Raid5 vs snapraid vs snapraid w/aufs

  • I know were are getting pretty close to beating a dead horse. But considering that I accidentally wiped half the data on my server I'm gonna back up what I have and start fresh.


    I would like pooled data so that I only have 1 spot for files but I'm not sure this is really necessary in the long run.


    Raid5 seems the easiest and most standard. Could I use rsnapshot with that to help protect against accidentally deletions? Does having all discs spun up effect reliability of the drive? My network is generally pretty fast and faster is always better but I'm not sure the speed improvements from the striping would be that great or needed here.


    There are many cases of snapraid working. I just had a fluke, must not have finished syncing or something. I like that each disk is its own disk and I can pull one out and all my data is still there and readable. I think I would like it pooled as one big volume but not really sure its necessary.


    And that leaves me to snapraid w/aufs plugin. the aufs plugin honestly confuses me when trying to set it up, And I'm not sure how the pooling will effect snapraid in the event I need to rebuild.


    Use is media streaming w/plex tv,movies,music and backups of files and folders from computers.


    Thanks guys

  • That made it really simple.


    What happens if the system goes down. If you build a new server and set it up the same way will all of your data still be readable? I'm assuming aufs stripes and if you pull one drive out the data would be unreadable.

    • Offizieller Beitrag

    aufs does not stripe. All it does is decide which drive to write to. If you remove aufs, nothing changes. It just decides where to write. It also pools the reads. aufs presents NO danger. The worst thing aufs can do is fill one drive before the other.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Zitat von "ryecoaaron"

    aufs presents NO danger. The worst thing aufs can do is fill one drive before the other.


    I have reinstalled OMV 3 or 4 times because I didn't understood how aufs works.
    A couple of times it filled a single HD while in the others drives of the pool there was plenty of space.
    Can you help me (us) to configure aufs with a "recommended" setting ?
    Thanks in advance.

    • Offizieller Beitrag

    Do you have different size drives? Do you have the mfs box checked or unchecked?

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Yes, it has different sizes. My ex-NasLite box has:
    -4 x 2tb drives
    -2 x 1.5tb drives
    -1 x 1tb drive
    -1 x 250gb drive for the OS (OMV)
    For the aufs settings, I have left all to default. (Now I can't access the OMV box, but i think that the mfs box default is NOT checked).
    The last time, it was filled one of the 2tb drives.
    I have uninstalled the aufs plugin and I could access the data again. After the 4 OMV reinstallations I haven't lost any data, only this problem with aufs.

    • Offizieller Beitrag

    Which drives were in the aufs pool? If you did not check the mfs box, it uses pmfs which will try to keep the files in the parent directory until the drive is full. This is useful if you ever want to split the drives. Even with mfs checked, it will probably fill the 2 tb drives until they have 1.5 tb left. Then it will start swapping between the 2tb and 1.5tb drives. It won't use the 1 tb drive until all six of the other drives have 1 tb space left. It does this because it is picking the drive with the most free space (mfs) left.


    As for accessing the data, I'm not sure why you couldn't access the data or why you had to uninstall the plugin to access it. Deleting the pool should be enough.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Sorry that I've to ask...


    Where was your backup, when you lost half your data? RAID, aufs and others are not build for datasecurity, they're just for availability. Whatever you do with RAID & Co: Backup important data!!

  • All the drives were in the pool, minus the 250gb OS drive.
    The problem came out because I started a nightly backup with lftp.
    In the morning I have found the lftp job stopped with an error, so I have tried to free space moving data from the full hard drive to the others.
    I think that I have made the mistake to move the data from the physical shared folder, not from the "poolshare" folder (I used "mc").
    In the end it was all "messed up" and I don't remember what I have done precisely, I acted too impulsively, ruining the OMV installation.
    Unfortunately I don't remember the error code, and I haven't took any screenshot.
    Later tonight I will install again OMV with the drives rescued from my broken Synology Nas and I will try to check the "mfs" checkbox.
    This time I will use only the "poolshare" folder to move the data and I will not use the single drive's shared folder.


    Dennis, I think you are referring to "GTvert90", luckily I haven't lost any data, the ex-NasLite box (now working with OMV) was/is the backup for my
    Synology Nas (that has broken 3 days ago).
    Also the data on the broken Synology Nas is safe and accessible using an XPEnology "trick", and this morning I have discovered that all the data
    is accesible also booting with a Parted Magic usb stick (I thought that the Synology Hybrid Raid was inaccessible outside a Synology box, but I was wrong :) )
    The important data was also backed up offsite using external hard drives.

    • Offizieller Beitrag

    aufs is kind of touchy when moving data around on individual drives especially when you have udba set to reval (which is default). So, everything should be done with the pool.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!