Btrfs-progs release 3.17

Discussion:

David Sterba

2014-10-20 16:54:15 UTC

Hi,

the version 3.17 of btrfs-progs has been released.

git://git.kernel.org/pub/scm/linux/kernel/git/kdave/btrfs-progs.git v3.17
https://www.kernel.org/pub/linux/kernel/people/kdave/btrfs-progs/btrfs-progs-v3.17.tar.xz

Among other fixes and updates, there are many fsck improvements, most notably a
fix for the bug introduced in 3.17 regarding inconsistencies after read-only
snapshots (fix https://patchwork.kernel.org/patch/5086521/).

User visible changes:
* check: --init-csum-tree acutally does something useful, rebuilds the whole
csum tree
* /dev scanning for btrfs devices is gone
* /proc/partitions scanning is gone, blkid is used exclusively
* new subcommand 'subvolume sync'
* filesystem df: new options to set unit format
* convert: allow to copy label from the origin, or specify a new one
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Petr Janecek

2014-10-21 21:28:22 UTC

Permalink

Hello,

Post by David Sterba
the version 3.17 of btrfs-progs has been released.

on a system with 3-disk raid1 and 4 and 5-disk raid10 fs,
"btrfs filesystem show" now stalls for approx. half a minute after the
listing, just before the version information. During that time, it
often prints something like

[...]
devid 4 size 2.73TiB used 806.01GiB path /dev/sdh
devid 5 size 3.64TiB used 123.00GiB path /dev/sdp

parent transid verify failed on 2636556500992 wanted 2264 found 2268
parent transid verify failed on 2636556500992 wanted 2264 found 2268
parent transid verify failed on 2636556500992 wanted 2264 found 2268
parent transid verify failed on 2636556500992 wanted 2264 found 2268
Ignoring transid failure
extent buffer leak: start 2636556091392 len 16384
Btrfs v3.17

I've had a balance running that time, but "btrfs fi sh /any/btrfs/mount"
returned always quickly as in previous versions.

Now, with the balance paused, the stalls continue to appear, but no
messages as above. During the stalls, btrfs is reading the disks at rate
several MB/s, probably seek limited. /proc/<btrfs fi sh pid>/fd shows
that it scans again the devices it has already written about.

Umounted all the btrfs, checked them -- nothing wrong.

Resumed the balance -- error messages are (sometimes) back:

[...]
Label: 'BTR3' uuid: 0bf93650-fd12-4f82-8086-2c7484141e97
Total devices 5 FS bytes used 1.63TiB
devid 1 size 2.73TiB used 800.00GiB path /dev/sde
devid 2 size 2.73TiB used 800.00GiB path /dev/sdf
devid 3 size 2.73TiB used 800.03GiB path /dev/sdg
devid 4 size 2.73TiB used 801.00GiB path /dev/sdh
devid 5 size 3.64TiB used 143.03GiB path /dev/sdp

Check tree block failed, want=2730665181184, have=9727841172931608576
Check tree block failed, want=2730665181184, have=9727841172931608576
Check tree block failed, want=2730665181184, have=8388343671073157408
Check tree block failed, want=2730665181184, have=9727841172931608576
Check tree block failed, want=2730665181184, have=9727841172931608576
read block failed check_tree_block
Label: 'BTR1' uuid: 2ec4b5a1-a2c4-4687-9015-cf5b519297f3
Total devices 3 FS bytes used 2.47TiB
devid 1 size 3.64TiB used 1.79TiB path /dev/sdm
devid 2 size 3.64TiB used 1.79TiB path /dev/sdn
devid 3 size 3.64TiB used 1.79TiB path /dev/sdo

Btrfs v3.17

(or the "parent transid verify failed..." + "extent buffer leak..." as
in the first case as well)

Kernel version 3.18-rc1

Regards,

Petr
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Petr Janecek

2014-10-22 00:32:18 UTC

Permalink

Hello,
one more thing: I just overwrote part of one disk.
"btrfs filesystem show" could be more helpful diagnosing this:

# btrfs fi sh
Label: 'BTRFSROOT' uuid: d877125e-9b8d-47ea-b57b-7411292fd26c
Total devices 1 FS bytes used 2.91GiB
devid 1 size 29.44GiB used 5.04GiB path /dev/sdi1

Label: 'BTR1' uuid: 2ec4b5a1-a2c4-4687-9015-cf5b519297f3
Total devices 3 FS bytes used 2.47TiB
devid 1 size 3.64TiB used 1.79TiB path /dev/sdn
devid 2 size 3.64TiB used 1.79TiB path /dev/sdo
devid 3 size 3.64TiB used 1.79TiB path /dev/sdp

Label: 'BTR0' uuid: 0ec83db3-4574-4e40-8d57-ebbe9fe246e1
Total devices 4 FS bytes used 2.41TiB
devid 1 size 2.73TiB used 1.21TiB path /dev/sda
devid 2 size 2.73TiB used 1.21TiB path /dev/sdb
devid 3 size 2.73TiB used 1.21TiB path /dev/sdc
devid 4 size 2.73TiB used 1.21TiB path /dev/sdd

Label: 'BTR3' uuid: 89f2016b-984f-4811-99d9-df3b8b5ce318
Total devices 4 FS bytes used 896.00KiB
devid 1 size 2.73TiB used 2.02GiB path /dev/sde
devid 2 size 2.73TiB used 2.00GiB path /dev/sdf
devid 3 size 2.73TiB used 1.01GiB path /dev/sdg
devid 4 size 2.73TiB used 1.01GiB path /dev/sdh

warning, device 3 is missing
warning devid 3 not found already
Btrfs v3.17

-- there are three disks with devid 3, can't tell which one it is. The
fs uuid can be displayed with very little effort:

diff --git a/volumes.c b/volumes.c
index 5b007fc..02866bc 100644
--- a/volumes.c
+++ b/volumes.c
@@ -30,6 +30,7 @@
#include "print-tree.h"
#include "volumes.h"
#include "math.h"
+#include "utils.h"

struct stripe {
struct btrfs_device *dev;
@@ -1625,9 +1626,11 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
map->stripes[i].dev = btrfs_find_device(root, devid, uuid,
NULL);
if (!map->stripes[i].dev) {
+ char uuidbuf[BTRFS_UUID_UNPARSED_SIZE];
+ uuid_unparse(root->fs_info->fsid, uuidbuf);
map->stripes[i].dev = fill_missing_device(devid);
- printf("warning, device %llu is missing\n",
- (unsigned long long)devid);
+ printf("warning, device %llu from fs %s is missing\n",
+ (unsigned long long)devid, uuidbuf);
}

}
@@ -1720,8 +1723,10 @@ static int read_one_dev(struct btrfs_root *root,

device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
if (!device) {
- printk("warning devid %llu not found already\n",
- (unsigned long long)devid);
+ char uuidbuf[BTRFS_UUID_UNPARSED_SIZE];
+ uuid_unparse(fs_uuid, uuidbuf);
+ printk("warning devid %llu from fs %s not found already\n",
+ (unsigned long long)devid, uuidbuf);
device = kzalloc(sizeof(*device), GFP_NOFS);
if (!device)
return -ENOMEM;

Regards,

Petr
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Anand Jain

2014-10-22 08:49:12 UTC

Permalink

Hi,

You have mentioned two issues when balance and fi show running
concurrently
- stalling and
- errors.

first of all..
3.17 replaced our own system wide disk scan methods by lblkid scan
methods. lblkid with its feature-rich is slower as reported here.
https://www.mail-archive.com/linux-***@vger.kernel.org/msg37839.html

Next, as of 3.17 it still miss the poor man's optimizations patches
(without seeking for a new kernel interface)

[PATCH 1/1] btrfs-progs: code optimize cmd_scan_dev() use
btrfs_register_one_device()
[PATCH 1/2] btrfs-progs: introduce btrfs_register_all_device()
[PATCH 2/2] btrfs-progs: optimize btrfs_scan_lblkid() for multiple calls

If you could, pls..
Now on 3.17 apply above 3 patches and see if you see any better
performance for the stalling issue.
can you do same steps on 3.16 and report what you observe

Now about the 2nd issue errors when balance is running.. here is my
initial analysis with some guess work..

Post by Petr Janecek
just before the version information. During that time, it
often prints something like
parent transid verify failed on 2636556500992 wanted 2264 found 2268

Post by Petr Janecek
Check tree block failed, want=2730665181184, have=9727841172931608576

The default btrfs fi show will scan kernel for mounted FS first and
then for umounted btrfs disks.

When we already know about the mounted disks its just unnecessary that
we would still read the mounted disks and later drop them as seen,
looks like we need to write a real fix to replace this workaround fix..
btrfs filesystem show: skip duplicate fsids

-Anand

Post by Petr Janecek
Hello,

Post by David Sterba
the version 3.17 of btrfs-progs has been released.

on a system with 3-disk raid1 and 4 and 5-disk raid10 fs,
"btrfs filesystem show" now stalls for approx. half a minute after the
listing, just before the version information. During that time, it
often prints something like
[...]
devid 4 size 2.73TiB used 806.01GiB path /dev/sdh
devid 5 size 3.64TiB used 123.00GiB path /dev/sdp
parent transid verify failed on 2636556500992 wanted 2264 found 2268
parent transid verify failed on 2636556500992 wanted 2264 found 2268
parent transid verify failed on 2636556500992 wanted 2264 found 2268
parent transid verify failed on 2636556500992 wanted 2264 found 2268
Ignoring transid failure
extent buffer leak: start 2636556091392 len 16384
Btrfs v3.17
I've had a balance running that time, but "btrfs fi sh /any/btrfs/mount"
returned always quickly as in previous versions.
Now, with the balance paused, the stalls continue to appear, but no
messages as above. During the stalls, btrfs is reading the disks at rate
several MB/s, probably seek limited. /proc/<btrfs fi sh pid>/fd shows
that it scans again the devices it has already written about.
Umounted all the btrfs, checked them -- nothing wrong.
[...]
Label: 'BTR3' uuid: 0bf93650-fd12-4f82-8086-2c7484141e97
Total devices 5 FS bytes used 1.63TiB
devid 1 size 2.73TiB used 800.00GiB path /dev/sde
devid 2 size 2.73TiB used 800.00GiB path /dev/sdf
devid 3 size 2.73TiB used 800.03GiB path /dev/sdg
devid 4 size 2.73TiB used 801.00GiB path /dev/sdh
devid 5 size 3.64TiB used 143.03GiB path /dev/sdp
Check tree block failed, want=2730665181184, have=9727841172931608576
Check tree block failed, want=2730665181184, have=9727841172931608576
Check tree block failed, want=2730665181184, have=8388343671073157408
Check tree block failed, want=2730665181184, have=9727841172931608576
Check tree block failed, want=2730665181184, have=9727841172931608576
read block failed check_tree_block
Label: 'BTR1' uuid: 2ec4b5a1-a2c4-4687-9015-cf5b519297f3
Total devices 3 FS bytes used 2.47TiB
devid 1 size 3.64TiB used 1.79TiB path /dev/sdm
devid 2 size 3.64TiB used 1.79TiB path /dev/sdn
devid 3 size 3.64TiB used 1.79TiB path /dev/sdo
Btrfs v3.17
(or the "parent transid verify failed..." + "extent buffer leak..." as
in the first case as well)
Kernel version 3.18-rc1
Regards,
Petr
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
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 ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Petr Janecek

2014-10-23 06:57:09 UTC

Permalink

Hello,

Post by Anand Jain
You have mentioned two issues when balance and fi show running
concurrently

my mail was a bit chaotic, but I get the stalls even on idle system.
Today I got

parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
Ignoring transid failure
leaf parent key incorrect 1559973888000

from 'btrfs fi sh' while I was just copying something, no balance running.

[...]

Post by Anand Jain
[PATCH 1/1] btrfs-progs: code optimize cmd_scan_dev() use
btrfs_register_one_device()
[PATCH 1/2] btrfs-progs: introduce btrfs_register_all_device()
[PATCH 2/2] btrfs-progs: optimize btrfs_scan_lblkid() for multiple calls
If you could, pls..
Now on 3.17 apply above 3 patches and see if you see any better
performance for the stalling issue.

no perceptible change: takes ~40 seconds both before and after
applying. Old version <1 sec.

Post by Anand Jain
can you do same steps on 3.16 and report what you observe

So many rejects -- do you have older versions of these patches?

Thanks,

Petr
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Anand Jain

2014-10-23 08:13:25 UTC

Permalink

Some of the disks on my system were missing and I was able to hit
this issue.

----------------
Check tree block failed, want=12582912, have=0
read block failed check_tree_block
Couldn't read chunk root
warning devid 2 not found already
Check tree block failed, want=143360, have=0
read block failed check_tree_block
Couldn't read chunk root
warning, device 4 is missing
warning, device 3 is missing
warning, device 2 is missing
warning, device 1 is missing
----------------

Did a bisect and it leads to this following patch.

----------------
commit 915902c5002485fb13d27c4b699a73fb66cc0f09

btrfs-progs: fix device missing of btrfs fi show with seed devices
----------------

Also this patch stalls ~2sec in the cmd btrfs fi show, on my system
with 48 disks.

Also a simple test case hits some warnings...

----------------
mkfs.btrfs -draid1 -mraid1 /dev/sdb /dev/sdc
mount /dev/sdb /btrfs && fillfs /btrfs 100 && umount /btrfs
wipefs -a /dev/sdb
modprobe -r btrfs && modprobe btrfs
mount -o degraded /dev/sdb /btrfs
btrfs fi show
Label: none uuid: 9844cd05-1c8c-473e-a84b-bac95aab7bc9
Total devices 2 FS bytes used 1.59MiB
devid 2 size 967.87MiB used 104.75MiB path /dev/sdc
*** Some devices missing

warning, device 1 is missing
warning, device 1 is missing
warning devid 1 not found already
----------------

Post by Petr Janecek
Hello,

Post by Anand Jain
You have mentioned two issues when balance and fi show running
concurrently

my mail was a bit chaotic, but I get the stalls even on idle system.
Today I got
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
Ignoring transid failure
leaf parent key incorrect 1559973888000
from 'btrfs fi sh' while I was just copying something, no balance running.
[...]

no perceptible change: takes ~40 seconds both before and after
applying. Old version <1 sec.

Post by Anand Jain
can you do same steps on 3.16 and report what you observe

So many rejects -- do you have older versions of these patches?
Thanks,
Petr
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
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 ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Gui Hecheng

2014-10-23 08:52:25 UTC

Permalink

Post by Anand Jain
Some of the disks on my system were missing and I was able to hit
this issue.
----------------
Check tree block failed, want=12582912, have=0
read block failed check_tree_block
Couldn't read chunk root
warning devid 2 not found already
Check tree block failed, want=143360, have=0
read block failed check_tree_block
Couldn't read chunk root
warning, device 4 is missing
warning, device 3 is missing
warning, device 2 is missing
warning, device 1 is missing
----------------
Did a bisect and it leads to this following patch.
----------------
commit 915902c5002485fb13d27c4b699a73fb66cc0f09
btrfs-progs: fix device missing of btrfs fi show with seed devices
----------------
Also this patch stalls ~2sec in the cmd btrfs fi show, on my system
with 48 disks.
Also a simple test case hits some warnings...
----------------
mkfs.btrfs -draid1 -mraid1 /dev/sdb /dev/sdc
mount /dev/sdb /btrfs && fillfs /btrfs 100 && umount /btrfs
wipefs -a /dev/sdb
modprobe -r btrfs && modprobe btrfs
mount -o degraded /dev/sdb /btrfs
btrfs fi show
Label: none uuid: 9844cd05-1c8c-473e-a84b-bac95aab7bc9
Total devices 2 FS bytes used 1.59MiB
devid 2 size 967.87MiB used 104.75MiB path /dev/sdc
*** Some devices missing
warning, device 1 is missing
warning, device 1 is missing
warning devid 1 not found already
----------------

Hi Anand and Petr,

Oh, it seems that there are btrfs with missing devs that are bringing
troubles to the @open_ctree_... function.
This should be a missing case of the patch above which should only take
effects when seeding devices are present.
I will try my best to follow this case, suggestions are welcome, Thanks!

-Gui

Post by Anand Jain

Post by Petr Janecek
Hello,

Post by Anand Jain
You have mentioned two issues when balance and fi show running
concurrently

my mail was a bit chaotic, but I get the stalls even on idle system.
Today I got
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
Ignoring transid failure
leaf parent key incorrect 1559973888000
from 'btrfs fi sh' while I was just copying something, no balance running.
[...]

no perceptible change: takes ~40 seconds both before and after
applying. Old version <1 sec.

Post by Anand Jain
can you do same steps on 3.16 and report what you observe

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Petr Janecek

2014-10-23 13:23:57 UTC

Permalink

Hello Gui,

Post by Gui Hecheng
Oh, it seems that there are btrfs with missing devs that are bringing

what do you mean by "missing devs"? I have no degraded fs.

The time "btrfs fi sh" spends scanning disks of a filesystem seems to
be proportional to the amount of data stored on them: on a completely
idle system, of ~20s total time it spends 10s scanning each of /mnt/b
and /mnt/b0, and almost no time on /mnt/b3 (which is the biggest)

Filesystem Size Used Avail Use% Mounted on
/dev/sdm 5.5T 2.4T 2.1T 54% /mnt/b
/dev/sda 5.5T 2.5T 3.1T 45% /mnt/b0
/dev/sde 7.3T 90G 5.4T 2% /mnt/b3

Thanks,

Petr
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Anand Jain

2014-10-23 13:36:57 UTC

Permalink

there is no point in re-creating so many btrfs kernel's logic in user
space. its just unnecessary, when kernel is already doing it. use
some interface to get info from kernel after device is registered,
(not necessarily mounted). so progs can be as sleek as possible.
to me it started as just one more bug now we have fixed so many many.
It all needs one good interface for kernel which provides anything
anything from the kernel.

Post by Gui Hecheng

Hi Anand and Petr,
Oh, it seems that there are btrfs with missing devs that are bringing
This should be a missing case of the patch above which should only take
effects when seeding devices are present.
I will try my best to follow this case, suggestions are welcome, Thanks!
-Gui

Post by Anand Jain

Post by Petr Janecek
Hello,

Post by Anand Jain
You have mentioned two issues when balance and fi show running
concurrently

my mail was a bit chaotic, but I get the stalls even on idle system.
Today I got
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
parent transid verify failed on 1559973888000 wanted 1819 found 1821
Ignoring transid failure
leaf parent key incorrect 1559973888000
from 'btrfs fi sh' while I was just copying something, no balance running.
[...]

no perceptible change: takes ~40 seconds both before and after
applying. Old version <1 sec.

Post by Anand Jain
can you do same steps on 3.16 and report what you observe

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
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 ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html