Some xfstests failures on non-crc xfs with latest xfsprogs (v4.5.0-rc1)

Some xfstests failures on non-crc xfs with latest xfsprogs (v4.5.0-rc1)

[Eryu Guan]

[-- Attachment #1: Type: text/plain, Size: 3761 bytes --]

Hi,

I noticed some xfstests failures while testing v4 XFS with latest
v4.5.0-rc1 xfsprogs on v4.5-rc6 kernel, and all these failures are gone
if I use v3.2.4 xfsprogs, or if test on v5 XFS.

So either xfstests needs update or xfsprogs breaks something for non-crc
XFS. But I'm not sure which is which, so I post them out for broader
review.

(All tests are checked with 'MKFS_OPTIONS="-m crc=0" ./check <sometest>')

== 1. xfs/032 fsck failure (xfs_db check) on ppc64 host ==

xfs/032 fails in the same way as the following steps on ppc64 host:

 mkfs -t xfs  -m crc=0 -s size=32768 -b size=65536 -d size=1g -f /dev/sda5
 xfs_db -c check /dev/sda5

This prints

 bad magic number 0 for inode 1792
 ...
 bad magic number 0 for inode 2047
 root inode 1792 is not a directory
 block 0/7 type unknown not expected
 allocated inode 1792 has 0 link count
 allocated inode 1793 has 0 link count
 allocated inode 1794 has 0 link count


== 2. xfs/033 fsck failure ==

xfs_repair detects more nlinks error. And this happens after commit
"c2c5096 libxfs: update to 3.16 kernel code"

# diff -u tests/xfs/033.out /root/xfstests/results//xfs_4k/xfs/033.out.bad
--- tests/xfs/033.out   2015-07-16 17:28:15.996000000 +0800
+++ /root/xfstests/results//xfs_4k/xfs/033.out.bad      2016-03-04 18:54:39.461000000 +0800
@@ -66,6 +66,7 @@
         - traversal finished ...
         - moving disconnected inodes to lost+found ...
 Phase 7 - verify and correct link counts...
+resetting inode INO nlinks from 0 to 1
 done
 Corrupting rt summary inode - setting bits to 0
 Wrote X.XXKb (value 0x0)
@@ -96,6 +97,7 @@
         - traversal finished ...
         - moving disconnected inodes to lost+found ...
 Phase 7 - verify and correct link counts...
+resetting inode INO nlinks from 0 to 1
 done
 Corrupting root inode - setting bits to -1
 Wrote X.XXKb (value 0xffffffff)
@@ -160,6 +162,7 @@
         - traversal finished ...
         - moving disconnected inodes to lost+found ...
 Phase 7 - verify and correct link counts...
+resetting inode INO nlinks from 0 to 1
 done
 Corrupting rt summary inode - setting bits to -1
 Wrote X.XXKb (value 0xffffffff)
@@ -191,4 +194,5 @@
         - traversal finished ...
         - moving disconnected inodes to lost+found ...
 Phase 7 - verify and correct link counts...
+resetting inode INO nlinks from 0 to 1
 done


== 3. xfs/108 xfs/196 xfs/261 xfs/244 fsck failure. ==

These tests fail in a similiar way, and they are all quota related.
Note that xfs/244 only fails on ppc64, others fail on all arches.

*** xfs_repair -n output ***
Phase 1 - find and verify superblock...
Phase 2 - using internal log
        - zero log...
        - scan filesystem freespace and inode maps...
        - found root inode chunk
Phase 3 - for each AG...
        - scan (but don't clear) agi unlinked lists...
        - process known inodes and perform inode discovery...
        - agno = 0
        - agno = 1
        - agno = 2
        - agno = 3
        - process newly discovered inodes...
Phase 4 - check for duplicate blocks...
        - setting up duplicate extent list...
        - check for inodes claiming duplicate blocks...
        - agno = 0
        - agno = 3
        - agno = 1
        - agno = 2
No modify flag set, skipping phase 5
Phase 6 - check inode connectivity...
        - traversing filesystem ...
        - traversal finished ...
        - moving disconnected inodes to lost+found ...
disconnected inode 131, would move to lost+found
Phase 7 - verify link counts...
No modify flag set, skipping filesystem flush and exiting.
*** end xfs_repair outpu

== 4. xfs/186 xfs/187 attr test failure ==

xfs/186 is missing almost all the attr values.
xfs/187 reports test needs update.


Please see attachments for details.

Thanks,
Eryu

[-- Attachment #2: xfs-032-ppc64.full.bz2 --]
[-- Type: application/x-bzip2, Size: 4463 bytes --]

[-- Attachment #3: xfs-033.out.bad.bz2 --]
[-- Type: application/x-bzip2, Size: 1160 bytes --]

[-- Attachment #4: xfs-108.full.bz2 --]
[-- Type: application/x-bzip2, Size: 1580 bytes --]

[-- Attachment #5: xfs-186.out.bad.bz2 --]
[-- Type: application/x-bzip2, Size: 318 bytes --]

[-- Attachment #6: xfs-187.out.bad.bz2 --]
[-- Type: application/x-bzip2, Size: 198 bytes --]

[-- Attachment #7: xfs-196.full.bz2 --]
[-- Type: application/x-bzip2, Size: 1229 bytes --]

[-- Attachment #8: xfs-244-ppc64.full.bz2 --]
[-- Type: application/x-bzip2, Size: 1862 bytes --]

[-- Attachment #9: xfs-261.full.bz2 --]
[-- Type: application/x-bzip2, Size: 1183 bytes --]

[-- Attachment #10: Type: text/plain, Size: 121 bytes --]

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: Some xfstests failures on non-crc xfs with latest xfsprogs (v4.5.0-rc1)

[Eric Sandeen]

On 3/4/16 5:23 AM, Eryu Guan wrote:
> Hi,
> 
> I noticed some xfstests failures while testing v4 XFS with latest
> v4.5.0-rc1 xfsprogs on v4.5-rc6 kernel, and all these failures are gone
> if I use v3.2.4 xfsprogs, or if test on v5 XFS.
> 
> So either xfstests needs update or xfsprogs breaks something for non-crc
> XFS. But I'm not sure which is which, so I post them out for broader
> review.
> 
> (All tests are checked with 'MKFS_OPTIONS="-m crc=0" ./check <sometest>')

Eryu, sorry you didn't get a reply on this in March.  We should have
listened!  ;)  The patches I sent in the pat day or two should
fix some of these ...

> == 1. xfs/032 fsck failure (xfs_db check) on ppc64 host ==

but not this one.  Do you still see this?

> == 2. xfs/033 fsck failure ==
> 
> xfs_repair detects more nlinks error. And this happens after commit
> "c2c5096 libxfs: update to 3.16 kernel code"

this should be fixed by
[PATCH] xfs_repair: set rsumino version to 2
 
> == 3. xfs/108 xfs/196 xfs/261 xfs/244 fsck failure. ==

this should be fixed by
[PATCH] xfs_repair: don't call xfs_sb_quota_from_disk twice

although I'm not sure about the ppc-specific failure.
 
> These tests fail in a similiar way, and they are all quota related.
> Note that xfs/244 only fails on ppc64, others fail on all arches.

> == 4. xfs/186 xfs/187 attr test failure ==
> 
> xfs/186 is missing almost all the attr values.

[PATCH] xfs/186: fix test for crc=0, ftype=1

> xfs/187 reports test needs update.

Not sure on this one yet.

-Eric

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: Some xfstests failures on non-crc xfs with latest xfsprogs (v4.5.0-rc1)

[Eryu Guan]

On Fri, Jun 24, 2016 at 05:37:10PM -0500, Eric Sandeen wrote:
> On 3/4/16 5:23 AM, Eryu Guan wrote:
> > Hi,
> > 
> > I noticed some xfstests failures while testing v4 XFS with latest
> > v4.5.0-rc1 xfsprogs on v4.5-rc6 kernel, and all these failures are gone
> > if I use v3.2.4 xfsprogs, or if test on v5 XFS.
> > 
> > So either xfstests needs update or xfsprogs breaks something for non-crc
> > XFS. But I'm not sure which is which, so I post them out for broader
> > review.
> > 
> > (All tests are checked with 'MKFS_OPTIONS="-m crc=0" ./check <sometest>')
> 
> Eryu, sorry you didn't get a reply on this in March.  We should have
> listened!  ;)  The patches I sent in the pat day or two should

Thanks for looking at them! :)

> fix some of these ...
> 
> > == 1. xfs/032 fsck failure (xfs_db check) on ppc64 host ==
> 
> but not this one.  Do you still see this?

My recent test with 4.7-rc4 kernel and latest xfsprogs (June 22nd) shows
xfs/032 passed on ppc64 host. But I'll retest to be sure.

> 
> > == 2. xfs/033 fsck failure ==
> > 
> > xfs_repair detects more nlinks error. And this happens after commit
> > "c2c5096 libxfs: update to 3.16 kernel code"
> 
> this should be fixed by
> [PATCH] xfs_repair: set rsumino version to 2
>  
> > == 3. xfs/108 xfs/196 xfs/261 xfs/244 fsck failure. ==
> 
> this should be fixed by
> [PATCH] xfs_repair: don't call xfs_sb_quota_from_disk twice
> 
> although I'm not sure about the ppc-specific failure.

My recent test on RHEL7 shows xfs/244 is not ppc-specific. I'll retest
with latest kernel and xfsprogs on ppc64 and report back.

>  
> > These tests fail in a similiar way, and they are all quota related.
> > Note that xfs/244 only fails on ppc64, others fail on all arches.
> 
> > == 4. xfs/186 xfs/187 attr test failure ==
> > 
> > xfs/186 is missing almost all the attr values.
> 
> [PATCH] xfs/186: fix test for crc=0, ftype=1

Thanks! I've queued it in fstests next pull request.

Thanks,
Eryu

> 
> > xfs/187 reports test needs update.
> 
> Not sure on this one yet.
> 
> -Eric
> 
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: Some xfstests failures on non-crc xfs with latest xfsprogs (v4.5.0-rc1)

[Eric Sandeen]

> On Jun 24, 2016, at 11:44 PM, Eryu Guan <eguan@redhat.com> wrote:
> 
>> On Fri, Jun 24, 2016 at 05:37:10PM -0500, Eric Sandeen wrote:
>>> On 3/4/16 5:23 AM, Eryu Guan wrote:
>>> Hi,
>>> 
>>> I noticed some xfstests failures while testing v4 XFS with latest
>>> v4.5.0-rc1 xfsprogs on v4.5-rc6 kernel, and all these failures are gone
>>> if I use v3.2.4 xfsprogs, or if test on v5 XFS.
>>> 
>>> So either xfstests needs update or xfsprogs breaks something for non-crc
>>> XFS. But I'm not sure which is which, so I post them out for broader
>>> review.
>>> 
>>> (All tests are checked with 'MKFS_OPTIONS="-m crc=0" ./check <sometest>')
>> 
>> Eryu, sorry you didn't get a reply on this in March.  We should have
>> listened!  ;)  The patches I sent in the pat day or two should
> 
> Thanks for looking at them! :)
> 
>> fix some of these ...
>> 
>>> == 1. xfs/032 fsck failure (xfs_db check) on ppc64 host ==
>> 
>> but not this one.  Do you still see this?
> 
> My recent test with 4.7-rc4 kernel and latest xfsprogs (June 22nd) shows
> xfs/032 passed on ppc64 host. But I'll retest to be sure.

I wasn't thinking.  Brian's recent sparse inode patches fix this one.

Eric

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: Some xfstests failures on non-crc xfs with latest xfsprogs (v4.5.0-rc1)

[Eryu Guan]

On Sat, Jun 25, 2016 at 12:21:26AM -0500, Eric Sandeen wrote:
> 
> > On Jun 24, 2016, at 11:44 PM, Eryu Guan <eguan@redhat.com> wrote:
> > 
> >> On Fri, Jun 24, 2016 at 05:37:10PM -0500, Eric Sandeen wrote:
> >>> On 3/4/16 5:23 AM, Eryu Guan wrote:
> >>> Hi,
> >>> 
> >>> I noticed some xfstests failures while testing v4 XFS with latest
> >>> v4.5.0-rc1 xfsprogs on v4.5-rc6 kernel, and all these failures are gone
> >>> if I use v3.2.4 xfsprogs, or if test on v5 XFS.
> >>> 
> >>> So either xfstests needs update or xfsprogs breaks something for non-crc
> >>> XFS. But I'm not sure which is which, so I post them out for broader
> >>> review.
> >>> 
> >>> (All tests are checked with 'MKFS_OPTIONS="-m crc=0" ./check <sometest>')
> >> 
> >> Eryu, sorry you didn't get a reply on this in March.  We should have
> >> listened!  ;)  The patches I sent in the pat day or two should
> > 
> > Thanks for looking at them! :)
> > 
> >> fix some of these ...
> >> 
> >>> == 1. xfs/032 fsck failure (xfs_db check) on ppc64 host ==
> >> 
> >> but not this one.  Do you still see this?
> > 
> > My recent test with 4.7-rc4 kernel and latest xfsprogs (June 22nd) shows
> > xfs/032 passed on ppc64 host. But I'll retest to be sure.
> 
> I wasn't thinking.  Brian's recent sparse inode patches fix this one.

Confirmed xfsprogs v4.7-rc1 passed the test.

And xfs/244 is not ppc-specific, x86_64 fails it either. I think your
patch could fix it as well.

Thanks,
Eryu

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs