Jump to content

AlexL

Members
  • Posts

    29
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by AlexL

  1. x-post

    Hi Christopher,
    I am aware of the command line options, but deem it an ugly solution since the reason for using Windows for NAS is to get away from the cmd line hassle ;)

    Since Drivepool already scans the filesystem like all the time, I would like to propose creating .txt files as place holders every time Drivepool moves a file.

    BR,
    A

  2. Hi Christopher,
    I am aware of the command line options, but deem it an ugly solution since the reason for using Windows for NAS is to get away from the cmd line hassle ;)

    Since Drivepool already scans the filesystem like all the time, I would like to propose creating .txt files as place holders every time Drivepool moves a file.

    BR,
    A

  3. Crossposting here from HW forum, for visibility.

    Quote


    Yesterday one of my disks turned to RAW, then the partition turned to a blank partition, so I had to forcefully remove it from DrivePool!

    Since its NTFS, I downloaded TestDisk ( https://www.cgsecurity.org/wiki/TestDisk_Step_By_Step ) and repaired the partition list. After a reboot, drive turned up all good and drivepool picked it up without a hitch!

    Thank god for keeping it simple with NTFS and DrivePool!

     

    EDIT:
    One of the things that really annoyed me with DrivePool was that I couldn't see what was missing once the drive took a dive. You have something in the works about that @Christopher (Drashna) ?

     

     

  4. Yesterday one of my disks turned to RAW, then the partition turned to a blank partition, so I had to forcefully remove it from DrivePool!

    Since its NTFS, I downloaded TestDisk ( https://www.cgsecurity.org/wiki/TestDisk_Step_By_Step ) and repaired the partition list. After a reboot, drive turned up all good and drivepool picked it up without a hitch!

    Thank god for keeping it simple with NTFS and DrivePool!

     

    EDIT:
    One of the things that really annoyed me with DrivePool was that I couldn't see what was missing once the drive took a dive. You have something in the works about that @Christopher (Drashna) ?

  5. On 11/15/2017 at 9:02 PM, ubern00b said:

    Hey AlexL

    I have this enclosure too. Working great for me so far, just a little noisy as you have found out! Thanks for the info on the fans, I will be buying the same ones and doing some surgery. Just a question for you though. 

    Are both fans 3pin? I was wondering if they were PWM and maybe that's why the 80mm wasn't spinning up? Or maybe it's just a higher voltage needed to get it spinning?

    I assume the enclosure does not mind if a fan is not plugged into the PSU fan header? Did you simply split the 140mm fan header to both fans?

    Thanks!

    Hi!
    After a few months one of the drives went haywire making the whole array to behave inconsistently, but a good old "chkdsk c: /r /scan /perf" brought it back to life.

    As for the fans, they are 3-pins yes, and worked fine in other mobos, but I guess the pins on the power supply wasnt outputting what needed to be outputed..

    The enclose doesnt mind if a fan doesnt exist inside the PSU no. I took the cable from the PSU, split that for both fans to work. This does mean that BOTH fans stop working if the unit is plugged in but without any of the disks are turned on, so use with caution!

    BR,
    Alex

  6. Sorry, I panicked in the dead of the night and didn't get a chance to explain. Let me rephrase my user story.

     

    Drive Pool:

    10x NTFS drives (Icybox diskN)

    1x Storage Spaces (SSD Pool) - containing 2 SSD's one Intel, one Micron.

     

    Balancers:

    SSD Optimizer - SSD Pool selected as SSD, rest as archive. Fill SSD and archives to 99% 

    Scanner

    Duplication space optimizer - Not sure if i enabled this after the fact.

    Balancing was set to 70%, no real time duplication enabled, no file rules, balancing immediately.

     

    SSD pool was empty in the morning, started disk image upload via SMB, verified that I was writing to the SSDs and left for work.

    17:00, I come home, I had managed to move 200gb by 14:00 or so the logs say, no balancing activity was being done on DP, even tho I did have more than 200GB left on a few drives. 

     

    I start my usual test of copy a big file from non-DP system ssd disk to the storage spaces and see the transfer rate go UP (Intel slab) then it goes down (Micron slab) to 30mb/s from 230! I want to troubleshoot this, naturally so I click remove SSD and it starts moving the files from the SSDs to the archive disks.

    In the bkg the mac keep trying to reconnect to continue backing up and all of the sudden I get an error msg on the mac saying 'access denied' on the upload. Odd, so I verify if I can create a new folder in the share which is a NO although this folder has been duplicated many days ago to several drives. I logon to the machine itself and try creating a folder withing the same folder, access denied. I can see the duplicated folder on a few drives so i should receive any errors should my writes/reads now go to the same folder on other drives. So I panic, write this...not well worded error report...and run the troubleshooter emailing as drivepool@l......e

     

    After a while DP manages to move all the data off from the SSD Storage Space, I check the Micron and no SMART errors and no weirdness, just super slow. I test another USB enclosure and I get the same slow speed. I replace that disk and install to the main DP which seems to work.

    Since I need the files OFF that ssd storage spaces pool, I set the Balance ratio to 1%, I also disabled Read stripping because I read somewhere it might be slowing me down. And a bunch of other radical changes as the default behavior obviously doesn't suit me.

     

    So thats my user story, along with full logs, at least those I could get, I wish I had the logs from the whole day before.

     

    As for the balance ratio, I get the desired super aggresive behavior of having a balance ratio of 1% or 1gb, but I should set it to 100% instead ?

     

    Now your answer indicates I should have 2x SSDs selected , so I am reading:

    "If you are using duplicated files, then you should specify multiple SDD disks or else the system will fall back
    to an Archive disk for one or more of the file parts."

    I am using file duplication, but the problem seems to me that I was writing big files to the SSD, that where never balanced off, or duplicated for that matter. And as soon as I clicked remove it felt like DP went all "read-only!!" on me.

  7. Are these SSDs in a Storage Spaces array, as well as in the pool?


    If so, that ... may be why your'e seeing the performance hit. 


    > 1 of the disks was bad


     


    Also, what version of StableBit DrivePool are you on? > 651 BETA


     


    And was the drive still being removed when you started getting the access denied errors? > Yes, sent all logs through the tool as well.


    Thing is, balancing is now super aggressive so the files should've been duplicated/moved but seemed to be only on the ssd storage pool.


    However, this for me is a huge show stopper, that I all of the sudden cannot access my own drives.


     


    Should I try STABLE (ie 1.x) ?


  8. So I noticed severe drops in speed on my storage space SSDs, and wanted to check it out. My disk images where doing their thing to archive folders [since the ssds had not been balanced out for the whole day, another strange issue], and when I removed the SSDs from DP, Trueimage comes back saying it gets access denied. Trying to create a new folder on that windows share gave permission errors.

     

    DP now has permission denied from within the OS as well.

     

    Troubleshooter uploading now to get you guys some view, but so far it seems I might have to stop using DP when the trial ends :(

  9. Crash in beta, full lockup even...

     

    Huge incoming file, going from realtime dup to nothing crashes the UI. Tried restarting the svc - locked up the whole server.

    guessing it was moving files in the bkg, and i cant touch it until its done, even with the file transfer stopped :)

     

    thnx for answering!

  10. An update here, since even having 1 ssd drivepool in the main drivepool as cache wont work (i guess), i made a storage space of the 2 ssds and added them to the main drivepool.

     

    Now I got 330gb of SSD write cache. But I need to force DrivePool move EVERYTHING from that "disk" (again 2 ssds in SS) as soon as I am done writing to the file to make room for the next file. How do I achieve this Christopher ?

  11. For posterity's sake:

    So I have worked with this enclosure for quite some time without any hitch, achieving the max that my old drives can achieve (tested without drivepool, 50mb/sec).

    Now for the box itself, you get it with 2 fans, one 140mm [placeholder for name and specs here] for the harddrives and a 80mm [placeholder for name and specs here] for the power supply, that had about 20 more dB output that I hated, living in an apartment and all.

    So I replaced these with a "Noctua NF-P14s redux-900 140mm" where the installation without a hitch, and a "Noctua NF-R8 redux-1200 80mm" where I found that it could not be started without help, so I bought Y-cable connected to the 140mm as well (https://www.kjell.com/se/sortiment/dator-natverk/datorkomponenter/kylning/adapterkablar/grenkabel-for-datorflaktar-3-pin-p98538) because the pin connector ON the PSU won't be able to drive my 80mm Noctua(!).

    All in all, the sound level in my room went from 56.3 dB to 39.6 db, which is like a breeze for me. The HDD's are at a stable 26C even at high load under a long time, which makes me confident buying fan regulators from aliexpress..

    Next up, I had to rename all drives "Icybox <drive bay> (<version of the drive>/<size>/USB version>)" to make sure I knew what to hotswap in the unlikely event ( :D ) that a disk goes nuts.  "version of the drive" to not confuse the disks if they are ever connected in the same computer...

    Snapraid is installed, as well as Elucidate for the GUI, with 2 parity drives for my 10 disk array, one disk being same size as the biggest in the array. I ran this the first time, and plan on running it once the balancing is done. 

    Speaking of fail-safe, I also run Duplication in DP for now, just for the piece of mind until I have fully tested the snapraid app in a recovery mode.

    I have 2 SSDs for the RAID write cache. For this I installed SSD Optimizer but it didn't kick in until a restart.

    Thanks!

    A

  12. Hi,

    Is there any way of getting API access to at least the balancer API ?

    I am thinking of 2 examples I'd like to code;

     

    1. Snapraid trigger

    As soon as a Balancing activity has ended, start snapraid check

     

     

    2. File placement "list"

    As soon as a Balancing activity has ended, create a folder á la https://blogs.dropbox.com/business/2016/04/announcing-project-infinite/ symlinking to the "real" folder.

    Coincidently, this removes the need to rebuild the driver as you stated in another thread.

     

    This can also mean that a tiered storage can be achieved by moving less-acessed files to slow storage, and often-accessed files to SSDs.

     

     

    Thanks for checking in :)

  13. I have an IcyBox 3810u3 (10disk USB unit) that is working great for me. But both the Scanner service and the UI are lagging so much, that I can barely use it. I did manage to do a full scan on 10 drives, but that took almost 2 days.. removing the temp and disk pool columns helps for the most part but the scanner still freezes.

     

    Running Windows 10.

     

    Thanks,

    A

  14. An update seems in order. It has dropped USB once and made my disks RAW because one of the drives is dead. I formated the disks with GPT now and checking scanners SMART as well to be sure. Stress testing all drives one by one doesnt seem to bring the interface down, so I guess it was just a timeout on the drive causing this.

     

    Also, I will be replacing the 120mm fan, because that thing makes me dream of F35's! ;) 

  15.  

    First, I'm sorry to hear about the failed drive, as that is never a pleasant experience. 

     

    This topic is covered in detail here:

     

    http://community.covecube.com/index.php?/topic/746-drive-must-have-died-is-there-a-way-to-tell-what-was-on-it/

     

    If duplication was enabled, then nothing is lost. Just remove the missing drive and let it reduplicate. 

     

    If it wasn't duplicated, well, if you have an inventorying software (such as sickbeard, sonarr, radarr, etc), just rescan the library, and make sure it's set to monitor these folders.  The software should go and re-download the missing content. 

    (In fact, that's why I personally recommend this software, even if you're ripping the media yourself, as it is fantastic inventory software)

     

     

    As for "for the future",

    http://community.covecube.com/index.php?/topic/1587-check-pool-fileparts/

    http://community.covecube.com/index.php?/topic/1865-howto-file-location-catalog/

     

     

     

    See above.

     

    1. This can help, definitely.  And for adding data, it will definitely end up "together".

       

      As for the folder level splits, just because Alex doesn't see a use for it, doesn't mean it won't end up in the product.  I've stated that repeatedly.   And more to the point, Alex is the developer and has a far better idea of how the software works, more so than myself or you.  But yes, he's not infallible.  And again, that's why I keep pushing this topic.   We've discussed it internally EXTENSIVELY. And it is possible that it may make it in a 2.3 or 2.4 release.  

    2. Not a bad program. There are a bunch of stuff that will do this, as well.  And "alternativeto.net" is a great resource for finding software. 

     

     

    Can't wait! :D

×
×
  • Create New...