Re: [gentoo-user] btrfs fails to balance
On 19/01/15 22:08, Marc Stuermer wrote: > Am 19.01.2015 um 12:02 schrieb Bill Kenworthy: > >> Brilliant, you have hit on the answer! - The ancient 300GB system disk >> was sda at one point and moved to sdb - possibly at the time I changed >> to using UUID's. Ive just resized all the disks and its now moved past >> 300G for the first time as well as the other two falling in step with >> the data moving. > > Actually this is not my doing, I am not touching Btrfs at the moment > with a ten foot pole yet because yet. > > Russel Coker wrote about the problem you had, just take a look at it > here and also mentioned the solution: > > http://etbe.coker.com.au/2014/12/05/btrfs-status-dec-2014/ > > He had the same problem with a bigger drive. His Btrfs status updates > are something I do follow indeed. > Yes, I read those but failed to "connect the dots" :) Balance almost done: rattus backups # btrfs fi usage -T /mnt/btrfs-root/ Overall: Device size: 5.46TiB Device allocated: 3.81TiB Device unallocated:1.65TiB Used: 3.80TiB Free (estimated):843.92GiB (min: 843.92GiB) Data ratio: 2.00 Metadata ratio: 2.00 Global reserve: 512.00MiB (used: 97.84MiB) DataMetadata System singleRAID1 RAID1RAID1 Unallocated /dev/sda - 1.26TiB 6.00GiB 32.00MiB 562.99GiB /dev/sdc - 1.26TiB 8.00GiB - 562.02GiB /dev/sdd - 1.27TiB 4.00GiB 32.00MiB 561.99GiB = === = === Total512.00MiB 1.90TiB 9.00GiB 32.00MiB 1.65TiB Used 97.84MiB 1.90TiB 6.42GiB 304.00KiB rattus backups #
Re: [gentoo-user] btrfs fails to balance
Am 19.01.2015 um 12:02 schrieb Bill Kenworthy: > Brilliant, you have hit on the answer! - The ancient 300GB system disk > was sda at one point and moved to sdb - possibly at the time I changed > to using UUID's. Ive just resized all the disks and its now moved past > 300G for the first time as well as the other two falling in step with > the data moving. Actually this is not my doing, I am not touching Btrfs at the moment with a ten foot pole yet because yet. Russel Coker wrote about the problem you had, just take a look at it here and also mentioned the solution: http://etbe.coker.com.au/2014/12/05/btrfs-status-dec-2014/ He had the same problem with a bigger drive. His Btrfs status updates are something I do follow indeed.
Re: [gentoo-user] btrfs fails to balance
On 19/01/15 18:45, Marc Stürmer wrote: > Am 19.01.2015 um 09:32 schrieb Bill Kenworthy: > >> Can someone suggest what is causing a balance on this raid 1, 3 disk >> volume to successfully complete but leave the data unevenly distributed? >> Content is mostly VM images. >> >> sdc and sdd are 2TB WD greens, and sda is a 2TB WD red. > > Question: was /dev/sda a smaller HDD before the 2 TB WD red? > > If your sda was around 250 GB before you changed it with 2 TB, did you > just issue a "btrfs balance" after that? If so, Btrfs just configured > itself for 2*2 TB + 1*250 GB, that's why. > > The proper Btrfs way if replacing a smaller hdd for a bigger one in Raid > 1 is to issue "btrfs filesystem resize" to make it use all of the > available space. > > This would be one possible explanation for the behaviour of your array. > Brilliant, you have hit on the answer! - The ancient 300GB system disk was sda at one point and moved to sdb - possibly at the time I changed to using UUID's. Ive just resized all the disks and its now moved past 300G for the first time as well as the other two falling in step with the data moving. I moved to UUID's as the machine has a number of sata ports and a PCI-e sata adaptor and the sd* drive numbering kept moving around when I added the WD red. BillK
Re: [gentoo-user] btrfs fails to balance
On 19/01/15 17:46, Marc Stürmer wrote: > > Zitat von Bill Kenworthy : > >> Can someone suggest what is causing a balance on this raid 1, 3 disk >> volume to successfully complete but leave the data unevenly distributed? >> Content is mostly VM images. > > On which kernel version are you? > 3.17.7 ... will move to 3.18 tomorrow as there are some brtfs fixes, but nothing that would relate to this problem which I think has been present for quite awhile. BillK
Re: [gentoo-user] btrfs fails to balance
Am 19.01.2015 um 09:32 schrieb Bill Kenworthy: Can someone suggest what is causing a balance on this raid 1, 3 disk volume to successfully complete but leave the data unevenly distributed? Content is mostly VM images. sdc and sdd are 2TB WD greens, and sda is a 2TB WD red. Question: was /dev/sda a smaller HDD before the 2 TB WD red? If your sda was around 250 GB before you changed it with 2 TB, did you just issue a "btrfs balance" after that? If so, Btrfs just configured itself for 2*2 TB + 1*250 GB, that's why. The proper Btrfs way if replacing a smaller hdd for a bigger one in Raid 1 is to issue "btrfs filesystem resize" to make it use all of the available space. This would be one possible explanation for the behaviour of your array.
Re: [gentoo-user] btrfs fails to balance
Zitat von Bill Kenworthy : Can someone suggest what is causing a balance on this raid 1, 3 disk volume to successfully complete but leave the data unevenly distributed? Content is mostly VM images. On which kernel version are you?
Re: [gentoo-user] btrfs fails to balance
Am Mon, 19 Jan 2015 16:32:53 +0800 schrieb Bill Kenworthy : > Can someone suggest what is causing a balance on this raid 1, 3 disk > volume to successfully complete but leave the data unevenly distributed? > Content is mostly VM images. > > sdc and sdd are 2TB WD greens, and sda is a 2TB WD red. > > > rattus backups # btrfs fi sh > Label: none uuid: 04d8ff4f-fe19-4530-ab45-d82fcd647515 > Total devices 1 FS bytes used 8.25GiB > devid1 size 271.36GiB used 23.04GiB path /dev/sdb3 > > Label: none uuid: f5a284b6-442f-4b3d-aa1a-8d6296f517b1 > Total devices 3 FS bytes used 1.90TiB > devid1 size 1.82TiB used 1.77TiB path /dev/sdc > devid2 size 1.82TiB used 1.77TiB path /dev/sdd > devid4 size 1.82TiB used 270.03GiB path /dev/sda > > Btrfs v3.18 > rattus backups # btrfs fi df /mnt/btrfs-root/ > Data, RAID1: total=1.90TiB, used=1.89TiB > System, RAID1: total=32.00MiB, used=464.00KiB > Metadata, RAID1: total=8.00GiB, used=6.53GiB > GlobalReserve, single: total=512.00MiB, used=0.00B > rattus backups # btrfs fi usage /mnt/btrfs-root/ > Overall: > Device size: 5.46TiB > Device allocated: 3.81TiB > Device unallocated:1.65TiB > Used: 3.80TiB > Free (estimated):845.00GiB (min: 845.00GiB) > Data ratio: 2.00 > Metadata ratio: 2.00 > Global reserve: 512.00MiB (used: 154.42MiB) > > Data,RAID1: Size:1.90TiB, Used:1.89TiB >/dev/sda 265.00GiB >/dev/sdc1.77TiB >/dev/sdd1.77TiB OK, that is odd. While my experience with BTRFS RAID1 was that the FS is never perfectly balanced, this is just... weird. If I were you, I would ask on the BTRFS ML. > Metadata,RAID1: Size:9.00GiB, Used:6.96GiB >/dev/sda6.00GiB >/dev/sdc7.00GiB >/dev/sdd5.00GiB This I would see as perfectly possible with BTRFS RAID1. > System,RAID1: Size:32.00MiB, Used:464.00KiB >/dev/sda 32.00MiB >/dev/sdd 32.00MiB > > Unallocated: >/dev/sda1.55TiB >/dev/sdc 47.02GiB >/dev/sdd 47.99GiB > rattus backups # HTH -- Marc Joliet -- "People who think they know everything really annoy those of us who know we don't" - Bjarne Stroustrup pgp7pIeqKAaqm.pgp Description: Digitale Signatur von OpenPGP
