IIRC, it is recommended (or at least, a common convention) to keep a copy of the content file in /var directory. The plugin doesn't seem to allow this.
Is it possible to have, say, a checkbox on the plugin configuration to put a copy of the content file in /var/snapraid or somewhere similar?
We verified with the snapraid author that this is not necessary. You can put the content file on all of the data drives and even the parity drive. No need to put in /var/snapraid
I know it's not necessary. I just thought it might be a useful feature. My OS is on an SSD (which I'm assuming is quite common)... I wonder if having the content file on an SSD makes any kind of speed difference.
Hi,
in a few days i will setup a media server with video files for xbmc and music files for squeeze box touch.
In this way i think first to make a raid 5 with omv to have pool for samba sharing to xbmc clients and squeezebox touch as client for logitech media server.
But after reading the explanations to snapraid and its intention for big and not so often changeables files i think its a better idea to go with snapraid and aufs plugin.
The media server will have one ssd for omv and four 4 TB WD RED, three for Data and one for parity. To setup procedure i will follow up the steps of solo. But the question for me is to built the snapraid array over hours before the setup of aufs or at the same time in the WebGui of OMV.
Excuse me for the bloody beginner question.
In this way i have a big petition: Would it be possible to got an howto snapraid with aufs and samba shares
Thank you all for the great omv project.
Gerald
Hi,
thank you very much for this advice ryecoaaron. I will follow these steps to built my media server.
Gerald
ryecoaaron,
For AUFS it says to create a folder called poolshare on any of the data drives. Stupid question but do we put any data in the poolshare folder or is it just a reference point.
For em I have 3 data drives with:
disk1: User shares, Music, Photos
disk2: TV Series
disk3: Music VIdeos
So if I put a folder called poolshare on say disk1 how will that point to say TV Series on disk2?
Kryspy
The pool share is how you reference the pooled drives in other plugins. Step 6 tells you to use the poolshare in other plugins such as samba. Technically, it is just a mount point but all of your pool files will be available there.
Okay, Label me thick but I can't get it to work. I can see a poolshare added to /media. But when I create a pool with the name poolshare and my 3 disk (labelled disk1,disk2,disk3) and poolshare folder within disk1 as thew bind point my shares are empty. I am using disks with data on them so they aren't empty. If I create a share of the actual folder all is fine but won't get me pooled data.
What am I missing?
Kryspy
Did you update to the latest aufs plugin I just released a few hours ago? Check Updates. It has a fix that required a change with OMV.
Running 0.5.5
Does anything go in the poolshare shred folder? I am not getting how this will pool data? Under ubuntu you merely created a directory of say poolshare and then added disk1,disk2,disk3 to it in fstab.
Kryspy
Post the output from:
cat /etc/fstab
# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a
# device; this may be used with UUID= as a more robust way to name devices
# that works even if disks are added and removed. See fstab(5).
#
# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0
# / was on /dev/sda1 during installation
UUID=55b7e576-676a-4b6d-9f61-50a3820e7088 / ext4 errors=remount-ro 0 1
# swap was on /dev/sda5 during installation
UUID=b20d3437-73aa-481b-99b2-68ce0d0f26c8 none swap sw 0 0
/dev/sdb1 /media/usb0 auto rw,user,noauto 0 0
tmpfs /tmp tmpfs defaults 0 0
# >>> [openmediavault]
UUID=d799002c-71e4-412a-ba79-d71b58c2d98d /media/d799002c-71e4-412a-ba79-d71b58c2d98d ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2
UUID=1fa4c8a2-d4c3-48cf-bb08-ec5669b457dc /media/1fa4c8a2-d4c3-48cf-bb08-ec5669b457dc ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2
UUID=f24372ec-2499-4bf8-9a0b-17bfe8ea616d /media/f24372ec-2499-4bf8-9a0b-17bfe8ea616d ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2
UUID=148c4c78-ff13-472c-a1f6-71fc64b47085 /media/148c4c78-ff13-472c-a1f6-71fc64b47085 ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2
UUID=49196c8f-cb29-4184-8bd7-98f0aee798cc /media/49196c8f-cb29-4184-8bd7-98f0aee798cc ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2
none /media/poolshare aufs br:/media/1fa4c8a2-d4c3-48cf-bb08-ec5669b457dc/disk1/=rw:/media/148c4c78-ff13-472c-a1f6-71fc64b47085/disk2/=rw:/media/49196c8f-cb29-4184-8bd7-98f0aee798cc/disk3/=rw,sum,create=pmfs,udba=reval 0 0
/media/poolshare /media/1fa4c8a2-d4c3-48cf-bb08-ec5669b457dc/poolshare/ none bind 0 0
# <<< [openmediavault]
Ubuntu mount would be like --> none /storage aufs br:/media/disk1=rw:/media/disk2=rw:/media/disk3=rw,sum,create=pmfs 0 0
Am I wrong in thinking that it is somehow listing each drive twice? Once by UUID and then by it's label? eg. disk1?
Kryspy
This line creates the pool:
none /media/poolshare aufs br:/media/1fa4c8a2-d4c3-48cf-bb08-ec5669b457dc/disk1/=rw:/media/148c4c78-ff13-472c-a1f6-71fc64b47085/disk2/=rw:/media/49196c8f-cb29-4184-8bd7-98f0aee798cc/disk3/=rw,sum,create=pmfs,udba=reval 0 0
This line bind the pool to the bind share:
/media/poolshare /media/1fa4c8a2-d4c3-48cf-bb08-ec5669b457dc/poolshare/ none bind 0 0
Once you use the share (with nfs or ftp), you see it bind mounted to other locations. If you can't see anything in /media/poolshare, it either didn't mount or has permission problems. I'm guessing it isn't mounted. If you reboot, it will definitely mount. Otherwise, type: mount -a
What should the write permissions be?
/media/poolshare is mounted for sure. I verified it with the "mount" command.
Any chance that it isn't mounting before the individual drives? This was a problem in Ubuntu and the solution was to add it to /etc/rc.local
Kryspy
Must be permissions because fstab order was change needed in omv. it is working on my systems but they didnt have pre-existing data.
I've got a backup drive of all my data so I think I will start fresh and format my drives.
Kryspy
Just out of interest:
What is the reason for only allowing ext4 for the parity disk?
The problem of XFS with large filesystems on 2.6 and older kernel? So, would it be possible to use XFS in OMV 0.6 (or maybe now with backport kernel)?
Or is there a performance (or other) advantage of ext4 over XFS for parity?
I had try to set up SnapRAID with 3 XFS drives. I had to look in the source code, to find out that not allowing ext4 for parity is a feature. 
Maybe a hint in the plugin page or in this howto would be helpful to others.
kalle
According to the snapraid FAQ, xfs and the 2.6.32 kernel don't work. So, the plugin should be able to use XFS under Wheezy.
Which filesystem is recommended for SnapRAID ?
For the data disks, you can use whatever filesystem you like. For the parity disks, follow these suggestions:
FileSystem Result
ext4 OK
ext3 No. It doesn't support the fallocate() command needed to allocated the parity files. You can anyway use it if you limit the use of the parity disk to contain only the parity file.
XFS OK (but use a recent kernel if you want to use disks of 3TB or bigger. There are problem reported with kernel 2.6.32 and fallocate() with XFS)
JFS No. It doesn't support the fallocate() command needed to allocated the parity files. It also doesn't seem to implement at the best the posix_fadvise() command.
NTFS OK
FAT No. You cannot store files bigger than 4GB. It's also not journaled. Use NTFS instead.I've edited Solo's script a little so it would Scrub the array every 10 runs of the script.
My modifications are marked by "ADD DSpevak".
See the code there: http://pastebin.com/x2aK67TZ
The idea is as follows: the Cronjob runs the script every night, so every 10-th night, if the Sync was OK, it also performs a Scrub operation.
I have 2 questions:
1) is it correct to run scrubs like that or there are some factors I miss?
2) The counter uses a text file to store the number between sessions. Is it OK or there's a more elegant solution?
Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!