client connected via a network share? Can anyone confirm that it would and works for all writes, including something like my use case? Thanks in advance!

So, I thought I'd update this. Tonight I shutdown the system to install a TPM module, and when I rebooted the system, DrivePool magically measured correctly. No other hardware changes other than the TPM module, and I am sure that Windows did not re-detect the disks or anything like that. Prior shutdowns/restarts didn't fix it, so I have no idea what interaction might have occurred between the TPM module and DrivePool.

I can confirm that each drive only has one poolpart folder. Thanks!

I tried doing that, and no change. The two disks still show as other. I also tried having Drivepool recheck the duplicate folder, and no change there either. I can confirm that all the files are in the PoolPart folders; nothing is outside those folders except for the NTFS system folders. Thanks.

Bump.

I have tried remeasuring multiple times, with no change. Thanks.

Two of my drives are suddenly showing all the data as other. The data on them does seem normally accessible via the pooled drive, so it seems to be cosmetic, but still annoying. These drives are both 4TB drives, and they both hold duplicate copies of a folder. One of the drives also contains a whole lot more data in the pooled folder. I've tried remeasuring with no success. I updated to the latest release version with no improvement either. This is running on Windows Server 2008 R2, which is also fully up to date. Note that I have a backup server with the same OS, same versions of Drivepool, as well as the same configuration of pooled drives (2x Toshiba 3TB, 4x HGST 4TB and 4x Toshiba 8TB), as well as the same duplication settings. No problems on that server. Thanks for any help!

Sorry for not being clearer. This was actually a test run to see how things would behave with missing disk(s). I'm in the process of setting up another NAS for backups and file sharing with friends/family, so I needed to see how Drivepool would handle these kinds of things. Disk Management comes up and responds fine when disks are removed, whether it happens while the system is running, or prior to boot. There are no errors from either Drivepool or from Windows volume manager/virtual disk manager when a disk is missing during bootup. I can still try to upload the system event logs you think if it would help. Also, this is on two different Windows 2008 R2 servers; my primary one, and the new backup NAS I'm setting up. Thanks for the help!

So, I thought I'd post an update since I finally managed to get it fixed, but it was a weird one. First, running the chkdsk command with the volume name you gave in the system recovery did not work; it gave a volume could not be opened error. After another couple of tries with chkdsk /f c:, I tried a simple chkdsk c: in Windows, and was finally able to see the error about the volume bitmap. The error only comes up in read only mode. I also discovered an error in Event Viewer from the volume shadow service that would occur anytime I did the chkdsk in read only mode. It basically said that the drive for shadow copies of c: was set to the e: drive, and that drive was not NTFS or an error occured while trying to detect it.. Drive e: is my drive created by Drivepool. I have no idea how it got set to that drive, but perhaps Windows made the decision to do that at some point prior to me using Drivepool due to more free space than I have on the c: drive? At any rate, I set the volume shadow service to use the c: drive for shadow copies of drive c:. Apparently that's supposed to be the default anyway, so again, I have no idea how it got changed to e:. Well, once that was done and I rebooted, chkdsk in read only mode no longer showed the volume bitmap error, and now Scanner finally shows the file system as healthy. Long story short, weird Windows issue that shouldn't have happened, but did.

During a normal boot, the Drivepool UI is available as soon as the Windows UI is up and ready. However, with the disk missing upon bootup, the Drivepool UI will not properly come up, even after waiting around 20 mins. However, the pool (at least what's on the remaining disks) is readable once Windows is up and ready, just like when all the disks there. I have 6 disks in the pool. Would it help for me to upload a log after a boot with a missing disk so we can see what is going on? If a disk is bad and unavailable upon boot, this is really something I need to be notified of, and as of now, this isn't working.

If a disk goes missing while the computer is running, Drivepool handles this as expected. It pops up a notification and e-mails me, and I can take action. However, I've noticed its handling of a disk that's missing upon bootup is less than optimal. When the UI is started, it simply says that the service is not started, though the service is actually running. What's on the disks still available also seems to be readable. And, of course, no notification of any sort, which is a problem on a headless server. I am running v2.2.0.651 BETA. Hopefully this will be fixed in a future version. I'd also like to make a request for both Drivepool and Scanner. I'd like an option for a when a notification of a drive problem or missing disk appears, a sound can be made through the motherboard speaker. This could be helpful for headless installations where someone other than me may be using it or I may not have my phone by me.

Any further ideas on my issue of why the chkdsk doesn't seem to be running on bootup? Interestingly enough, it finally got around to scanning the other disk in the mirror and is actually showing healthy, which is a first. But it still shows damaged for the other one. I went ahead and uploaded the most recent log with the other mirror disk showing healthy. Not sure if the comparison will help or not.

Using mountvol, it simply shows it as as C:\ When executing that chkdsk command, it shows that the volume is mounted and a system disk and offers to run it during the next system boot. However, it never does, nor does it show anything in event log, even an error. Now, I notice two different long strings that you posted, one ending in c1b9, the other 6963. What is each referring to? Thanks!