Hi, I have a BTRFS filesystem that seems to be suffering from a few problems. I''ll post the first one, which looks most bizarre to me. The filesystem is mounted at /media/Media. It consists of 4 devices in RAID1 (both metadata and data), of sizes 3*2TB and 1*1TB. The result of sudo btrfs fi df /media/Media/ is as follows: Data, RAID1: total=3.06TB, used=8.74TB System, RAID1: total=32.00MB, used=1.81MB System: total=4.00MB, used=0.00 Metadata, RAID1: total=130.29GB, used=9.41GB I''m running Ubuntu Quantal, on the stock 3.5 kernel (with stock btrfs-tools), but the filesystem has been around for a over a year now, since Oneiric or maybe Natty. It has been causing a few problems recently with the odd kernel crash, but this line in bold seems the most odd. It reports Total as smaller than Used. That just seems very wrong to me. I''d also like to rescue my filesystem rather than kill it and restore from backup if possible (since it''s pretty big and restoring would take a while / be error-prone). I''m no kernel developer, but I am able to compile and and willing to help debug issues given instructions. I also have the following mentions of btrfs in the syslog following boot: Nov 12 22:10:43 HomeServer kernel: [ 6.756770] btrfs: bdev /dev/disk/by-uuid/f2178b12-fa9c-4fa0-9efb-11ad188156c0 errs: wr 0, rd 0, flush 0, corrupt 0, gen 0 Nov 12 22:10:43 HomeServer kernel: [ 14.613574] btrfs: bdev /dev/sdd errs: wr 72, rd 0, flush 0, corrupt 0, gen 0 Nov 12 22:10:43 HomeServer kernel: [ 14.613578] btrfs: bdev /dev/sdc errs: wr 542570, rd 559160, flush 0, corrupt 891, gen 0 Nov 12 22:10:43 HomeServer kernel: [ 14.613582] btrfs: bdev /dev/sdb errs: wr 156, rd 1166758, flush 0, corrupt 0, gen 0 Nov 12 22:10:43 HomeServer kernel: [ 14.613584] btrfs: bdev /dev/sda errs: wr 44273, rd 2378244, flush 0, corrupt 429, gen 0 If I attempt to rebalance, the kernel crashes after a while, well before it has had a chance to complete. Any assistance greatly appreciated, and I hope I can also help identify a bug with help. Thanks, Ross -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Nov 12, 2012 at 10:37:58PM +0800, Ross Moore wrote:> Hi, > > I have a BTRFS filesystem that seems to be suffering from a few > problems. I''ll post the first one, which looks most bizarre to me. > > The filesystem is mounted at /media/Media. It consists of 4 devices in > RAID1 (both metadata and data), of sizes 3*2TB and 1*1TB. The result > of sudo btrfs fi df /media/Media/ is as follows: > Data, RAID1: total=3.06TB, used=8.74TB > System, RAID1: total=32.00MB, used=1.81MB > System: total=4.00MB, used=0.00 > Metadata, RAID1: total=130.29GB, used=9.41GB > > I''m running Ubuntu Quantal, on the stock 3.5 kernel (with stock > btrfs-tools), but the filesystem has been around for a over a year > now, since Oneiric or maybe Natty. It has been causing a few problems > recently with the odd kernel crash, but this line in bold seems the > most odd. It reports Total as smaller than Used. That just seems very > wrong to me. I''d also like to rescue my filesystem rather than kill it > and restore from backup if possible (since it''s pretty big and > restoring would take a while / be error-prone). >Can you provide the output of ''btrfs-debug-tree /dev/sdd''? thanks, liubo> I''m no kernel developer, but I am able to compile and and willing to > help debug issues given instructions. > > I also have the following mentions of btrfs in the syslog following boot: > Nov 12 22:10:43 HomeServer kernel: [ 6.756770] btrfs: bdev > /dev/disk/by-uuid/f2178b12-fa9c-4fa0-9efb-11ad188156c0 errs: wr 0, rd > 0, flush 0, corrupt 0, gen 0 > Nov 12 22:10:43 HomeServer kernel: [ 14.613574] btrfs: bdev /dev/sdd > errs: wr 72, rd 0, flush 0, corrupt 0, gen 0 > Nov 12 22:10:43 HomeServer kernel: [ 14.613578] btrfs: bdev /dev/sdc > errs: wr 542570, rd 559160, flush 0, corrupt 891, gen 0 > Nov 12 22:10:43 HomeServer kernel: [ 14.613582] btrfs: bdev /dev/sdb > errs: wr 156, rd 1166758, flush 0, corrupt 0, gen 0 > Nov 12 22:10:43 HomeServer kernel: [ 14.613584] btrfs: bdev /dev/sda > errs: wr 44273, rd 2378244, flush 0, corrupt 429, gen 0 > > If I attempt to rebalance, the kernel crashes after a while, well > before it has had a chance to complete. > > Any assistance greatly appreciated, and I hope I can also help > identify a bug with help. > > Thanks, > Ross > -- > To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html-- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
That''s a big file - is that a dump of all the metadata? It''s around 800MB. Two files provided, one with the errors/warnings, one the actual output. Time-limited link to download, can you let me know when you have downloaded a copy please? http://akiajwevbuo5lh2yiu7q.sharing.s3.amazonaws.com/btrfs-debug-tree?AWSAccessKeyId=AKIAJWEVBUO5LH2YIU7Q&Expires=1352934869&Signature=LAX0noCgddqJY3pmfYh9v1moFqg%3D http://akiajwevbuo5lh2yiu7q.sharing.s3.amazonaws.com/btrfs-debug-tree_errors?AWSAccessKeyId=AKIAJWEVBUO5LH2YIU7Q&Expires=1352934869&Signature=no9GxLnaaS1gc%2BBLJmxSlAAmNz0%3D On 12 November 2012 23:08, Liu Bo <bo.li.liu@oracle.com> wrote:> On Mon, Nov 12, 2012 at 10:37:58PM +0800, Ross Moore wrote: >> Hi, >> >> I have a BTRFS filesystem that seems to be suffering from a few >> problems. I''ll post the first one, which looks most bizarre to me. >> >> The filesystem is mounted at /media/Media. It consists of 4 devices in >> RAID1 (both metadata and data), of sizes 3*2TB and 1*1TB. The result >> of sudo btrfs fi df /media/Media/ is as follows: >> Data, RAID1: total=3.06TB, used=8.74TB >> System, RAID1: total=32.00MB, used=1.81MB >> System: total=4.00MB, used=0.00 >> Metadata, RAID1: total=130.29GB, used=9.41GB >> >> I''m running Ubuntu Quantal, on the stock 3.5 kernel (with stock >> btrfs-tools), but the filesystem has been around for a over a year >> now, since Oneiric or maybe Natty. It has been causing a few problems >> recently with the odd kernel crash, but this line in bold seems the >> most odd. It reports Total as smaller than Used. That just seems very >> wrong to me. I''d also like to rescue my filesystem rather than kill it >> and restore from backup if possible (since it''s pretty big and >> restoring would take a while / be error-prone). >> > > Can you provide the output of ''btrfs-debug-tree /dev/sdd''? > > thanks, > liubo > >> I''m no kernel developer, but I am able to compile and and willing to >> help debug issues given instructions. >> >> I also have the following mentions of btrfs in the syslog following boot: >> Nov 12 22:10:43 HomeServer kernel: [ 6.756770] btrfs: bdev >> /dev/disk/by-uuid/f2178b12-fa9c-4fa0-9efb-11ad188156c0 errs: wr 0, rd >> 0, flush 0, corrupt 0, gen 0 >> Nov 12 22:10:43 HomeServer kernel: [ 14.613574] btrfs: bdev /dev/sdd >> errs: wr 72, rd 0, flush 0, corrupt 0, gen 0 >> Nov 12 22:10:43 HomeServer kernel: [ 14.613578] btrfs: bdev /dev/sdc >> errs: wr 542570, rd 559160, flush 0, corrupt 891, gen 0 >> Nov 12 22:10:43 HomeServer kernel: [ 14.613582] btrfs: bdev /dev/sdb >> errs: wr 156, rd 1166758, flush 0, corrupt 0, gen 0 >> Nov 12 22:10:43 HomeServer kernel: [ 14.613584] btrfs: bdev /dev/sda >> errs: wr 44273, rd 2378244, flush 0, corrupt 429, gen 0 >> >> If I attempt to rebalance, the kernel crashes after a while, well >> before it has had a chance to complete. >> >> Any assistance greatly appreciated, and I hope I can also help >> identify a bug with help. >> >> Thanks, >> Ross >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html-- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Ross Moore <ross <at> rossmoore.net> writes: [...]> >> It reports Total as smaller than Used. That just seems very > >> wrong to me. I''d also like to rescue my filesystem rather than kill it > >> and restore from backup if possible (since it''s pretty big and > >> restoring would take a while / be error-prone). > >>I think this is the same problem I am having. I really do not want to restore from backup because it is a lot of data and will therefore involve a van. Pls post solution if you find it, thx. -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Nov 20, 2012 at 05:54:35PM +0800, Ross Moore wrote:> Hi, > Any update on this before I wipe the disks and start again?Hi Ross, Sorry but I failed to reproduce it locally, although I''ve tried various disk operations like balance, dev add/del etc... So have you also tried the latest btrfs or the last stable one? Does they show the same ''btrfs fi df'' output? thanks, liubo -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Nov 20, 2012 at 09:25:02PM +0800, Ross Moore wrote:> Strange. Does that imply a problem with disks or disk cables physically? > Our just the data rather than the metadata? >It seems to be a kernel bug since another user also reported it sometime ago.> I''ve compiled the latest btrfs-tools from git trunk add of a week ago when > I reported the bug, same results. I haven''t tried an updated kernel, just > the 3.5. Should I upgrade to 3.7 rc? Will that hold any hope of change?To be honest, I''m not sure if it''ll help. Hope that others can come up with where goes wrong. thanks, liubo> On 20 Nov 2012 21:20, "Liu Bo" <bo.li.liu@oracle.com> wrote: > > > On Tue, Nov 20, 2012 at 05:54:35PM +0800, Ross Moore wrote: > > > Hi, > > > Any update on this before I wipe the disks and start again? > > > > Hi Ross, > > > > Sorry but I failed to reproduce it locally, although I''ve tried various > > disk operations like balance, dev add/del etc... > > > > So have you also tried the latest btrfs or the last stable one? > > Does they show the same ''btrfs fi df'' output? > > > > thanks, > > liubo > >-- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html