BTRFS: Testing restore of file compressed with lzo falied

Bug #1559836 reported by bugproxy on 2016-03-21
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
btrfs-tools (Ubuntu)
Medium
Unassigned

Bug Description

Problem Description
===================================
Using xfstests, testing btrfs/041 testcase which does btrfs-restore operation .

Analysis : This testcase create extents, write data to it, compress it then restore.But it is failing with lzo compression.

root@powerkvm5:~/xfstests# ./check btrfs/041
FSTYP -- btrfs
PLATFORM -- Linux/ppc64le powerkvm5 4.4.0-13-generic
MKFS_OPTIONS -- /dev/sdg2
MOUNT_OPTIONS -- /dev/sdg2 /mnt/scratch

btrfs/041 [failed, exit status 1] - output mismatch (see /root/xfstests/results//btrfs/041.out.bad)
    --- tests/btrfs/041.out 2016-03-15 01:28:12.782520209 -0500
    +++ /root/xfstests/results//btrfs/041.out.bad 2016-03-18 02:47:17.598581773 -0500
    @@ -11,30 +11,5 @@
     wrote 100/100 bytes at offset 99000
     XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
     67edd038aaa42adb5a1aa78f2eb1d2b6 SCRATCH_MNT/foo
    -67edd038aaa42adb5a1aa78f2eb1d2b6
    -Testing restore of file compressed with zlib
    -wrote 100000/100000 bytes at offset 0
    -XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
    ...
    (Run 'diff -u tests/btrfs/041.out /root/xfstests/results//btrfs/041.out.bad' to see the entire diff)
Ran: btrfs/041
Failures: btrfs/041
Failed 1 of 1 tests

System : powerkvm5, Bare metal
OS : Ubuntu 16.04
Version : 4.4.0-13-generic

root@powerkvm5:~/xfstests# cat /root/xfstests/results//btrfs/041.out.bad
QA output created by 041
Testing restore of file compressed with lzo
wrote 100000/100000 bytes at offset 0
XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
wrote 100000/100000 bytes at offset 100000
XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
wrote 2/2 bytes at offset 10000
XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
wrote 11/11 bytes at offset 33000
XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
wrote 100/100 bytes at offset 99000
XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
67edd038aaa42adb5a1aa78f2eb1d2b6 SCRATCH_MNT/foo
failed: '/bin/btrfs restore /dev/sdg2 /mnt/test/btrfs-test-041'
(see /root/xfstests/results//btrfs/041.full for details)
root@powerkvm5:~/xfstests# vi tests/btrfs/041
root@powerkvm5:~/xfstests# uname -a
Linux powerkvm5 4.4.0-13-generic #29-Ubuntu SMP Fri Mar 11 19:30:44 UTC 2016 ppc64le ppc64le ppc64le GNU/Linux

Steps to reproduce :
======================================
1. Install xfstests.
2. Create 5 file images of 10gb each, using dd command
3. Create loop devices using above images
4. Create btrfs fs on these loop devices
5. test setup:
1. suites has already built with 2 directories : xfsprogs & xfstests
2. create /mnt/test and /mnt/scratch directories
3. cd xfstests and edit local.config file
sample file for btrfs test :

export TEST_DEV=/dev/loop0
export TEST_DIR=/mnt/test
export SCRATCH_DEV_POOL="/dev/loop1 /dev/loop2 /dev/loop3 /dev/loop4"
export SCRATCH_MNT=/mnt/scratch

6. create fsgqa user : 'useradd fsgqa'
7. ./check btrfs/041

== Comment: #2 - Chandan B. Rajendra <email address hidden> - 2016-03-18 09:07:07 ==
The issue was recently fixed by Fiona. Please refer to the commit b025392ffca98b3a74947cf8f772535cf8bfceb3 in the git repository available at git://repo.or.cz/btrfs-progs-unstable/devel.git.

b025392ffca98b3a74947cf8f772535cf8bfceb3
Author: Feifei Xu <email address hidden>
AuthorDate: Thu Mar 10 16:43:20 2016 +0800
Commit: David Sterba <email address hidden>
CommitDate: Mon Mar 14 13:42:47 2016 +0100

btrfs-progs: Replace hardcoded PAGE_CACHE_SIZE with sectorsize.

PAGE_CACHE_SIZE is hardcoded to 4K in cmds-restore.c. Correct
value should be sector size. Fix this through replacing
hardcoded 4K to sectorsize.

Signed-off-by: Feifei Xu <email address hidden>
Signed-off-by: David Sterba <email address hidden>

We should mirror this bug to Canonical and request them to apply the commit to their btrfs-progs package.

bugproxy (bugproxy) on 2016-03-21
tags: added: architecture-ppc64le bugnameltc-139295 severity-high targetmilestone-inin---
Changed in ubuntu:
assignee: nobody → Taco Screen team (taco-screen-team)
vaishnavi (vaishnavi) on 2016-03-21
affects: ubuntu → btrfs (Ubuntu)
Steve Langasek (vorlon) wrote :

Accepting this as a medium-priority bug for Ubuntu. We do not use btrfs as the default filesystem in Ubuntu for reasons far beyond a single backup/restore test failure; we do not commit to fixing this bug for 16.04 release.

affects: btrfs (Ubuntu) → btrfs-tools (Ubuntu)
Changed in btrfs-tools (Ubuntu):
importance: Undecided → Medium
status: New → Triaged
bugproxy (bugproxy) on 2016-06-13
tags: added: targetmilestone-inin16041
removed: targetmilestone-inin---

------- Comment From <email address hidden> 2016-07-07 02:40 EDT-------
== comment from Pavithra ==
Hi,

Issue could be reproduced in 16.04.01 build.

ubuntu@tuleta4u-lp14:~$ uname -a
Linux tuleta4u-lp14 4.4.0-24-generic #43-Ubuntu SMP Wed Jun 8 19:25:36 UTC 2016 ppc64le ppc64le ppc64le GNU/Linux

Log
===
root@tuleta4u-lp14:/home/ubuntu/xfstests# ./check btrfs/041
FSTYP -- btrfs
PLATFORM -- Linux/ppc64le tuleta4u-lp14 4.4.0-24-generic
MKFS_OPTIONS -- /dev/loop1
MOUNT_OPTIONS -- /dev/loop1 /mnt/scratch

btrfs/041 [failed, exit status 1] - output mismatch (see /home/ubuntu/xfstests/results//btrfs/041.out.bad)
--- tests/btrfs/041.out 2016-06-15 07:26:04.347320804 -0400
+++ /home/ubuntu/xfstests/results//btrfs/041.out.bad 2016-06-15 08:00:41.948122839 -0400
@@ -11,30 +11,5 @@
wrote 100/100 bytes at offset 99000
XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
67edd038aaa42adb5a1aa78f2eb1d2b6 SCRATCH_MNT/foo
-67edd038aaa42adb5a1aa78f2eb1d2b6
-Testing restore of file compressed with zlib
-wrote 100000/100000 bytes at offset 0
-XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
...
(Run 'diff -u tests/btrfs/041.out /home/ubuntu/xfstests/results//btrfs/041.out.bad' to see the entire diff)
Ran: btrfs/041
Failures: btrfs/041
Failed 1 of 1 tests

Thanks,
Pavithra

Andrew Cloke (andrew-cloke) wrote :

Taco-screen-team still assigned, although as per comment #1, the bug has been screened, assessed and triaged.

Changed in btrfs-tools (Ubuntu):
assignee: Taco Screen team (taco-screen-team) → nobody
bugproxy (bugproxy) wrote :

------- Comment From <email address hidden> 2017-02-17 04:36 EDT-------
Hi Canonical,

Can you please include the commit b025392ffca98b3a74947cf8f772535cf8bfceb3 that fixes this issue ?

Thank you.

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers