AUFS Issue.

Meridock · 29. Januar 2016

I had a working AUFS drivepool that was getting full, and added a new drive today. I:

formatted the new one,
added to union filesystem just like I did the last time I added a drive to the same system, it showed 10TB available on the pool
started copying files to the pool

and the pool filled up on the original 3 drives and nothing went to the new one. So I read some of the posts on this forum and it hinted at a permissions issue. I dug thru the various drives and saw that the new one did not have the drive pool folder created that the other drives had. I manually added the directory and set perms. (even tried to set via the GUI the share and such perms.) Nothing seems to set that new drive to be in use.

I have the brilliant idea to tear down the entire pool and reformat (this is a backup drive to my RAID) and rebuilt. Now only a SINGLE dive is being used.

Thoughts?

sno0k · 29. Januar 2016

Had the same issues a year ago. You can read it up in the unionfilesystem plugin thread.
Meanwhile I switched to mergerfs and get similar speeds in read (above 100Mb/s) and write (average of 75Mb/s or higher) compared to an aufs pool. To reach this speeds you have to add something over cli (see this post).
I'm satisfied with the result.

Meridock · 29. Januar 2016

i pulled the fstab to see if the some of the early fixes were there but it seems like the =rw is solid.

Code

# /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>
# / was on /dev/sda1 during installation
UUID=3a02401b-f8e6-4f43-9775-93ebbcdcdacf /               ext4    errors=remount-ro 0       1
# swap was on /dev/sda5 during installation
UUID=3413ecb8-5711-4b56-bbdb-17b4ef2b6e1c none            swap    sw              0       0
/dev/sr0        /media/cdrom0   udf,iso9660 user,noauto     0       0
tmpfs           /tmp            tmpfs   defaults        0       0
# >>> [openmediavault]
UUID=7e5ffd2d-926e-4198-abcc-23253775824a /media/7e5ffd2d-926e-4198-abcc-23253775824a ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0 0 2
UUID=9c5a3020-f6d3-4d6d-865e-0c06c3566b19 /media/9c5a3020-f6d3-4d6d-865e-0c06c3566b19 ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl 0 2
UUID=e36b057b-73e6-4fbb-9a8a-595bea46294c /media/e36b057b-73e6-4fbb-9a8a-595bea46294c ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl 0 2
UUID=33602b2f-ca24-4d1e-a8dc-3ade1dace599 /media/33602b2f-ca24-4d1e-a8dc-3ade1dace599 ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl 0 2
UUID=8a48e98a-9dc3-4d93-b046-e8d28fbc7d5f /media/8a48e98a-9dc3-4d93-b046-e8d28fbc7d5f ext4 defaults,nofail,acl,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl 0 2
none /media/43dd1777-c05a-4a83-8715-1ab764b3fbb4 aufs br:/media/33602b2f-ca24-4d1e-a8dc-3ade1dace599=rw:/media/8a48e98a-9dc3-4d93-b046-e8d28fbc7d5f=rw:/media/e36b057b-73e6-4fbb-9a8a-595bea46294c=rw:/media/9c5a3020-f6d3-4d6d-865e-0c06c3566b19=rw,sum,create=tdp 0 0
/media/7e5ffd2d-926e-4198-abcc-23253775824a/media/ /export/media none bind 0 0
/media/7e5ffd2d-926e-4198-abcc-23253775824a/backups/ /export/backups none bind 0 0
# <<< [openmediavault]

Alles anzeigen

My aufs is drive : 43dd1777-c05a-4a83-8715-1ab764b3fbb4
none /media/43dd1777-c05a-4a83-8715-1ab764b3fbb4 aufs br:/media/33602b2f-ca24-4d1e-a8dc-3ade1dace599=rw:/media/8a48e98a-9dc3-4d93-b046-e8d28fbc7d5f=rw:/media/e36b057b-73e6-4fbb-9a8a-595bea46294c=rw:/media/9c5a3020-f6d3-4d6d-865e-0c06c3566b19=rw,sum,create=tdp 0 0

but

ryecoaaron · 29. Januar 2016

Never reformat the drives because aufs isn't working. That is not the problem. aufs and the other union filesystems work on top of the filesystem.

Getting rid of the .wh files would have probably fixed the problem but I would just switch to mergerfs. I did.

Meridock · 29. Januar 2016

ha. I was not mad to reformat. it probably cleared 200 gigs of debris from the folders.

That said, I spent the last 20 min or so selecting each drive to be the only drive in the pool, and each one mounted and accepted a file or two. I just put all 4in one drive and am testing.

and we have no joy in this. I am seeing a single drive being loaded with files.

So...

Meridock · 29. Januar 2016

Ok... based on my archaic form of OMV I moved over to mergerfs... seems like the same problem one drive loaded up first.

Currently copying @ 10+MB/s 2 file streams but even with the mergerfs only one of my drives is filling,

Meridock · 29. Januar 2016

okay. reset the policy. and it seems to be filling as I expect.

ryecoaaron · 29. Januar 2016

You should reboot whenever changing union filesystem type.

Meridock · 29. Januar 2016

I reboot at most changes since I don't CLI so i rely on most of OVM implementations that being stsaid... 300GB into my 7TB copy and it is looking solid.

Jetzt mitmachen!