BTRFS constantly reports "No space left on device" even with a huge unallocated space

Hi guys,

I'm facing a daily problem with BTRFS. Almost everyday, I get the message "No space left on device". Sometimes I can recover by balancing the system but sometimes even balancing does not work due to the lack of space. In this case, only a hard reset works if I can't delete some files. The problem is that I have a huge unallocated space as you can see here:

# btrfs fi usage /
Overall:
    Device size:		   1.26TiB
    Device allocated:		 119.07GiB
    Device unallocated:		   1.14TiB
    Device missing:		     0.00B
    Used:			 115.08GiB
    Free (estimated):		   1.14TiB	(min: 586.21GiB)
    Data ratio:			      1.00
    Metadata ratio:		      2.00
    Global reserve:		 512.00MiB	(used: 0.00B)

Data,single: Size:113.01GiB, Used:111.19GiB
   /dev/sda6  113.01GiB

Metadata,DUP: Size:3.00GiB, Used:1.94GiB
   /dev/sda6    6.00GiB

System,DUP: Size:32.00MiB, Used:16.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.14TiB

It is not easy to trigger the problem. But I do find some correlation between two things:

1) When I started to create jails to build openSUSE packages locally, then the problem happens more often. In these jails, some directories like /dev/, /dev/pts, /proc, are mounted inside the jail.

2) When I open my KVM, I also see this problem more often. Notice, however, that the KVM disk is stored in another EXT4 partition.

I would be glad if anyone can help me to fix it. In the following, I'm providing more information about my system:

# uname -a
Linux ronanarraes-osd 4.7.0-1-default #1 SMP PREEMPT Mon Jul 25
08:42:47 UTC 2016 (89a2ada) x86_64 x86_64 x86_64 GNU/Linux

# btrfs --version
btrfs-progs v4.6.1+20160714

# btrfs fi show
Label: none  uuid: 80381f7f-8cef-4bd8-bdbc-3487253ee566
	Total devices 1 FS bytes used 113.13GiB
	devid    1 size 1.26TiB used 119.07GiB path /dev/sda6

# btrfs fi df /
Data, single: total=113.01GiB, used=111.19GiB
System, DUP: total=32.00MiB, used=16.00KiB
Metadata, DUP: total=3.00GiB, used=1.94GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

Regards,
Ronan Arraes

Tons of unallocated space. What kernel messages do you get for the enospc? It sounds like this will be one of the mystery -28 error file systems. So far as I recall the only work around is recreating the file system. There are two additional things you can try: mount with enospc_debug mount option and see if you can gather more information about the problem. Or try a 4.8rc1 kernel which as a large number of enospc changes.

Chris Murphy

Em Sex, 2016-08-12 às 12:02 -0600, Chris Murphy escreveu:

Unfortunately no log was written due to the lack of space :) Next time it happens, I will take a screenshot of the message. Do you think that if I reinstall my openSUSE it will be fixed?

Regards,
Ronan Arraes

a. journalctl -f in a Terminal window or tab should still record everything. So long as the OS isn't totally face planting when the enospc happens, you may still be able to copy paste it into a file that you can save on another file system volume. It might have some noisy messages from systemd-journald being unable to flush to disk but the enospc itself should all be in the window even though they don't get committed to disk.

b. Modify /etc/systemd/journald.conf so that Storage=volatile and now the journal is only in memory, and you can flush it to another file system yourself with something like 'journalctl -b -o short-monotonic

c. create a ~1GiB separate file system and mount it at /var/log/

d. Run journalctl -f from a 2nd computer.

Maybe. enospc_debug tends to spit out more than the usual amount of stuff that'll fit on a single screen.

Probably but the nature of this probem isn't well understood as far as I know. It's not that common or it'd be easy for a dev to reproduce and then figure out what's going on.

Chris Murphy

Hopefully it's obvious I mean run journalctl -f on the affected computer remotely via ssh.

Since this file system has relatively small metadata size, just under 2GiB, it might be useful to take a btrfs-image of it and put it up somewhere like a google drive, or wherever it can remain for a while. Options -t 4 -c9 -s are fairly standard and sanitize file names. Data itself is not included in the image. From this I think a dev might be able to figure out what's unique about this file system that results in the bogus enospc. If you do this, I recommend filing a bugzilla.kernel.org bug and include URL to the image and URL to this thread, and then the bugzilla URL in a post on this thread that way everything is cross referenced.

Chris Murphy

Looks similar to this:
https://lkml.org/lkml/2016/3/28/230

Can you describe the workload happening at the time?

Chris Murphy

Em Seg, 2016-08-15 às 17:24 -0600, Chris Murphy escreveu:

I was copying my /home using rsyinc when this happened. Unfortunately I needed to format this machine because it is a production system. If I see any problems related to that, I will report to this mailing list.

Regards,
Ronan Arraes

New information guys! I formatted using the latest Tumbleweed snapshot (btrfs-progs v4.7+20160729) and I still have the same problem.

I notice two things. First, when I see the "No space left on device", it is fixed when the Metadata space increases **a lot**. For example, when the error first occurred, I had:

Metadata, DUP: total=2.00GiB, used=811.52MiB

After waiting a while (could not run balance), it was automatically fixed and then I have:

Metadata, DUP: total=9.50GiB, used=811.52MiB

During the error, when I ran the balance command, I see these messages in dmesg:

Ago 22 16:00:03 ronanarraes-osd kernel: BTRFS info (device sda6):
relocating block group 9323937792 flags 34
Ago 22 16:00:04 ronanarraes-osd kernel: BTRFS info (device sda6): found 1 extents
Ago 22 16:00:04 ronanarraes-osd kernel: BTRFS info (device sda6): 1 enospc errors during balance
Ago 22 16:00:24 ronanarraes-osd kernel: BTRFS info (device sda6):
relocating block group 36201037824 flags 34
Ago 22 16:00:24 ronanarraes-osd kernel: BTRFS info (device sda6): 2 enospc errors during balance
Ago 22 16:00:45 ronanarraes-osd kernel: BTRFS info (device sda6):
relocating block group 36234592256 flags 34
Ago 22 16:00:46 ronanarraes-osd kernel: BTRFS info (device sda6): found 1 extents
Ago 22 16:00:46 ronanarraes-osd kernel: BTRFS info (device sda6): 4 enospc errors during balance
Ago 22 16:01:20 ronanarraes-osd kernel: BTRFS info (device sda6):
relocating block group 38415630336 flags 34
Ago 22 16:01:21 ronanarraes-osd kernel: BTRFS info (device sda6): found 1 extents
Ago 22 16:01:21 ronanarraes-osd kernel: BTRFS info (device sda6): 8 enospc errors during balance

Does it add anything relevant to the problem?

Regards,
Ronan Arraes

The same thing just happened again! And now it was also fixed automatically, but now I have:

Metadata,DUP: Size:33.50GiB, Used:812.78MiB

This is really weird. I'm running 4.7.0 (Fedora) and I'm not experiencing problems, let alone this. What is this kernel's provenance? Is it a plain mainline 4.7.0 that you built? I'm not really sure what to recommend except maybe going back to 4.5.7 or 4.6.7 as it's a production machine. Heck even 4.4.19 is OK for me in this regard.

Chris Murphy

Em Seg, 2016-08-22 às 14:49 -0600, Chris Murphy escreveu:

Well, I'm using the default openSUSE kernel here. And I have been seen this errors for sometimes. When I reported it, I was using v4.6.1. Hence, I think the version of btrfs-progs is not the problem.

The openSUSE Tumbleweed kernel is effectively vanilla for btrfs. The 4.6-based kernel had two btrfs patches, the crc32c implementation publishing patch that landed in 4.7, and the super_operation callback that lets us publish the per-root anon dev via stat(). The problem you're encountering isn't due to our patches.

-Jeff

Jeff Mahoney
SUSE Labs

Looks also similar to the subject of the lenghty thread titled "6TB partition, Data only 2TB - aka When you haven't hit the "usual" problem" that started with:

Regards,

Lutz Vieweg

I'm thinking it might be a conflict with the OP doing builds, implies heavy writes, maybe especially heavy metadata based on the large increase in metadata allocation; along with the default on opensuse using snapper to make read only snapshots. That it can't be triggered on demand makes sense, if the build doesn't overlap with snapper making a snapshot of /home.

http://www.spinics.net/lists/linux-btrfs/msg52670.html

Anyway it's a known problem, I don't think it's fixed still. There's a lot of enospc work in 4.8 so eventually it'll make sense to give it a shot with that kernel.

Chris Murphy

hi,

assuming that I'm willing to try that, will a successful rebalance with 4.8 fix a filesystem, or is the recommended way still "backup, format, restore, lose all snapshots"?

Greetings
Marc

-----------------------------------------------------------------------------
Marc Haber         | "I don't trust Computers. They | Mailadresse im Header
Leimen, Germany    |  lose things."    Winona Ryder | Fon: *49 6224 1600402
Nordisch by Nature |  How to make an American Quilt | Fax: *49 6224 1600421

hello,

When strange ENOSPC errors occur, I think "btrfs fi usage" or "btrfs di df" do not help too much. Their output do not reflect btrfs kernel current status :)

Would you please provide attribute files' values in /sys/fs/btrfs/$UUID/allocation/data and /sys/fs/btrfs/$UUID/allocation/metadata when ENOSPC error occurs.

Regards,
Xiaoguang Wang

Hi!

Em Seg, 2016-08-29 às 20:12 +0800, Wang Xiaoguang escreveu:

Sure! As soon as I see the error again, I will send this results. Now, I see that if I move my jail directory to a ext4 partition, then I do not see the problem anymore, but I need more test to validade this assumption.

Best regards,
Ronan Arraes

Hi guys,

I just have the problem again. Now, it happens during the lunch time when the machine was idle. Only the system processes were running. It was not the first time that I saw this problem just after lunch when the machine stayed idle for a long period (+- 1h).

Here is the information requested:

/sys/fs/btrfs/$UUID/allocation/data

./bytes_may_use
0
./bytes_pinned
0
./bytes_reserved
0
./bytes_used
36128374784
./disk_total
37589352448
./disk_used
36128374784
./flags
1
./total_bytes
37589352448
./total_bytes_pinned
20339560448
./single/total_bytes
37589352448
./single/used_bytes
36128374784

/sys/fs/btrfs/$UUID/allocation/metadata

./bytes_may_use
84974452736
./bytes_pinned
0
./bytes_reserved
0
./bytes_used
977354752
./disk_total
4294967296
./disk_used
1954709504
./flags
4
./total_bytes
2147483648
./total_bytes_pinned
-57851904
./dup/total_bytes
2147483648
./dup/used_bytes
977354752

# btrfs fi usage /
Overall:
    Device size:		   1.26TiB
    Device allocated:		  39.07GiB
    Device unallocated:		   1.22TiB
    Device missing:		     0.00B
    Used:			  35.29GiB
    Free (estimated):		   1.22TiB	(min: 625.93GiB)
    Data ratio:			      1.00
    Metadata ratio:		      2.00
    Global reserve:		 320.00MiB	(used: 0.00B)

Data,single: Size:35.01GiB, Used:33.47GiB
   /dev/sda6   35.01GiB

Metadata,DUP: Size:2.00GiB, Used:932.00MiB
   /dev/sda6    4.00GiB

System,DUP: Size:32.00MiB, Used:16.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.22TiB

# btrfs fi df /
Data, single: total=35.01GiB, used=33.47GiB
System, DUP: total=32.00MiB, used=16.00KiB
Metadata, DUP: total=2.00GiB, used=932.09MiB
GlobalReserve, single: total=320.00MiB, used=0.0

I also saw the following information in journalctl:

Ago 29 10:25:33 ronanarraes-osd kernel: ------------[ cut here ]-------
-----
Ago 29 10:25:33 ronanarraes-osd kernel: WARNING: CPU: 4 PID: 30424 at ../fs/btrfs/extent-tree.c:4303 btrfs_free_reserved_data_space_noquota+0xfe/0x110 [btrfs] Ago 29 10:25:33 ronanarraes-osd kernel: Modules linked in: fuse nf_log_ipv6 xt_pkttype nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_
Ago 29 10:25:33 ronanarraes-osd kernel:  mei_wdt sysimgblt iTCO_vendor_support i2c_i801 tpm_infineon tpm_tis tpm ioatdma crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw sparse_keymap
Ago 29 10:25:33 ronanarraes-osd kernel: CPU: 4 PID: 30424 Comm:
kworker/u65:1 Tainted: P           O    4.7.1-1-default #1 Ago 29 10:25:33 ronanarraes-osd kernel: Hardware name: Hewlett-Packard HP Z820 Workstation/158B, BIOS J63 v03.65 12/19/2013
Ago 29 10:25:33 ronanarraes-osd kernel: Workqueue: writeback wb_workfn (flush-btrfs-1)
Ago 29 10:25:33 ronanarraes-osd kernel:  0000000000000000 ffffffff81393104 0000000000000000 0000000000000000
Ago 29 10:25:33 ronanarraes-osd kernel:  ffffffff8107ca1e ffff88100027c800 0000000000001000 ffff88082ff06400
Ago 29 10:25:33 ronanarraes-osd kernel:  ffff88100c7af784 0000000000001000 ffff8805bd60f6cc ffffffffa025098e
Ago 29 10:25:33 ronanarraes-osd kernel: Call Trace:
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8102fe41>] show_stack+0x21/0x40
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff81393104>] dump_stack+0x5c/0x78
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa025098e>] btrfs_free_reserved_data_space_noquota+0xfe/0x110 [btrfs] Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa026d036>] btrfs_clear_bit_hook+0x296/0x380 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028a755>] clear_state_bit+0x55/0x1d0 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028aa0d>] __clear_extent_bit+0x13d/0x3f0 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028b8d2>] extent_clear_unlock_delalloc+0x62/0x280 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa0273722>] run_delalloc_nocow+0x962/0xba0 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa0273cbf>] run_delalloc_range+0x35f/0x3b0 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028c090>] writepage_delalloc.isra.40+0x100/0x170 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028e9d3>] __extent_writepage+0xc3/0x340 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028ee8b>] extent_write_cache_pages.isra.36.constprop.53+0x23b/0x350 [btrfs] Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffffa028f4fe>] extent_writepages+0x4e/0x60 [btrfs]
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8123c64d>] __writeback_single_inode+0x3d/0x3b0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8123ce8a>] writeback_sb_inodes+0x20a/0x440
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8123d147>] __writeback_inodes_wb+0x87/0xb0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8123d49d>] wb_writeback+0x28d/0x330
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8123dbe2>] wb_workfn+0x222/0x3f0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff810950ed>] process_one_work+0x1ed/0x4e0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff81095427>] worker_thread+0x47/0x4c0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8109affd>] kthread+0xbd/0xe0
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff816bb71f>] ret_from_fork+0x1f/0x40
Ago 29 10:25:33 ronanarraes-osd kernel: DWARF2 unwinder stuck at ret_from_fork+0x1f/0x40
Ago 29 10:25:33 ronanarraes-osd kernel:
Ago 29 10:25:33 ronanarraes-osd kernel: Leftover inexact backtrace:
Ago 29 10:25:33 ronanarraes-osd kernel:  [<ffffffff8109af40>] ? kthread_worker_fn+0x170/0x170

Ago 29 10:34:51 ronanarraes-osd kernel: ------------[ cut here ]-------
-----
Ago 29 10:34:51 ronanarraes-osd kernel: WARNING: CPU: 6 PID: 27335 at ../fs/btrfs/inode.c:9306 btrfs_destroy_inode+0x23f/0x2b0 [btrfs] Ago 29 10:34:51 ronanarraes-osd kernel: Modules linked in: fuse nf_log_ipv6 xt_pkttype nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_
Ago 29 10:34:51 ronanarraes-osd kernel:  mei_wdt sysimgblt iTCO_vendor_support i2c_i801 tpm_infineon tpm_tis tpm ioatdma crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw sparse_keymap
Ago 29 10:34:51 ronanarraes-osd kernel: CPU: 6 PID: 27335 Comm: Cache2 I/O Tainted: P        W  O    4.7.1-1-default #1
Ago 29 10:34:51 ronanarraes-osd kernel: Hardware name: Hewlett-Packard HP Z820 Workstation/158B, BIOS J63 v03.65 12/19/2013
Ago 29 10:34:51 ronanarraes-osd kernel:  0000000000000000 ffffffff81393104 0000000000000000 0000000000000000
Ago 29 10:34:51 ronanarraes-osd kernel:  ffffffff8107ca1e 0000000000000000 ffff88071b592a80 ffff881000221800
Ago 29 10:34:51 ronanarraes-osd kernel:  0000000000000000 ffff88071b592a80 00000000ffffff9c ffffffffa027dabf
Ago 29 10:34:51 ronanarraes-osd kernel: Call Trace:
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff8102fe41>] show_stack+0x21/0x40
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff81393104>] dump_stack+0x5c/0x78
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffffa027dabf>] btrfs_destroy_inode+0x23f/0x2b0 [btrfs]
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff8121f6d1>] do_unlinkat+0x131/0x310
Ago 29 10:34:51 ronanarraes-osd kernel:  [<ffffffff816bb4f6>] entry_SYSCALL_64_fastpath+0x1e/0xa8
Ago 29 10:34:51 ronanarraes-osd kernel: DWARF2 unwinder stuck at entry_SYSCALL_64_fastpath+0x1e/0xa8
Ago 29 10:34:51 ronanarraes-osd kernel:
Ago 29 10:34:51 ronanarraes-osd kernel: Leftover inexact backtrace:
Ago 29 10:34:51 ronanarraes-osd kernel: ---[ end trace 5774bd3049f78a61 ]---

Ago 29 11:21:19 ronanarraes-osd kernel: ------------[ cut here ]-------
-----
Ago 29 11:21:19 ronanarraes-osd kernel: WARNING: CPU: 18 PID: 16759 at ../fs/btrfs/extent-tree.c:4303 btrfs_free_reserved_data_space_noquota+0xfe/0x110 [btrfs] Ago 29 11:21:19 ronanarraes-osd kernel: Modules linked in: fuse nf_log_ipv6 xt_pkttype nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_
Ago 29 11:21:19 ronanarraes-osd kernel:  mei_wdt sysimgblt iTCO_vendor_support i2c_i801 tpm_infineon tpm_tis tpm ioatdma crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw sparse_keymap
Ago 29 11:21:19 ronanarraes-osd kernel: CPU: 18 PID: 16759 Comm:
kworker/u65:2 Tainted: P        W  O    4.7.1-1-default #1 Ago 29 11:21:19 ronanarraes-osd kernel: Hardware name: Hewlett-Packard HP Z820 Workstation/158B, BIOS J63 v03.65 12/19/2013
Ago 29 11:21:19 ronanarraes-osd kernel: Workqueue: writeback wb_workfn (flush-btrfs-1)
Ago 29 11:21:19 ronanarraes-osd kernel:  0000000000000000 ffffffff81393104 0000000000000000 0000000000000000
Ago 29 11:21:19 ronanarraes-osd kernel:  ffffffff8107ca1e ffff881000221800 0000000000001000 ffff88082ff06400
Ago 29 11:21:19 ronanarraes-osd kernel:  ffff8807b11b6784 0000000000001000 ffff8806acb1f73c ffffffffa025098e
Ago 29 11:21:19 ronanarraes-osd kernel: Call Trace:
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8102fe41>] show_stack+0x21/0x40
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff81393104>] dump_stack+0x5c/0x78
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa025098e>] btrfs_free_reserved_data_space_noquota+0xfe/0x110 [btrfs] Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa026d036>] btrfs_clear_bit_hook+0x296/0x380 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028a755>] clear_state_bit+0x55/0x1d0 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028aa0d>] __clear_extent_bit+0x13d/0x3f0 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028b8d2>] extent_clear_unlock_delalloc+0x62/0x280 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa0272c19>] cow_file_range+0x299/0x440 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa0273cf2>] run_delalloc_range+0x392/0x3b0 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028c090>] writepage_delalloc.isra.40+0x100/0x170 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028e9d3>] __extent_writepage+0xc3/0x340 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028ee8b>] extent_write_cache_pages.isra.36.constprop.53+0x23b/0x350 [btrfs] Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffffa028f4fe>] extent_writepages+0x4e/0x60 [btrfs]
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8123c64d>] __writeback_single_inode+0x3d/0x3b0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8123ce8a>] writeback_sb_inodes+0x20a/0x440
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8123d147>] __writeback_inodes_wb+0x87/0xb0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8123d49d>] wb_writeback+0x28d/0x330
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8123dbe2>] wb_workfn+0x222/0x3f0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff810950ed>] process_one_work+0x1ed/0x4e0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff81095427>] worker_thread+0x47/0x4c0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8109affd>] kthread+0xbd/0xe0
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff816bb71f>] ret_from_fork+0x1f/0x40
Ago 29 11:21:19 ronanarraes-osd kernel: DWARF2 unwinder stuck at ret_from_fork+0x1f/0x40
Ago 29 11:21:19 ronanarraes-osd kernel:
Ago 29 11:21:19 ronanarraes-osd kernel: Leftover inexact backtrace:
Ago 29 11:21:19 ronanarraes-osd kernel:  [<ffffffff8109af40>] ? kthread_worker_fn+0x170/0x170
Ago 29 11:21:19 ronanarraes-osd kernel: ---[ end trace 5774bd3049f78a62 ]---

Ago 29 12:06:07 ronanarraes-osd kernel: ------------[ cut here ]-------
-----
Ago 29 12:06:07 ronanarraes-osd kernel: WARNING: CPU: 3 PID: 27335 at ../fs/btrfs/inode.c:9306 btrfs_destroy_inode+0x23f/0x2b0 [btrfs] Ago 29 12:06:07 ronanarraes-osd kernel: Modules linked in: fuse nf_log_ipv6 xt_pkttype nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_
Ago 29 12:06:07 ronanarraes-osd kernel:  mei_wdt sysimgblt iTCO_vendor_support i2c_i801 tpm_infineon tpm_tis tpm ioatdma crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw sparse_keymap
Ago 29 12:06:07 ronanarraes-osd kernel: CPU: 3 PID: 27335 Comm: Cache2 I/O Tainted: P        W  O    4.7.1-1-default #1
Ago 29 12:06:07 ronanarraes-osd kernel: Hardware name: Hewlett-Packard HP Z820 Workstation/158B, BIOS J63 v03.65 12/19/2013
Ago 29 12:06:07 ronanarraes-osd kernel:  0000000000000000 ffffffff81393104 0000000000000000 0000000000000000
Ago 29 12:06:07 ronanarraes-osd kernel:  ffffffff8107ca1e 0000000000000000 ffff88071b5eeb00 ffff881000221800
Ago 29 12:06:07 ronanarraes-osd kernel:  0000000000000000 ffff88071b5eeb00 00000000ffffff9c ffffffffa027dabf
Ago 29 12:06:07 ronanarraes-osd kernel: Call Trace:
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff8102fe41>] show_stack+0x21/0x40
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff81393104>] dump_stack+0x5c/0x78
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffffa027dabf>] btrfs_destroy_inode+0x23f/0x2b0 [btrfs]
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff8121f6d1>] do_unlinkat+0x131/0x310
Ago 29 12:06:07 ronanarraes-osd kernel:  [<ffffffff816bb4f6>] entry_SYSCALL_64_fastpath+0x1e/0xa8
Ago 29 12:06:07 ronanarraes-osd kernel: DWARF2 unwinder stuck at entry_SYSCALL_64_fastpath+0x1e/0xa8
Ago 29 12:06:07 ronanarraes-osd kernel:
Ago 29 12:06:07 ronanarraes-osd kernel: Leftover inexact backtrace:
Ago 29 12:06:07 ronanarraes-osd kernel: ---[ end trace 5774bd3049f78a63 ]---

Best regards,
Ronan Arraes

I was going to suggest that this was due to the fsync speedup patch that we were carrying but has since landed upstream in v4.8, but the Tumbleweed kernel doesn't contain that patch.

It looks like we have some digging to do.

-Jeff

Jeff Mahoney
SUSE Labs

hello,

For metadata, "bytes_may_use" is about 80GB, it's very big, I think this value is very abnormal.

So this explains why you have huge unallocated space, you still get ENOSPC error. In kernel btrfs, there is a function should_alloc_chunk() to determine whether to allocate new chunks(new device space)
num_bytes = total_bytes - bytes_readonly; it's 2147483648 num_allocated = bytes_used + bytes_reserved; it's 977354752

if num_allocated < num_bytes * 0.8, it will not allocate new device space :) even you have huge unallocated space.

I think the root reason is that bytes_may_use has some computation error and is not be converted to bytes_used or bytes_reserved.

I just explain why you get ENOSPC error even with huge unallocated space from codes :)

Yes, I know these WARNINGs, but indeed they are already results, we don't know the procedures which cause these results.

Can you work out a reproducer for this ENOSPC error, then I can dig into codes to figure out the true reason.

Regards,
Xiaoguang Wang

Hi!

Em Ter, 2016-08-30 às 10:12 +0800, Wang Xiaoguang escreveu:

Thanks! At least we known why ENOSPC is happening.

Unfortunately I failed in every attempt to trigger the problem. It happens randomly and I could not figure out yet what was triggering it. First, I though it was related to a build process inside a chroot jail, but then I see the problem happening after the computer being idle for a long time (+- 1h). So, no clues yet :(

Is there any workaround I can do?

Best regards,
Ronan Arraes

It sounds related to read-only snapshots to me. I wonder if this system has something busy that's writing to a file, database, even maybe something just spamming journald, and then there's a read-only snapshot during the write, which then triggers the enospc.

Ronan, if you're given a work around, then it's even less likely the bug gets fixed. But if you can disable snapper snapshots entirely and the problem doesn't happen; or if you can increase the frequency of snapper snapshots and the problem happens more often, that might help narrow it down to a point where it's more easily reproduced. If it's not related, that's still useful to know.

Chris Murphy

Em Ter, 2016-08-30 às 10:44 -0600, Chris Murphy escreveu:

I saw the problem yesterday after lunch time (13:00) and the last snapper snapshot was taken at 10:17:

snapper list
Tipo   | #  | Pre # | Data                         | Usuário | Limpeza | Descrição           | Dados de usuário
-------+----+-------+------------------------------+----------+------
---+-----------------------+------------------

single | 0  |       |                              |
root     |         | current               |                  
single | 1  |       | Ter 16 Ago 2016 15:07:25 BRT |
root     |         | first root filesystem |                  
single | 2  |       | Ter 16 Ago 2016 15:15:57 BRT | root     |
number  | after installation    | important=yes    
pre    | 4  |       | Ter 16 Ago 2016 15:26:44 BRT | root     |
number  | zypp(y2base)          | important=yes    
post   | 5  | 4     | Ter 16 Ago 2016 16:12:46 BRT | root     |
number  |                       | important=yes    
pre    | 29 |       | Ter 16 Ago 2016 18:02:43 BRT | root     |
number  | zypp(zypper)          | important=yes    
post   | 30 | 29    | Ter 16 Ago 2016 18:07:34 BRT | root     |
number  |                       | important=yes    
pre    | 45 |       | Seg 22 Ago 2016 13:59:45 BRT | root     |
number  | zypp(zypper)          | important=yes    
post   | 46 | 45    | Seg 22 Ago 2016 14:11:17 BRT | root     |
number  |                       | important=yes    
pre    | 89 |       | Seg 29 Ago 2016 09:56:19 BRT | root     |
number  | yast sw_single        |                  
pre    | 90 |       | Seg 29 Ago 2016 10:00:00 BRT | root     |
number  | zypp(y2base)          | important=no     
post   | 91 | 90    | Seg 29 Ago 2016 10:01:11 BRT | root     |
number  |                       | important=no     
pre    | 92 |       | Seg 29 Ago 2016 10:07:01 BRT | root     |
number  | zypp(y2base)          | important=no     
post   | 93 | 92    | Seg 29 Ago 2016 10:07:10 BRT | root     |
number  |                       | important=no     
pre    | 94 |       | Seg 29 Ago 2016 10:12:32 BRT | root     |
number  | zypp(y2base)          | important=no     
post   | 95 | 94    | Seg 29 Ago 2016 10:14:25 BRT | root     |
number  |                       | important=no     
post   | 96 | 89    | Seg 29 Ago 2016 10:17:17 BRT | root     |
number  |                       |                 

I agree with you. The problem is that since this is a production machine, it is kind very problematic to have so many reboots that occurs randomly.

I will install something using zypper, which will trigger snapper, and see if the problem will be triggered. I will be out of the office this afternoon, so the machine will be on idle.

Best regards,
Ronan Arraes

Hi guys!

And the problem happened again. This time, I was only using Mozilla Firefox. I could get the very first message after the error. I hope it brings more information:

[28039.672199] ------------[ cut here ]------------
[28039.672253] WARNING: CPU: 3 PID: 31800 at ../fs/btrfs/qgroup.c:2667
btrfs_qgroup_free_meta+0x88/0x90 [btrfs]
[28039.672255] Modules linked in: fuse nf_log_ipv6 xt_pkttype

nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_raw ipt_REJECT nf_reject_ipv4 iptable_raw xt_CT nvidia_drm(PO) nvidia_modeset(PO) iptable_filter nvidia(PO) ip6table_mangle nf_conntrack_netbios_ns nf_conntrack_broadcast drm_kms_helper nf_conntrack_ipv4 drm nf_defrag_ipv4 fb_sys_fops snd_hda_codec_hdmi joydev snd_hda_codec_realtek ip_tables syscopyarea snd_hda_codec_generic xt_conntrack snd_hda_intel sysfillrect intel_rapl sb_edac edac_core snd_hda_codec hp_wmi x86_pkg_temp_thermal intel_powerclamp snd_hda_core snd_hwdep nf_conntrack sparse_keymap sysimgblt coretemp kvm_intel kvm rfkill irqbypass snd_pcm snd_timer crct10dif_pclmul

[28039.672305]  e1000e crc32_pclmul ghash_clmulni_intel snd aesni_intel

ip6table_filter aes_x86_64 lrw gf128mul glue_helper ablk_helper iTCO_wdt iTCO_vendor_support mei_wdt ioatdma pcspkr cryptd ip6_tables ptp lpc_ich fjes i2c_i801 dca mfd_core soundcore pps_core shpchp tpm_infineon tpm_tis tpm mei_me mei x_tables btrfs xor raid6_pq hid_generic usbhid crc32c_intel serio_raw xhci_pci ehci_pci sr_mod firewire_ohci xhci_hcd ehci_hcd cdrom firewire_core crc_itu_t isci usbcore usb_common libsas ata_generic mpt3sas raid_class scsi_transport_sas wmi button sg

[28039.672373] CPU: 3 PID: 31800 Comm: gnome-terminal- Tainted:
P        W  O    4.7.1-1-default #1
[28039.672375] Hardware name: Hewlett-Packard HP Z820 Workstation/158B,
BIOS J63 v03.65 12/19/2013
[28039.672378]  0000000000000000 ffffffff81393104 0000000000000000
0000000000000000
[28039.672382]  ffffffff8107ca1e ffff881008780800 0000000000014000
ffff881008780800
[28039.672386]  ffffffffffffffe4 ffff88100b297c00 ffff88053b7e3540
ffffffffa02c9f58
[28039.672390] Call Trace:
[28039.672406]  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
[28039.672413]  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
[28039.672419]  [<ffffffff8102fe41>] show_stack+0x21/0x40
[28039.672425]  [<ffffffff81393104>] dump_stack+0x5c/0x78
[28039.672430]  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
[28039.672461]  [<ffffffffa02c9f58>] btrfs_qgroup_free_meta+0x88/0x90
[btrfs]
[28039.672492]  [<ffffffffa0261023>] start_transaction+0x3c3/0x4f0
[btrfs]
[28039.672521]  [<ffffffffa0271078>] btrfs_create+0x38/0x1d0 [btrfs]
[28039.672528]  [<ffffffff8121e8fb>] path_openat+0x139b/0x14a0
[28039.672535]  [<ffffffff8121fb2e>] do_filp_open+0x7e/0xe0
[28039.672541]  [<ffffffff8120e7a4>] do_sys_open+0x124/0x1f0
[28039.672547]  [<ffffffff816bb4f6>]
entry_SYSCALL_64_fastpath+0x1e/0xa8
[28039.676186] DWARF2 unwinder stuck at

entry_SYSCALL_64_fastpath+0x1e/0xa8

Best regards,
Ronan

Does this file system have quota enabled?

I'm testing this right now and can't even figure out how to determine when quota is enabled on a Btrfs file system. There's enable, disable, and rescan. If it's enabled or disabled, I get the same message if I rescan. If I mount the file system with quota previously enabled, there is no mount time notification that quota is enabled.

I sincerely hope opensuse isn't enabled quota by default.

Chris Murphy

OK it looks like with -w flag I can get a reliable indication of whether quota is enabled or not:

[root@f24s ~]# btrfs quota enable /mnt/0
[root@f24s ~]# btrfs quota rescan -w /mnt/0
quota rescan started
[root@f24s ~]# btrfs quota disable /mnt/0
[root@f24s ~]# btrfs quota rescan -w /mnt/0
ERROR: quota rescan failed: Invalid argument

So if you did not enable quota support, and aren't sure if it's enabled you can try 'btrfs quota rescan -w <mp>' but this might actually be a bad idea, a rescan could take a while if you're actually using quotas, I have no idea because I don't use them.

Perhaps someone can point out an easier way to determine whether quotas are enabled?

Chris Murphy

It can take a while, but the code is smart enough not to get too much in the way of other activity. It maintains a progress marker and only does live accounting on extents that have already been scanned.

btrfs qgroup show <path>

If you get a message like:
ERROR: can't perform the search - No such file or directory
ERROR: can't list qgroups: No such file or directory

... it means there's no quota root and thus quotas aren't enabled.

-Jeff

Jeff Mahoney
SUSE Labs

Wow, thanks but that's not obvious at all. man btrfs quota is described as "btrfs-quota - control the global quota status of a btrfs filesystem" so it stands to reason the state command for whether it's enabled or disabled would be in that subcommand not in some other subcommand.

But this is sidetracking. Does Ronan's call trace showing /fs/btrfs/qgroup.c:2667

Chris Murphy

Agreed. The tools interface has some warts.

Yeah. That warning doesn't get checked unless they're enabled.

-Jeff

Jeff Mahoney
SUSE Labs

OK so Ronan, I'm gonna guess the simplest work around for your problem is to disable quota support, and see if the problem happens again.

If it doesn't happen again then it sounds like the reproduce steps are:

a. enable quota support
b. do something metadata heavy workload that's also maybe hitting fsync; from opensuse list the example that sometimes causes it:

  osc co home:Ronis_BR/julia
  cd home:Ronis_BR/julia
  osc build --root=`pwd`/jail openSUSE_Tumbleweed x86_64

I wonder if it's easier to hit it on a hard drive, slower fsyncs?

Chris Murphy

Hi!

Em Qua, 2016-08-31 às 17:09 -0600, Chris Murphy escreveu:

Look at the output of the command proposed by Jeff:

btrfs qgroup show /
qgroupid         rfer         excl
--------         ----         ----
0/5          16.00KiB     16.00KiB
0/257        16.00KiB     16.00KiB
0/258        16.30MiB     16.30MiB
0/259        11.65GiB    309.67MiB
0/260         2.34MiB      2.34MiB
0/261        16.00KiB     16.00KiB
0/262        13.19GiB     13.19GiB
0/263        16.00KiB     16.00KiB
0/264        60.00KiB     60.00KiB
0/265       480.00KiB    480.00KiB
0/266        16.00KiB     16.00KiB
0/267         2.00GiB      2.00GiB
0/268        16.00KiB     16.00KiB
0/269        16.00KiB     16.00KiB
0/270        16.00KiB     16.00KiB
0/271        16.00KiB     16.00KiB
0/272        16.00KiB     16.00KiB
0/273        16.00KiB     16.00KiB
0/274        16.00KiB     16.00KiB
0/275       205.78MiB    205.78MiB
0/276        16.00KiB     16.00KiB
0/277        48.00KiB     48.00KiB
0/278       328.41MiB    328.41MiB
0/283         3.92GiB     26.63MiB
0/285         3.93GiB      4.10MiB
0/294         7.84GiB    100.59MiB
0/330         7.98GiB      6.61MiB
0/332         8.32GiB     69.17MiB
0/353         9.53GiB     49.46MiB
0/355        10.51GiB    235.39MiB
0/415        11.54GiB      3.38MiB
0/416        11.54GiB    896.00KiB
0/417        11.57GiB      2.68MiB
0/418        11.57GiB    160.00KiB
0/419        11.54GiB      2.40MiB
0/420        11.54GiB    192.00KiB
0/421        11.62GiB      4.61MiB
0/422        11.83GiB    212.93MiB
0/427        11.64GiB      1.27MiB
0/428        11.65GiB      4.25MiB
1/0          16.11GiB      4.77GiB
255/262      13.19GiB     13.19GiB

This system was installed with Tumbleweed ISO and I did not change anything in btrfs options. Hence, it seems that openSUSE is enabling quotas by default. Now, I need to disable it and avoid triggering the problem. What is the best way I can do this? Is it OK to do just:

btrfs quota disable /

? Or do I need to format and recreate btrfs without quotas?

This sounds good! Actually, I'm using a 7200RPM hard driver.

Thank you all very much for all the help,
Ronan Arraes

Yes, you can just run btrfs quota disable / and it should work. This ironically reiterates that one of the bigger problems with BTRFS is that distros are enabling unstable and known broken features by default on install. I was pretty much dumbfounded when I first learned that OpenSUSE is enabling BTRFS qgroups by default since they are known to not work reliably and cause all kinds of issues.

Em Qui, 2016-09-01 às 09:21 -0400, Austin S. Hemmelgarn escreveu:

Thanks Austin! I executed the command and now I get:

btrfs qgroup show /
ERROR: can't perform the search - No such file or directory
ERROR: can't list qgroups: No such file or directory

as expected. Now I will wait for +- 1 week to see if the problem will occur and, if not, I will send an e-mail to openSUSE factory mailing list to start a discussion if it is better to not enable qgroups by default.

Best regards and thanks everyone for the help,
Ronan Arraes

I have a feeling that you'll probably have no issues.

As far as having qgroups enabled by default, I think the reasoning is to emulate having separate filesystems with their own space limits. I can entirely understand this use case, and TBH it's about the only use case I'd consider quota groups for (per-user subvolumes for home directories are great, but there are numerous perfectly legitimate reasons to have very large amounts of data in your home directory for very short periods of time, so I wouldn't personally use qgroups there). The problem arises from the fact that it doesn't _look_ like separate filesystems (single entry in df, all the mounts point at the same device, etc), and the standard of overloading ENOSPC to mean you've hit your quota leads to lots of confusion in this particular case (especially considering the free space issues that BTRFS is known to have from time to time).

Hi Jeff,

Em Qui, 2016-09-01 às 13:43 -0400, Jeff Mahoney escreveu:

Thanks for the prompt answer. I'm glad because space is not a concern here, at least now :) Hence, I have plenty time to wait for a proper fix. Until there, I will try to keep my snapshot count low.

Absolutely. It doesn't affect the ability to take, retain, or recover using snapshots. It only affects the ability to see how much space a particular snapshot is using on disk, both from the user wanting to know and snapper using it to make retention decisions. Snapper can handle qgroups not being there.

-Jeff

Jeff Mahoney
SUSE Labs

It's not. We use qgroups because that's the only way we can track how much space each subvolume is using, regardless of whether anyone wants to do enforcement. When it's working properly, snapper can make use of that information to make informed decisions on how much space will actually be released when removing old snapshots.

On SUSE-based kernels, the inodes on different subvolumes report the anonymous device associated with the subvolume.

That said, I have a WIP that creates (and auto-tears down) vfsmounts for each subvolume. It's not all the way to a working df that would use the qgroup information to report space usage, but it's a start.

-Jeff

Jeff Mahoney
SUSE Labs

This is all well and good, but it ignores a few specific things:
1. There are numerous known issues with qgroups right now. This includes among other things returning ENOSPC when it should return EDQUOT (this isn't your fault, but you haven't tried to fix it either), and all kinds of general usability issues (systems tend to misbehave when at or near the quotas for example).
2. Snapper's default snapshot creation configuration is absolutely pathological in nature, generating insane amounts of background resource usage and taking up huge amounts of space. If this were changed, you would be a lot less dependent on being able to free up snapshots based on space usage.
3. It is fully possible (now, it may not have been when this choice was made) to get this info without using qgroups. btrfs filesystem du can be used to determine essentially the same information (summing the values in the second column will give you a reasonable estimate of how much space deleting the snapshot will free).
4. Enabling such a marginal technology without user intervention with no warnings about it or other notice that it's being used is a pretty solid example of something that a developer should not do.

It's poor choices like this that fall into the category of 'Ooh, this looks cool, let's do it!' made by major distros that are most of the reason that BTRFS has such a bad reputation right now. This is not something that should reasonably be on a production system, especially considering that even most of the BTRFS developers don't use qgroups, and that apparently your own customer support people couldn't tell that qgroups were to blame (seriously, your _ABSOLUTE FIRST SUGGESTION_ should have been to disable qgroups and see if the issue went away).

I get that you want something on par with Windows Restore Points or the bootable snapshot functionality provided by ZFS on Solaris, but qgroups really aren't at all essential to that, and even if they were, such functionality isn't even remotely ready for production usage on Linux yet.

So in other words even more dependence on a feature that doesn't even work reliably?

Jeff, I'm a little bit irritated because I initially suspected in this thread that this was an opensuse issue. That I questioned the kernel as the source is really beside the point. You didn't even recognize this might be quota related based on what was going on, because you bounced him back to this list when I suggested he take the issue to the opensuse-factory list.

What Ronan was reporting was behavior that no one on this list has ever previously reported. And upstream does not have quotas enabled by default so there is no reason why any regular testers here would have come across this.

So now we come full circle and I have to call this a misfeature that's trying to make up for another one, which is neurotic levels of snapshots taken by snapper out of the box. There is no good goddamn reason for it to take 100 read only snapshots in two fucking days. It's any wonder why the results are pathological.

Chris Murphy

Hi Jeff,

Em Qui, 2016-09-01 às 13:12 -0400, Jeff Mahoney escreveu:

Given that, what am I loosing by disabling qgroups here? Will I still be able to recover my machine using snapshots (this saved my two or three times)?

Best regards,
Ronan Arraes

Yes, I've just confirmed this on the OpenSUSE Factory mailing list.
[1] This is default on Tumbleweed (devel) and Leap (stable), and also SLE 12 SP2.

The feature that depends on it, that's actually enabling it is snapper:
http://snapper.io/2016/05/18/space-aware-cleanup.html

That feature says "btrfs quota support looks mature enough" which is big news to me. If it's that mature, why not make it the mkfs default? Just turn it on for everyone out of the gate? And if it isn't that mature, is it really appropriate for broad, by default, silent deployment for opensuse stable, and SUSE enterprise? I'm surprised no one said on this list that qgroups were stable enough for widespread testing for list regulars first. It just suddenly ends up enabled across three major distro outputs?

Even the fucking error messages were misleading. It wasn't until the most recent call trace that qgroups was even considered as possibly being related to this. How is it that busting a quota limit doesn't cause a very explicit quota related message, rather than a generic enospc?

[1] https://lists.opensuse.org/opensuse-factory/2016-09/msg00033.html

Chris Murphy

That's diplomatic.

They know all of this already though, but instead of toning down snapper defaults, they're amping up the voluming by enabling quotas instead.

There is only one logical reason for this that I can thing of. They're trying to increase problem reports, presumably in order to smooth out noisy data, maybe even by getting better bug reports like Ronan's. But I think this is a specious policy.

Over on Factory list, they're trying to have this two ways. First they're saying quotas are stable as they've implemented them in the Leap 4.4 kernel. And they consider the btrfs-progs man page warning that quotas aren't yet stable even in 4.7, and aren't recommended unless the user will use them, is a bug that should be removed from their copy of the man page.

So, what are they using? Pulling out such warnings doesn't make upstream code backported to their 4.4 kernel magically stable. If they're using out of tree quota code, fine, remove the warnings. But then, what is this code? How does it interact with upstream kernels?

Chris Murphy

The　kernel warning is interesting.

It means qgroup is underflowing its reserved metadata space.
However although it's a warning, it won't really under flow the numbers, but decrease it to zero.

It shows there is something wrong with metadata allocation, but won't directly cause quota corruption.

Quota uses two isolated different system, one extent based for qgroup numbers, and one reserved space based for reserved space.

The latter one is only used to prevent user from exceeding qgroup limit, and if user doesn't use limit, it won't cause any qgroup corruption or ENOSPC.

Further more, if it's qgroup reserved space causing anything wrong, it won't return -ENOSPC, but -EDQUOT.

So, just as Wang suspected, there is something wrong with metadata allocation, causing the problem and triggering the qgroup warning.

Thankg,
Qu

Ok, so I think this is a race that can happen when one thread is starting a transaction and another thread is committing a transaction that involves creating a snapshot.

We reserve blocks at the top of start_transaction and that reservation stays with the root. In: btrfs_commit_transaction-> create_pending_snapshots-> create_pending_snapshot-> qgroup_account_snapshot-> commit_fs_roots, we clear that reservation from the root via btrfs_qgroup_free_meta_all, potentially while start_transaction is waiting to join a new transaction. Or not. It can happen asynchronously, which is the point of having the reservation prior to that.

So the thing is that this error can only occur if start_transaction fails after this race occurs. That, combined with your report that you were seeing ENOSPC instead of EDQUOT, leads me to believe that this is just a side effect of whatever is causing you to not hit ENOSPC. I expect that you'll see it again -- you just won't see the WARN_ON anymore since quotas are disabled. I suspect it's probably the btrfs_block_rsv_add call immediately after the reservation, but there's no way to tell without tracing.

-Jeff

Jeff Mahoney
SUSE Labs

There's no conspiracy to leverage the openSUSE user base to generate bug reports any more than enabling any other feature in Tumbleweed before SLES is. We've enabled qgroups by default so that snapper can make sane decisions based on space usage. That's it.

Yep. That's a bug in the man page. We do consider them stable. I see every btrfs bug that gets reported against SLE12 SP2, upon which the Leap kernel is based. Have there been qgroups bugs over the development cycle? You bet. There's a reason if you look at the commit log for qgroups over the past year, you'll see a bunch of fixes from SUSE developers.

I explained what I think Ronan's issue is in another part of the thread just now. I don't think that's a severe issue at all. Annoying? Sure, but I'm more concerned with the underlying ENOSPC issue. Without more info, I don't know what the cause of it is and when it was introduced.

We, like every other group of file system developers, run xfstests pretty religiously. Since qgroups are becoming a bigger part of the btrfs experience for our products, we test them specifically. Yes, there are xfstests /just/ for qgroups, but we also make it a point to run the entire xfstests suite with and without qgroups enabled. Since the requirement for snapper was to have accurate space tracking, that's what we've focused on.

I obviously can't open up the SLES bugzilla to the world, so you're going to have to take my word on this. For our 4.4-based kernel there are currently 3 qgroup related bugs. The first is a report about how annoying it is to see old qgroup items for removed subvolumes. The second is an accounting bug that is old and the developer just hasn't gotten around to closing it yet. The third is a real issue, where users can hit the qgroup limit and are then stuck, similar to how it used to be when you'd hit ENOSPC and couldn't remove files or subvolumes. My gut feeling is that it's the same kind of problem: Removing files involves allocating blocks to CoW the metadata and when you've hit your quota limit, you can't allocate the blocks. I expect the solution will be similar to the ENOSPC issue except that rather than keeping a pool around, we can just CoW knowing full well the intention is to release space. My team is working on that today and I expect a fix shortly.

-Jeff

Jeff Mahoney
SUSE Labs

Hi Jeff,

Em Sex, 2016-09-02 às 10:26 -0400, Jeff Mahoney escreveu:

Sorry, but I really need to humbly disagree with you. Look to what has already happened to me when the problem occurred (which is almost every day):

1) Firefox crash;
2) Libreoffice crash (auto-save stop working);
3) Can't save my work in any text editor (vim, neovim, gedit, etc.);
4) Sometimes I can't even log as root (in TTY or by su);
5) Sometimes only a hard-reset solves the problem;
6) I was left with a broken operational system when the problem occurred during a zypper dup.

I just can't tell you how much work I lost during those situations. So, I think we cannot call this issue just annoying. I think it is very severe.

Best regards,
Ronan Arraes

Sorry, I miscommunicated there. The WARN_ON is annoying. It's the underlying issue that's causing you to lose work that is the one that concerns me.

-Jeff

Jeff Mahoney
SUSE Labs

Hi Jeff,

Em Sex, 2016-09-02 às 10:48 -0400, Jeff Mahoney escreveu:

Oh, OK, I see, sorry about that :)

Thus, if disabling quotas does not help to fix my problem, is there any workaround you can think of to avoid the problem you suggested in the previous e-mail?

Best regards,
Ronan Arraes

Which part? The quota reservation race will go away with quotas disabled, so you won't get the WARN_ON. The ENOSPC issue needs more investigation before I can suggest a workaround/fix. I won't be able to get into that until Tuesday. (Start of a holiday weekend in the US).

-Jeff

Jeff Mahoney
SUSE Labs

Hi guys!

Jeff was right. I had the problem again today and quotas are disabled now. I couldn't get any useful message in log this time. Look at the metadata:

btrfs fi usage /
Overall:
    Device size:    1.26TiB
    Device allocated:   43.07GiB
    Device unallocated:    1.21TiB
    Device missing:      0.00B
    Used:   41.94GiB
    Free (estimated):    1.21TiB (min: 622.46GiB)
    Data ratio:       1.00
    Metadata ratio:       2.00
    Global reserve:  352.00MiB (used: 0.00B)

Data,single: Size:40.01GiB, Used:39.94GiB
   /dev/sda6   40.01GiB

Metadata,DUP: Size:1.50GiB, Used:1.00GiB
   /dev/sda6    3.00GiB

System,DUP: Size:32.00MiB, Used:16.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.21TiB

Any ideas to help me?

Regards,
Ronan Arraes

Hi again guys!

After I rebooted the computer, I still can't run balance on metatada:

btrfs balance start -musage=1 /
ERROR: error during balancing '/': No space left on device There may be more info in syslog - try dmesg | tail

dmesg shows:

[ 2022.530285] BTRFS info (device sda6): relocating block group
128509280256 flags 36
[ 2023.355206] BTRFS info (device sda6): relocating block group
127972409344 flags 36
[ 2024.265313] BTRFS info (device sda6): relocating block group
127435538432 flags 36
[ 2025.646712] BTRFS info (device sda6): relocating block group
126898667520 flags 36
[ 2026.794791] BTRFS info (device sda6): relocating block group
126361796608 flags 36
[ 2028.023517] BTRFS info (device sda6): relocating block group
125824925696 flags 36
[ 2028.881287] BTRFS info (device sda6): relocating block group
125288054784 flags 36
[ 2029.739342] BTRFS info (device sda6): relocating block group
124751183872 flags 36
[ 2030.631990] BTRFS info (device sda6): relocating block group
124214312960 flags 36
[ 2031.523176] BTRFS info (device sda6): relocating block group
123677442048 flags 36
[ 2032.407859] BTRFS info (device sda6): relocating block group
123140571136 flags 36
[ 2033.806672] BTRFS info (device sda6): relocating block group
122603700224 flags 36
[ 2035.237712] BTRFS info (device sda6): relocating block group
122066829312 flags 36
[ 2038.257268] BTRFS info (device sda6): relocating block group
122033274880 flags 34
[ 2039.911443] BTRFS info (device sda6): relocating block group
121496403968 flags 36
[ 2040.958106] BTRFS info (device sda6): relocating block group
120959533056 flags 36
[ 2041.841051] BTRFS info (device sda6): relocating block group
120422662144 flags 36
[ 2042.828359] BTRFS info (device sda6): relocating block group
119885791232 flags 36
[ 2044.297744] BTRFS info (device sda6): relocating block group
119348920320 flags 36
[ 2045.684932] BTRFS info (device sda6): relocating block group
118812049408 flags 36
[ 2046.761787] BTRFS info (device sda6): relocating block group
118275178496 flags 36
[ 2048.200756] BTRFS info (device sda6): relocating block group
117738307584 flags 36
[ 2049.806986] BTRFS info (device sda6): relocating block group
117201436672 flags 36
[ 2051.170470] BTRFS info (device sda6): relocating block group
116664565760 flags 36
[ 2051.910536] BTRFS info (device sda6): relocating block group
116127694848 flags 36
[ 2052.678395] BTRFS info (device sda6): relocating block group
115590823936 flags 36
[ 2053.737959] BTRFS info (device sda6): relocating block group
106363355136 flags 36
[ 2054.852065] BTRFS info (device sda6): relocating block group
105826484224 flags 36
[ 2055.911187] BTRFS info (device sda6): relocating block group
105222504448 flags 36
[ 2057.047407] BTRFS info (device sda6): 4 enospc errors during balance

and I have:

btrfs fi usage /
Overall:
    Device size:    1.26TiB
    Device allocated:   80.07GiB
    Device unallocated:    1.18TiB
    Device missing:      0.00B
    Used:   41.95GiB
    Free (estimated):    1.18TiB (min: 603.95GiB)
    Data ratio:       1.00
    Metadata ratio:       2.00
    Global reserve:  352.00MiB (used: 576.00KiB)

Data,single: Size:40.01GiB, Used:39.95GiB
   /dev/sda6   40.01GiB

Metadata,DUP: Size:20.00GiB, Used:1.00GiB
   /dev/sda6   40.00GiB

System,DUP: Size:32.00MiB, Used:16.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.18TiB

Hope this brings new information!

Best regards,
Ronan Arraes

Except for your software build case, I have about the same workload you have with two machines, one SSD one HDD, using 4.7.0 for a month, and then 4.7.2 for the last week. I haven't had any enospc on these two systems.

I think for you the path of least resistance that also permits further testing is to see if you can track down the leap 42.2 beta kernel which is 4.4.19-1-default. I'm not easily finding that particular one, but I did find something a bit more recent:
http://download.opensuse.org/repositories/Kernel:/openSUSE-42.2/standard/x86_64/

Chris Murphy

Hi!

Em Sex, 2016-09-02 às 15:34 -0600, Chris Murphy escreveu:

Unfortunately, it will not be possible since my actual hardware depends on kernel >= 4.6 :(

Just now, I saw the problem again. For the first time, it happened twice in a small period. I was copying the e-mail from one IMAP server to my local HD. I use offlineimap, but this time it changed the backend to sqlite and started to create tons of database files, I think. My HDD IO stayed at 60/70% for a very long period.

Hence, let's do a review of situations in which I saw the problem:

1) Local builds using osc;
2) During zypper dup;
3) When offlineimap created tons of database files;
4) During rsync-ing /home;
4) During usage of a virtual machine (the disk image was in an EXT4 partition).

I think we can conclude that this problem is tightly coupled with actions that require a lot of writing to the HDD. Here is the specification of my HDD:

hdparm -I /dev/sda

/dev/sda:

ATA device, with non-removable media

	Model Number:       ST2000DM001-1CH164                      
	Serial Number:      W1E73CF5            
	Firmware Revision:  HP34    
	Transport:          Serial, SATA 1.0a, SATA II Extensions, SATA
Rev 2.5, SATA Rev 2.6, SATA Rev 3.0
Standards:
	Used: unknown (minor revision code 0x001f) 
	Supported: 9 8 7 6 5 
	Likely used: 9

Configuration:
Logical max current
cylinders 16383 16383
heads 16 16
sectors/track 63 63
--
CHS current addressable sectors:   16514064
LBA    user addressable sectors:  268435455
LBA48  user addressable sectors: 3907029168 Logical  Sector size:                   512 bytes Physical Sector size:                  4096 bytes Logical Sector-0 offset:                  0 bytes device size with M = 1024*1024:     1907729 MBytes device size with M = 1000*1000:     2000398 MBytes (2000 GB) cache/buffer size  = unknown
Form Factor: 3.5 inch
Nominal Media Rotation Rate: 7200
Capabilities:
LBA, IORDY(can be disabled)
Queue depth: 32
Standby timer values: spec'd by Standard, no device specific
minimum
R/W multiple sector transfer: Max = 16 Current = ?
Advanced power management level: 128
Recommended acoustic management value: 208, current value: 0
DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 *udma5 Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
     Cycle time: no flow control=120ns  IORDY flow
control=120ns
Commands/features:
Enabled Supported:
   * SMART feature set
     Security Mode feature set
   * Power Management feature set
   * Write cache
   * Look-ahead
   * WRITE_BUFFER command
   * READ_BUFFER command
   * DOWNLOAD_MICROCODE
   * Advanced Power Management feature set
     Power-Up In Standby feature set
   * SET_FEATURES required to spinup after power up
   * 48-bit Address feature set
   * Device Configuration Overlay feature set
   * Mandatory FLUSH_CACHE
   * FLUSH_CACHE_EXT
   * SMART error logging
   * SMART self-test
   * General Purpose Logging feature set
   * 64-bit World wide name
   * WRITE_UNCORRECTABLE_EXT command
   * {READ,WRITE}_DMA_EXT_GPL commands
   * Segmented DOWNLOAD_MICROCODE
   * Gen1 signaling speed (1.5Gb/s)
   * Gen2 signaling speed (3.0Gb/s)
   * Gen3 signaling speed (6.0Gb/s)
   * Native Command Queueing (NCQ)
   * Phy event counters
   * READ_LOG_DMA_EXT equivalent to READ_LOG_EXT
   * DMA Setup Auto-Activate optimization
     Device-initiated interface power management
   * Software settings preservation
   * SMART Command Transport (SCT) feature set
   * SCT Read/Write Long (AC1), obsolete
   * SCT Error Recovery Control (AC3)
   * SCT Features Control (AC4)
   * SCT Data Tables (AC5)
     unknown 206[12] (vendor specific)
     unknown 206[13] (vendor specific)
Security:
Master password revision code = 65534
supported
not enabled
not locked
not frozen
not expired: security count
supported: enhanced erase
212min for SECURITY ERASE UNIT. 212min for ENHANCED SECURITY
ERASE UNIT.
Logical Unit WWN Device Identifier: 5000c50072f7ce86

	NAA		: 5
	IEEE OUI	: 000c50
	Unique ID	: 072f7ce86

Checksum: correct

Best regards,
Ronan Arraes

Worth a shot, considering the opensuse/SLE 4.4 kernel has a shittonne of backports. It seems unlikely to me opensuse intends to not support your hardware (skylake?)

I don't think there's anything remarkable about any of these. And I even do VM stuff on Btrfs. I also don't think it's the drive.

What it sounds like is possible, is the file system is now in some kind of weird metadata state and it keeps tripping up on that. There may be more than one bug going on, one that gets it into this state, and then one that face plants with enospc when it's encountered.

Chris Murphy

Hi guys!

Em Sex, 2016-09-02 às 16:39 -0600, Chris Murphy escreveu:

Actually it is a peripheral we use to program embedded systems here and the (proprietary) driver requires kernel >= 4.6. I barely use it. I am really thinking to transfer it to another machine just to be able to change my kernel.

I will post here one thing I already posted on openSUSE mailing list:

I think I forgot to mention one very important thing: I have been using Tumbleweed+BTRFS on this machine for a very very very long time. I think I installed it just after it changed to the current model. By that time, I was using the same machine but without one peripheral that requires a "new" kernel (HDD, processor, RAM, everything was the same). AFAIK, the first time I saw that problem was this year. So, I think it must be a regression after some kernel / btrfs-progs update.

Best regards,
Ronan Arraes

I suggest removing the hardware, and the proprietary driver, and retest the system with the existing Tumbleweed 4.7.0 kernel; and if that still fails, then try the Leap 4.4 kernel.

Proprietary kernels can do all kinds of crazy things they shouldn't so it's entirely possible that driver is a factor in the problem.

Chris Murphy

Hi Chris,

Em Sex, 2016-09-02 às 21:41 -0600, Chris Murphy escreveu:

Actually it is just a module that I load. It is only loaded when I need to work with it. However, I can assure this is not the problem because I installed the board one month ago +-, but I have been seeing ENOSPC since the beginning of the year IIRC. I am using Tumbleweed default kernel right now, but I just can try Leap when 42.2 is released.

If you want a work around sooner than later, pick up one of the latest Leap 42.2 kernels from the URL I provided, I haven't tried it but it ought to work. Leap 42.2 isn't going to be released for another 2.5 months.

Chris Murphy

Just like what Wang has mentioned, would you please paste all the output of the contents of /sys/fs/btrfs/<your fs uuid>/allocation?

It's recommended to use "grep . -IR <path>" to get all the data as it will show the file name.

Thanks,
Qu

Hi all!

Em Seg, 2016-09-05 às 16:49 +0800, Qu Wenruo escreveu:

So, one more time, I see the problem. This time I was just using Firefox and I cannot recover using btrfs balance. I think that, one more time, I will need to reboot this machine. This problem is really causing me a lot of troubles :(

I have disabled the quotas and the first error message after the problem was:

[ 2444.592255] ------------[ cut here ]------------
[ 2444.592314] WARNING: CPU: 4 PID: 289 at ../fs/btrfs/extent-
tree.c:4303 btrfs_free_reserved_data_space_noquota+0xfe/0x110 [btrfs]
[ 2444.592317] Modules linked in: fuse nf_log_ipv6 xt_pkttype

nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_raw nvidia_drm(PO) ipt_REJECT nf_reject_ipv4 snd_hda_codec_hdmi nvidia_modeset(PO) intel_rapl sb_edac edac_core x86_pkg_temp_thermal intel_powerclamp nvidia(PO) coretemp snd_hda_codec_realtek iTCO_wdt snd_hda_codec_generic iptable_raw drm_kms_helper snd_hda_intel drm xt_CT snd_hda_codec snd_hda_core snd_hwdep kvm_intel snd_pcm snd_timer joydev mei_wdt fb_sys_fops iTCO_vendor_support i2c_i801 lpc_ich kvm syscopyarea snd sysfillrect irqbypass mei_me hp_wmi sysimgblt iptable_filter crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw gf128mul glue_helper ablk_helper

[ 2444.592386]  cryptd soundcore mei sparse_keymap rfkill e1000e shpchp

pcspkr ioatdma mfd_core tpm_infineon tpm_tis dca tpm fjes ptp pps_core ip6table_mangle nf_conntrack_netbios_ns nf_conntrack_broadcast nf_conntrack_ipv4 nf_defrag_ipv4 ip_tables xt_conntrack nf_conntrack ip6table_filter ip6_tables x_tables btrfs xor raid6_pq hid_generic usbhid crc32c_intel serio_raw xhci_pci ehci_pci xhci_hcd ehci_hcd firewire_ohci sr_mod firewire_core cdrom crc_itu_t usbcore isci usb_common libsas ata_generic mpt3sas raid_class scsi_transport_sas wmi button sg

[ 2444.592447] CPU: 4 PID: 289 Comm: kworker/u65:7 Tainted:
P        W  O    4.7.1-1-default #1
[ 2444.592450] Hardware name: Hewlett-Packard HP Z820 Workstation/158B,
BIOS J63 v03.65 12/19/2013
[ 2444.592458] Workqueue: writeback wb_workfn (flush-btrfs-1)
[ 2444.592462]  0000000000000000 ffffffff81393104 0000000000000000
0000000000000000
[ 2444.592468]  ffffffff8107ca1e ffff88080de6d800 0000000000009000
ffff88080c437a00
[ 2444.592472]  ffff880634b379ac 0000000000009000 ffff88080dcfb73c
ffffffffa02af98e
[ 2444.592477] Call Trace:
[ 2444.592499]  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
[ 2444.592507]  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
[ 2444.592514]  [<ffffffff8102fe41>] show_stack+0x21/0x40
[ 2444.592523]  [<ffffffff81393104>] dump_stack+0x5c/0x78
[ 2444.592531]  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
[ 2444.592561]  [<ffffffffa02af98e>]
btrfs_free_reserved_data_space_noquota+0xfe/0x110 [btrfs]
[ 2444.592602]  [<ffffffffa02cc036>] btrfs_clear_bit_hook+0x296/0x380
[btrfs]
[ 2444.592642]  [<ffffffffa02e9755>] clear_state_bit+0x55/0x1d0 [btrfs]
[ 2444.592676]  [<ffffffffa02e9a0d>] __clear_extent_bit+0x13d/0x3f0
[btrfs]
[ 2444.592707]  [<ffffffffa02ea8d2>]
extent_clear_unlock_delalloc+0x62/0x280 [btrfs]
[ 2444.592739]  [<ffffffffa02d1c19>] cow_file_range+0x299/0x440 [btrfs]
[ 2444.592768]  [<ffffffffa02d2cf2>] run_delalloc_range+0x392/0x3b0
[btrfs]
[ 2444.592801]  [<ffffffffa02eb090>]
writepage_delalloc.isra.40+0x100/0x170 [btrfs]
[ 2444.592834]  [<ffffffffa02ed9d3>] __extent_writepage+0xc3/0x340
[btrfs]
[ 2444.592864]  [<ffffffffa02ede8b>]
extent_write_cache_pages.isra.36.constprop.53+0x23b/0x350 [btrfs]
[ 2444.592894]  [<ffffffffa02ee4fe>] extent_writepages+0x4e/0x60
[btrfs]
[ 2444.592900]  [<ffffffff8123c64d>]
__writeback_single_inode+0x3d/0x3b0
[ 2444.592907]  [<ffffffff8123ce8a>] writeback_sb_inodes+0x20a/0x440
[ 2444.592914]  [<ffffffff8123d147>] __writeback_inodes_wb+0x87/0xb0
[ 2444.592921]  [<ffffffff8123d49d>] wb_writeback+0x28d/0x330
[ 2444.592927]  [<ffffffff8123dbe2>] wb_workfn+0x222/0x3f0
[ 2444.592934]  [<ffffffff810950ed>] process_one_work+0x1ed/0x4e0
[ 2444.592942]  [<ffffffff81095427>] worker_thread+0x47/0x4c0
[ 2444.592947]  [<ffffffff8109affd>] kthread+0xbd/0xe0
[ 2444.592954]  [<ffffffff816bb71f>] ret_from_fork+0x1f/0x40
[ 2444.596679] DWARF2 unwinder stuck at ret_from_fork+0x1f/0x40

[ 2444.596683] Leftover inexact backtrace:

[ 2444.596689]  [<ffffffff8109af40>] ? kthread_worker_fn+0x170/0x170

I will also provide the information requested by Qu:

grep . -IR /sys/fs/btrfs/e9efaa0c-d477-4249-830f-
ee5956768b29/allocation
allocation/data/flags:1
allocation/data/bytes_pinned:0
allocation/data/bytes_may_use:0
allocation/data/total_bytes_pinned:202973265920
allocation/data/bytes_reserved:0
allocation/data/bytes_used:45623730176
allocation/data/single/used_bytes:45623730176
allocation/data/single/total_bytes:46179287040
allocation/data/total_bytes:46179287040
allocation/data/disk_total:46179287040
allocation/data/disk_used:45623730176
allocation/metadata/dup/used_bytes:1120698368
allocation/metadata/dup/total_bytes:6979321856
allocation/metadata/flags:4
allocation/metadata/bytes_pinned:0
allocation/metadata/bytes_may_use:88521768960
allocation/metadata/total_bytes_pinned:-44285952
allocation/metadata/bytes_reserved:0
allocation/metadata/bytes_used:1120698368
allocation/metadata/total_bytes:6979321856
allocation/metadata/disk_total:13958643712
allocation/metadata/disk_used:2241396736
allocation/global_rsv_size:385875968
allocation/global_rsv_reserved:385875968
allocation/system/dup/used_bytes:16384
allocation/system/dup/total_bytes:33554432
allocation/system/flags:2
allocation/system/bytes_pinned:0
allocation/system/bytes_may_use:0
allocation/system/total_bytes_pinned:0
allocation/system/bytes_reserved:0
allocation/system/bytes_used:16384
allocation/system/total_bytes:33554432
allocation/system/disk_total:67108864
allocation/system/disk_used:32768

Additional information:

btrfs fi usage /
Overall:
    Device size:    1.26TiB
    Device allocated:   56.07GiB
    Device unallocated:    1.20TiB
    Device missing:      0.00B
    Used:   44.58GiB
    Free (estimated):    1.20TiB (min: 616.41GiB)
    Data ratio:       1.00
    Metadata ratio:       2.00
    Global reserve:  368.00MiB (used: 0.00B)

Data,single: Size:43.01GiB, Used:42.49GiB
   /dev/sda6   43.01GiB

Metadata,DUP: Size:6.50GiB, Used:1.04GiB
   /dev/sda6   13.00GiB

System,DUP: Size:32.00MiB, Used:16.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.20TiB

Can anyone help me?

Best regards,
Ronan Arraes

I have a hunch the list is about to be flooded with similar reports if we don't find this one before 4.8.

commit d555b6c380c644af63dbdaa7cc14bba041a4e4dd

Author: Josef Bacik <[email protected]>
Date:   Fri Mar 25 13:25:51 2016 -0400

Btrfs: warn_on for unaccounted spaces

This commit isn't the source of the bug, but it's making it a lot more noisy. I spent a few hours last night trying to track down why xfstests was throwing these warnings and I was able to reproduce them at least as far back as 4.4-vanilla with -oenospc_debug enabled.

Speaking of which, can you turn on mounting with -oenospc_debug if you haven't already?

In my case, space_info->bytes_may_use was getting accounted incorrectly.

I am able to reproduce that even with the following commit:
commit 18513091af9483ba84328d42092bd4d42a3c958f

Author: Wang Xiaoguang <[email protected]>
Date:   Mon Jul 25 15:51:40 2016 +0800

btrfs: update btrfs_space_info's bytes_may_use timely

That adds up to ~ 189 GB. total_bytes is only about 42 GB.

... well that's certainly interesting. It looks like we'll need to see how that happened. It seems like we've messed up at least that portion of accounting.

-Jeff

Jeff Mahoney
SUSE Labs

And the btrfs_free_reserved_data_space_noquota WARN_ON I was seeing is fixed by:

commit ed7a6948394305b810d0c6203268648715e5006f

Author: Wang Xiaoguang <[email protected]>
Date:   Fri Aug 26 11:33:14 2016 +0800

btrfs: do not decrease bytes_may_use when replaying extents

... which shouldn't change anything for your issue, unfortunately.

I still see these:

WARNING: CPU: 2 PID: 8166 at ../fs/btrfs/extent-tree.c:9582
btrfs_free_block_groups+0x2a8/0x400 [btrfs]()
Modules linked in: loop dm_flakey af_packet iscsi_ibft iscsi_boot_sysfs

msr ext4 crc16 mbcache jbd2 ipmi_ssif dm_mod igb ptp pps_core acpi_cpufreq tpm_infineon kvm_amd ipmi_si kvm dca pcspkr ipmi_msghandler 8250_fintek sp5100_tco fjes irqbypass i2c_piix4 shpchp processor button amd64_edac_mod edac_mce_amd edac_core k10temp btrfs xor raid6_pq sd_mod ata_generic mgag200 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect ohci_pci sysimgblt ehci_pci serio_raw ohci_hcd fb_sys_fops pata_atiixp ehci_hcd ttm ahci libahci drm usbcore libata usb_common sg scsi_mod autofs4

CPU: 2 PID: 8166 Comm: umount Tainted: G        W
4.4.19-11.g81405db-vanilla #1
Hardware name: HP ProLiant DL165 G7, BIOS O37 10/17/2012
 0000000000000000 ffff880230317d10 ffffffff813170ec 0000000000000000
 ffffffffa0472528 ffff880230317d48 ffffffff8107d816 0000000000000000
 ffff88009ab03600 ffff8800ba106288 ffff8800ab75a000 ffff8800ba106200
Call Trace:
 [<ffffffff813170ec>] dump_stack+0x63/0x87
 [<ffffffff8107d816>] warn_slowpath_common+0x86/0xc0
 [<ffffffff8107d90a>] warn_slowpath_null+0x1a/0x20
 [<ffffffffa03de3a8>] btrfs_free_block_groups+0x2a8/0x400 [btrfs]
 [<ffffffffa03ef24b>] close_ctree+0x15b/0x330 [btrfs]
 [<ffffffffa03bfeb9>] btrfs_put_super+0x19/0x20 [btrfs]
 [<ffffffff811fe5bf>] generic_shutdown_super+0x6f/0x100
 [<ffffffff811fe662>] kill_anon_super+0x12/0x20
 [<ffffffffa03c4fa8>] btrfs_kill_super+0x18/0x120 [btrfs]
 [<ffffffff811fe003>] deactivate_locked_super+0x43/0x70
 [<ffffffff811fe076>] deactivate_super+0x46/0x60
 [<ffffffff81219dcf>] cleanup_mnt+0x3f/0x80
 [<ffffffff81219e62>] __cleanup_mnt+0x12/0x20
 [<ffffffff81099fb6>] task_work_run+0x86/0xb0
 [<ffffffff81078806>] exit_to_usermode_loop+0x73/0xa2
 [<ffffffff81003b2d>] syscall_return_slowpath+0x8d/0xa0
 [<ffffffff815f928c>] int_ret_from_sys_call+0x25/0x8f
---[ end trace 09a0cc2892b6305c ]---

BTRFS: space_info 1 has 7946240 free, is not full
BTRFS: space_info total=8388608, used=442368, pinned=0, reserved=0, may_use=4096, readonly=0

... where the value of may_use varies.

-Jeff

Jeff Mahoney
SUSE Labs

hello,

It maybe a irrelevant question, but do you have compression enabled?

Regards,
Xiaoguang Wang

Hi!

Em Ter, 2016-09-13 às 11:17 +0800, Wang Xiaoguang escreveu:

No, I do not have compression enabled. I'm using openSUSE's default configuration.

By the way, I was wrongly mounting the filesystem with enospc_debug. It turns out that I modified the fstab in a backup directory, sorry :) Now, I did it correctly so, hopefully, we will have much more information about the problem the next time I see it!

Best regards,
Ronan Arraes

What test are you seeing this with? Thanks,

Josef

Hi guys,

One more time I saw the problem. It begins to happen on a daily basis now. Unfortunately the enospc_debug flag did not help. I did not see any new information in the logs. This time, only a hard reset worked. I could not even reboot using gnome panel.

Best regards,
Ronan Arraes

I just started paying attention to this, the last kernel I saw you were running was 4.7. Have you tried a recent kernel, like chris's tree?

git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs.git for-linus-4.8

is what I would like you to try if not. Thanks,

Josef

btrfs/022 hits it every time for me.

-Jeff

Jeff Mahoney
SUSE Labs

Hi Josef,

Em Ter, 2016-09-13 às 17:01 -0400, Josef Bacik escreveu:

Unfortunately, since this is a production machine, I am not allowed to install unreleased kernels. If this is the only solution, I will need to wait for 4.8 or search if anyone has already backported the BTRFS patches for 4.7.

Best regards,
Ronan Arraes

Hi guys,

The problem happened again, but now it was way more serious. I was doing a big Tumbleweed update (4680 packages) and I got the ENOSPC during the update. To avoid being left with a broken system, as it has already happened in the past, I, unfortunately, needed to delete data that I really was not planning to. This is a disaster, because I have more than 1 TiB of **free space**.

After deleting 7GiB of data, I could run rebalance and the update finished successfully. However, the ENOSPC happened 3 more times (!) and I always needed to run rebalance to keep the update going.

Sometimes, during the rebalance, I saw the message:

[28736.688266] BTRFS info (device sda6): relocating block group
389998968832 flags 34
[28737.376302] BTRFS info (device sda6): found 4 extents
[28737.712815] BTRFS info (device sda6): relocating block group
343760961536 flags 36
[28738.010030] BTRFS info (device sda6): relocating block group
343224090624 flags 36
[28738.343461] BTRFS info (device sda6): relocating block group
342687219712 flags 36
[28738.660023] BTRFS info (device sda6): relocating block group
342150348800 flags 36
[28738.665241] use_block_rsv: 11 callbacks suppressed
[28738.665247] ------------[ cut here ]------------
[28738.665290] WARNING: CPU: 10 PID: 639 at ../fs/btrfs/extent-
tree.c:8097 btrfs_alloc_tree_block+0x3f1/0x4c0 [btrfs]
[28738.665292] BTRFS: block rsv returned -28
[28738.665295] Modules linked in: dm_mod fuse nf_log_ipv6 xt_pkttype

nf_log_ipv4 nf_log_common xt_LOG xt_limit af_packet iscsi_ibft iscsi_boot_sysfs msr ip6t_REJECT nf_reject_ipv6 xt_tcpudp nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_raw ipt_REJECT nf_reject_ipv4 iptable_raw xt_CT snd_hda_codec_hdmi snd_hda_codec_realtek nvidia_drm(PO) snd_hda_codec_generic snd_hda_intel nvidia_modeset(PO) snd_hda_codec snd_hda_core snd_hwdep iptable_filter nvidia(PO) joydev drm_kms_helper intel_rapl drm fb_sys_fops iTCO_wdt mei_wdt syscopyarea snd_pcm snd_timer iTCO_vendor_support sysfillrect sb_edac snd i2c_i801 mei_me lpc_ich edac_core sysimgblt ip6table_mangle x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel soundcore mei aes_x86_64

[28738.665359]  lrw gf128mul glue_helper ablk_helper cryptd e1000e

hp_wmi ioatdma fjes nf_conntrack_netbios_ns ptp shpchp pps_core sparse_keymap pcspkr mfd_core nf_conntrack_broadcast rfkill tpm_infineon tpm_tis dca tpm nf_conntrack_ipv4 nf_defrag_ipv4 ip_tables xt_conntrack nf_conntrack ip6table_filter ip6_tables x_tables btrfs xor raid6_pq hid_generic usbhid crc32c_intel serio_raw xhci_pci ehci_pci sr_mod firewire_ohci xhci_hcd ehci_hcd cdrom firewire_core crc_itu_t usbcore isci usb_common libsas ata_generic mpt3sas raid_class scsi_transport_sas wmi button sg

[28738.665419] CPU: 10 PID: 639 Comm: systemd-journal Tainted:
P        W  O    4.7.1-1-default #1
[28738.665421] Hardware name: Hewlett-Packard HP Z820 Workstation/158B,
BIOS J63 v03.65 12/19/2013
[28738.665425]  0000000000000000 ffffffff81393104 ffff88080bc63a68
0000000000000000
[28738.665430]  ffffffff8107ca1e ffff8804eaa73300 ffff88080bc63ab8
0000000000004000
[28738.665434]  0000000000000000 ffff88017be9a000 ffff880f51b31760
ffffffff8107ca8f
[28738.665438] Call Trace:
[28738.665464]  [<ffffffff8102ed5e>] dump_trace+0x5e/0x320
[28738.665472]  [<ffffffff8102f12c>] show_stack_log_lvl+0x10c/0x180
[28738.665478]  [<ffffffff8102fe41>] show_stack+0x21/0x40
[28738.665486]  [<ffffffff81393104>] dump_stack+0x5c/0x78
[28738.665496]  [<ffffffff8107ca1e>] __warn+0xbe/0xe0
[28738.665503]  [<ffffffff8107ca8f>] warn_slowpath_fmt+0x4f/0x60
[28738.665529]  [<ffffffffa029d911>] btrfs_alloc_tree_block+0x3f1/0x4c0
[btrfs]
[28738.665560]  [<ffffffffa02846a2>] btrfs_copy_root+0xf2/0x280 [btrfs]
[28738.665593]  [<ffffffffa02fd141>] create_reloc_root+0x171/0x1e0
[btrfs]
[28738.665623]  [<ffffffffa030316f>] btrfs_init_reloc_root+0x8f/0xa0
[btrfs]
[28738.665652]  [<ffffffffa02ac992>] record_root_in_trans+0xb2/0x110
[btrfs]
[28738.665679]  [<ffffffffa02adb11>]
btrfs_record_root_in_trans+0x41/0x70 [btrfs]
[28738.665704]  [<ffffffffa02afd00>] start_transaction+0xa0/0x4f0
[btrfs]
[28738.665732]  [<ffffffffa02b6153>] btrfs_dirty_inode+0x33/0xc0
[btrfs]
[28738.665741]  [<ffffffff8122aa59>] file_update_time+0x99/0xf0
[28738.665770]  [<ffffffffa02c11a3>] btrfs_page_mkwrite+0xa3/0x450
[btrfs]
[28738.665779]  [<ffffffff811bd2c9>] do_page_mkwrite+0x69/0xc0
[28738.665785]  [<ffffffff811c00f4>] handle_pte_fault+0xf4/0x1760
[28738.665792]  [<ffffffff811c1bfe>] handle_mm_fault+0x29e/0x5a0
[28738.665798]  [<ffffffff81064fc0>] __do_page_fault+0x1e0/0x510
[28738.665809]  [<ffffffff816bd608>] page_fault+0x28/0x30
[28738.669296] DWARF2 unwinder stuck at page_fault+0x28/0x30

[28738.669300] Leftover inexact backtrace:

[28738.669327] ---[ end trace 8ef9cfba38cc9bfc ]---

Look what happened to my METADATA during the update:

1) When the problem occured:

# btrfs fi usage /
Overall:
    Device size:		   1.26TiB
    Device allocated:		  63.07GiB
    Device unallocated:		   1.20TiB
    Device missing:		     0.00B
    Used:			  50.21GiB
    Free (estimated):		   1.20TiB	(min: 612.49GiB)
    Data ratio:			      1.00
    Metadata ratio:		      2.00
    Global reserve:		 400.00MiB	(used: 0.00B)

Data,single: Size:48.01GiB, Used:47.91GiB
   /dev/sda6   48.01GiB

Metadata,DUP: Size:7.50GiB, Used:1.15GiB
   /dev/sda6   15.00GiB

System,DUP: Size:32.00MiB, Used:16.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.20TiB

2) After deleting 7GiB of data and run rebalance:

# btrfs fi usage /
Overall:
    Device size:		   1.26TiB
    Device allocated:		 133.07GiB
    Device unallocated:		   1.13TiB
    Device missing:		     0.00B
    Used:			  43.16GiB
    Free (estimated):		   1.13TiB	(min: 584.46GiB)
    Data ratio:			      1.00
    Metadata ratio:		      2.00
    Global reserve:		 384.00MiB	(used: 0.00B)

Data,single: Size:48.01GiB, Used:40.94GiB
   /dev/sda6   48.01GiB

Metadata,DUP: Size:42.50GiB, Used:1.11GiB
   /dev/sda6   85.00GiB

System,DUP: Size:32.00MiB, Used:48.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.13TiB

3) After another rebalance (I saw the ENOSPC again):

# btrfs fi usage /
Overall:
    Device size:		   1.26TiB
    Device allocated:		 207.07GiB
    Device unallocated:		   1.05TiB
    Device missing:		     0.00B
    Used:			  43.87GiB
    Free (estimated):		   1.06TiB	(min: 540.83GiB)
    Data ratio:			      1.00
    Metadata ratio:		      2.00
    Global reserve:		 400.00MiB	(used: 0.00B)

Data,single: Size:42.01GiB, Used:41.57GiB
   /dev/sda6   42.01GiB

Metadata,DUP: Size:82.50GiB, Used:1.15GiB
   /dev/sda6  165.00GiB

System,DUP: Size:32.00MiB, Used:48.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6    1.05TiB

4) After another rebalance (I saw the ENOSPC again):

# btrfs fi usage /
Overall:
    Device size:		   1.26TiB
    Device allocated:		 344.07GiB
    Device unallocated:		 943.79GiB
    Device missing:		     0.00B
    Used:			  44.69GiB
    Free (estimated):		 944.45GiB	(min: 472.55GiB)
    Data ratio:			      1.00
    Metadata ratio:		      2.00
    Global reserve:		 416.00MiB	(used: 0.00B)

Data,single: Size:43.01GiB, Used:42.34GiB
   /dev/sda6   43.01GiB

Metadata,DUP: Size:150.50GiB, Used:1.17GiB
   /dev/sda6  301.00GiB

System,DUP: Size:32.00MiB, Used:80.00KiB
   /dev/sda6   64.00MiB

Unallocated:
   /dev/sda6  943.79GiB

Yes, 150 GiB of METADATA, 3x more than my actual data.

This problem is really causing me problems. I am starting to think that Tumbleweed, at least, should not choose BTRFS as the default file system, since this distribution is supposed to be stable. I think that BTRFS has some serious problems at least in kernels 4.6 and 4.7.

I reported this problem more than 1 month ago, and yet nobody could provide me at least a workaround so I can keep working here. I think the best will be to format this machine (**again**) and use EXT4 of XFS, if nobody could help me to fix or avoid this problem in the following days.

Best regards,
Ronan Arraes

All I can think of is the file system has gotten into a unique state through a combination of events. I'm still suspicious that qgroups is contributing to the problem even after being disabled. The workload you're talking about is completely ordinary and trivial.

The openSUSE layout is basically impossible to backup and restore, there's astrometric tons of snapshots, there's no recursive btrfs send/receive to try and migrate it to a new file system intact, so you'd pretty much just have to reinstall it no matter what. If it were me, reinstall with Btrfs same as now, and first thing before anything else I'd disable quotas. Or yeah, it's completely reasonable for you to move to a different file system, it's really a coin toss for ext4 vs XFS, but at least XFS now checksums metadata and the journal by default so if I thought about it at the time of the installation I'd do that.

Yeah FWIW, the devs seem to prefer the output from 'grep . -IR /sys/fs/btrfs/<fsuuid>/allocation/' so for these kinds of problems I'd report that.

Yeah holy crap weird.

But the fs is already in some funky state so at this point it's not surprising it continues to do crazy things. If the devs knew exactly what was going on, they'd say so. If they had a fix, they'd post it or at least an ETA. And while ostensibly the enospc work in 4.8 would work around this problem, it's unknown until it's tested.

If you *really* want to, you could grab a Fedora Rawhide nightly that has kernel 4.8 rc6 on it, with debug stuff enabled. If it face plants, it should catch useful stuff for Josef. If it doesn't, maybe it fixes enough things that you can get back to work for a while longer until a long term fix becomes available. The only way to know for sure is to test it. But it's completely sane to just switch to XFS and get back to work also.

Current
https://kojipkgs.fedoraproject.org/compose/rawhide/Fedora-Rawhide-20160914.n.0/compose/Everything/x86_64/iso/Fedora-Everything-netinst-x86_64-Rawhide-20160914.n.0.iso.n.0.iso

Use 'dd if=ISO of=USBstick bs=256K' that will boot anything, BIOS or UEFI. At the menu, choose Troubleshooting, then the Rescue option, at the next text menu choose 3 to get to a shell. And from there you can mount with enospc_debug, and do a balance of the file system. To get logs off the system, use a 2nd USB stick, or if you have wired ethernet use scp, or if you know nmcli you can maybe get the wireless up by command line.

Yep, completely reasonable.

Chris Murphy

Hi Chris,

Em Qua, 2016-09-14 às 16:25 -0600, Chris Murphy escreveu:

This seems reasonable. However, I formatted the computer and after two days, if I remember correctly, I started to see the problems again. I'm still thinking it should be also related to my HDD (7200 RPM). In all my other computers, everything is fine and I use SSD.

Thanks!

Yeah, unfortunately I forgot this one today :(

This seems good. However, I just have access to that machine during my working period, and I just does not have time to test this, sorry :(

Nevertheless, when you mentioned the dd command, I had a great idea that can help me to live with this problem until I have access to kernel 4.8. I will use dd to create, let's say, 100 files with 3 GiB each in my /home directory. Hence, when I see ENOSPC, I will just need to delete some of these files. I think this should work.

Thanks for all the advices Chris!

Best regards,
Ronan Arraes

hi,

btrfs/022 is not related to this enospc error.
Qu wenruo's patch “ btrfs: Fix leaking bytes_may_use after hitting EDQUOTA” has fixed this warning, please check his patch for detailed commit message.

Regards,
Xiaoguang Wang

Guys,

Something very strange happened. I have not seen the problem since Monday, which is pretty much the first time ever I work more than 3 days without seeing it.

Ok, it can be a coincidence. Notice that I did not change anything related to my work behavior. However, I did do two things:

_ Update the kernel to 4.7.2; and
_ Created 50 dummy files with 3.0 GiB each.

Can anyone, please, tell me if these things seems to be correlated?

Best regards,
Ronan Arraes

Yep, that's understood. This was just something I happened to encounter while looking at this.

-Jeff

Jeff Mahoney
SUSE Labs

Most likely the kernel upgrade fixed things. It's possible that the large allocation is impacting something and making it work, but I don't think that that is very likely.

Em qui, 2016-09-22 às 09:41 -0400, Austin S. Hemmelgarn escreveu:

The patches related to btrfs I could find in kernel 4.7.2 and 4.7.3 changelog are:

commit 8d32aaa89067225d4202a362dc201280e2514952
Author: Chris Mason <[email protected]>
Date:   Tue Jul 19 05:52:36 2016 -0700

    Btrfs: fix delalloc accounting after copy_from_user faults

commit f495a60eb6351bf2f29fdbc1854375df9fe4022b
Author: Paolo Valente <[email protected]> Date:   Wed Jul 27 07:22:05 2016 +0200

    block: add missing group association in bio-cloning functions

commit ff3235105fc7e4ecf04eb308940821d4a098c08d
Author: Jeff Mahoney <[email protected]> Date:   Wed Aug 17 21:58:33 2016 -0400

    btrfs: don't create or leak aliased root while cleaning up orphans

commit 64563a38fde57a26f4d68d488d0d4918f843547c
Author: Jeff Mahoney <[email protected]> Date:   Mon Aug 15 12:10:33 2016 -0400

    btrfs: properly track when rescan worker is running

commit 69b69167965e108a775ef20decabcc76fbe4fc08
Author: Jeff Mahoney <[email protected]> Date:   Mon Aug 8 22:08:06 2016 -0400

    btrfs: waiting on qgroup rescan should not always be interruptible

Best regards,
Ronan Arraes

This is what fixed it. I thought it was in 4.7 which is why I started paying attention, but I guess I was wrong. Glad your problem is resolved. Thanks,

Josef

Hi Josef,

Em qui, 2016-09-22 às 10:39 -0400, Josef Bacik escreveu:

Do you have any explanations why the problem solved by the patch was causing me the ENOSPC? Also, is it necessary to format my partition or should I consider it good for use after the installation of the new kernel?

Best regards,
Ronan Arraes

That patch fixed a problem where we would screw up the ENOSPC accounting, and would slowly leak space into one of the counters. So eventually (or often in your case) you'd hit ENOSPC, but have plenty of space available. If you unmounted and mounted again, or simply rebooted, everything would have been fine. You can still use the fs, the accounting is purely in memory so it's not like your FS is permanently screwed. Thanks,

Josef

Hi Josef,

Em qui, 2016-09-22 às 13:49 -0400, Josef Bacik escreveu:

Thank you very much for the explanation. I am very glad it is finally fixed here :)

Best regards,
Ronan Arraes

Hi guys!

After a week without experiencing the problem, I think we can mark this problem as solved. I want to thanks all the devs on this list. You were always very helpful. For anyone who is still experiencing the reported problem, upgrade to kernel 4.7.3 and I think you will be fine :)

Best regards and thank you all,
Ronan Arraes