Re: Metadata_csum is still enabled after trying to disable it.

Re: Metadata_csum is still enabled after trying to disable it.

[Jean-Christophe Guillain]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1128 bytes --]


​​​----- On Nov 22, 2018, at 4:01 AM, Darrick J. Wong darrick.wong@oracle.com wrote:
>> * I check : metadata checksum looks like disabled
>>     root@test# tune2fs -l /dev/my_vol tune2fs 1.43.4 (31-Jan-2017)
>>     Filesystem volume name:   <none>
>>     [...]
>>     Journal backup:           inode blocks
>> 
>> * I change filesystem UUID : It looks like metada checksum still is enabled
>>     root@test# tune2fs -U random /dev/my_vol
>>     tune2fs 1.43.4 (31-Jan-2017)
>>     Setting UUID on a checksummed filesystem could take some time.
>>     Proceed anyway (or wait 5 seconds) ? (y,N) y
>> 
>> Is it a bug ?
> 
> Yes.

Should I open a bug in bugzilla ?

> 
>> How can I actually disable this parameter ?
> 
> I think you have to turn off the metadata_csum_seed feature, but tbh
> tune2fs really shouldn't pester about !metadata_csum,metadata_csum_seed
> filesystems...
> 

I actually tried that (!metadata_csum,!metadata_csum_seed) : tune2fs has still the same behaviour.
How does it detect that the filesystem is (it shouldn't be anymore), or was checksummed ?

> --D
> 
>> Best regards,
>> jC Guillain
>> 
>> 

[-- Attachment #2: Type: text/plain, Size: 335 bytes --]



-- IMPORTANT NOTICE: 

The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

Re: Metadata_csum is still enabled after trying to disable it.

[Jean-Christophe Guillain]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 695 bytes --]


----- On Nov 24, 2018, at 4:48 AM, Theodore Y. Ts'o tytso@mit.edu wrote:

​​​​​> On Wed, Nov 21, 2018 at 07:01:19PM -0800, Darrick J. Wong wrote:
>> > Is it a bug ?
>> 
>> Yes.
>> 
>> > How can I actually disable this parameter ?
>>
> 
> Actually, the metadata_csum feature *was* disabled.  You would have
> seen that if you looked at the "Features: " line printed by dumpe2fs
> -h or tune2fs -l.  The bug was that we were erroneously warning that
> the tune2fs -U operation would take a long time when flex_bg was
> enabled, but metadata_csum was not.  So it was misleading wrong, but
> it actually turning off the metadata_csum feature.
> 

Thank you very much for your time !

jC


[-- Attachment #2: Type: text/plain, Size: 335 bytes --]



-- IMPORTANT NOTICE: 

The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

Re: Metadata_csum is still enabled after trying to disable it.

[Theodore Y. Ts'o]

On Wed, Nov 21, 2018 at 07:01:19PM -0800, Darrick J. Wong wrote:
> > Is it a bug ?
> 
> Yes.
> 
> > How can I actually disable this parameter ?
>

Actually, the metadata_csum feature *was* disabled.  You would have
seen that if you looked at the "Features: " line printed by dumpe2fs
-h or tune2fs -l.  The bug was that we were erroneously warning that
the tune2fs -U operation would take a long time when flex_bg was
enabled, but metadata_csum was not.  So it was misleading wrong, but
it actually turning off the metadata_csum feature.

Cheers,

					- Ted

From 0eeb17d0610fd512c1038dabb39370420ffb47a4 Mon Sep 17 00:00:00 2001
From: Theodore Ts'o <tytso@mit.edu>
Date: Fri, 23 Nov 2018 22:34:31 -0500
Subject: [PATCH] tune2fs: fix false warning that a UUID change will take a
 long time

If the file system only has the flex_bg feature enabled (with out the
metadata_csum feature enabled), it won't take a long time time fix up
the checksums after changing the UUID.  While it does need to
recalculate all of the checksums in the block group descriptors, that
doesn't take a long time.

Also, if the ea_data feature is enabled, changing the UUID will also
take a long time, and we weren't warning the user about that case.

Fix up the warning message so it doesn't mislead people, and is more
accurate.

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
---
 misc/tune2fs.c           | 17 +++++++++--------
 tests/t_dangerous/expect | 29 ++++++++++++++++++++++++-----
 tests/t_dangerous/script | 33 +++++++++++++++++++++++++++------
 3 files changed, 60 insertions(+), 19 deletions(-)

diff --git a/misc/tune2fs.c b/misc/tune2fs.c
index ec977b8c3..7c5ba0c77 100644
--- a/misc/tune2fs.c
+++ b/misc/tune2fs.c
@@ -3221,6 +3221,15 @@ _("Warning: The journal is dirty. You may wish to replay the journal like:\n\n"
 		char buf[SUPERBLOCK_SIZE] __attribute__ ((aligned(8)));
 		__u8 old_uuid[UUID_SIZE];
 
+		if (!ext2fs_has_feature_csum_seed(fs->super) &&
+		    (ext2fs_has_feature_metadata_csum(fs->super) ||
+		     ext2fs_has_feature_ea_inode(fs->super))) {
+			check_fsck_needed(fs,
+				_("Setting the UUID on this "
+				  "filesystem could take some time."));
+			rewrite_checksums = 1;
+		}
+
 		if (ext2fs_has_group_desc_csum(fs)) {
 			/*
 			 * Changing the UUID on a metadata_csum FS requires
@@ -3241,10 +3250,6 @@ _("Warning: The journal is dirty. You may wish to replay the journal like:\n\n"
 				try_confirm_csum_seed_support();
 				exit(1);
 			}
-			if (!ext2fs_has_feature_csum_seed(fs->super))
-				check_fsck_needed(fs,
-					_("Setting UUID on a checksummed "
-					  "filesystem could take some time."));
 
 			/*
 			 * Determine if the block group checksums are
@@ -3302,10 +3307,6 @@ _("Warning: The journal is dirty. You may wish to replay the journal like:\n\n"
 		}
 
 		ext2fs_mark_super_dirty(fs);
-		if (!ext2fs_has_feature_csum_seed(fs->super) &&
-		    (ext2fs_has_feature_metadata_csum(fs->super) ||
-		     ext2fs_has_feature_ea_inode(fs->super)))
-			rewrite_checksums = 1;
 	}
 
 	if (I_flag) {
diff --git a/tests/t_dangerous/expect b/tests/t_dangerous/expect
index a9903b750..31aaf4fad 100644
--- a/tests/t_dangerous/expect
+++ b/tests/t_dangerous/expect
@@ -43,10 +43,6 @@ tune2fs -I 512 test.img
 Resizing inodes could take some time.
 Proceed anyway (or wait 5 seconds to proceed) ? (y,N) 
 Exit status is 1
-tune2fs -U random test.img
-Setting UUID on a checksummed filesystem could take some time.
-Proceed anyway (or wait 5 seconds to proceed) ? (y,N) 
-Exit status is 1
 
 Change in FS metadata:
 Pass 1: Checking inodes, blocks, and sizes
@@ -79,7 +75,7 @@ Resizing inodes could take some time.
 Proceed anyway (or wait 5 seconds to proceed) ? (y,N) Setting inode size 512
 Exit status is 0
 tune2fs -U f0f0f0f0-f0f0-f0f0-f0f0-f0f0f0f0f0f0 test.img
-Setting UUID on a checksummed filesystem could take some time.
+Setting the UUID on this filesystem could take some time.
 Proceed anyway (or wait 5 seconds to proceed) ? (y,N) Exit status is 0
 Backing up journal inode block information.
 
@@ -99,3 +95,26 @@ Pass 4: Checking reference counts
 Pass 5: Checking group summary information
 
 Exit status is 0
+ 
+Testing with metadata checksum enabled
+Creating filesystem with 524288 1k blocks and 65536 inodes
+Superblock backups stored on blocks: 
+	8193, 24577, 40961, 57345, 73729, 204801, 221185, 401409
+
+Allocating group tables:      \b\b\b\b\bdone                            
+Writing inode tables:      \b\b\b\b\bdone                            
+Creating journal (16384 blocks): done
+Creating 445 huge file(s) with 1024 blocks each: done
+Writing superblocks and filesystem accounting information:      \b\b\b\b\bdone
+
+Pass 1: Checking inodes, blocks, and sizes
+Pass 2: Checking directory structure
+Pass 3: Checking directory connectivity
+Pass 4: Checking reference counts
+Pass 5: Checking group summary information
+
+Exit status is 0
+tune2fs -U random test.img
+Setting the UUID on this filesystem could take some time.
+Proceed anyway (or wait 5 seconds to proceed) ? (y,N) 
+Exit status is 1
diff --git a/tests/t_dangerous/script b/tests/t_dangerous/script
index b71421881..f684d56d3 100644
--- a/tests/t_dangerous/script
+++ b/tests/t_dangerous/script
@@ -17,6 +17,18 @@ cat > $CONF << ENDL
 		hugefiles_size = 1M
 		zero_hugefiles = false
 	}
+	ext4m = {
+		features = has_journal,extent,huge_file,^flex_bg,uninit_bg,dir_nlink,extra_isize,sparse_super,filetype,dir_index,ext_attr,resize_inode,64bit,metadata_csum
+		blocksize = 1024
+		inode_size = 256
+		make_hugefiles = true
+		hugefiles_dir = /
+		hugefiles_slack = 32M
+		hugefiles_name = aaaaa
+		hugefiles_digits = 4
+		hugefiles_size = 1M
+		zero_hugefiles = false
+	}
 ENDL
 
 echo "tune2fs dangerous prompts test" > $OUT
@@ -62,12 +74,6 @@ echo 'n' | TUNE2FS_FORCE_PROMPT=1 $TUNE2FS -I 512 $TMPFILE >> $OUT 2>&1
 status=$?
 echo Exit status is $status >> $OUT
 
-# change uuid
-echo "tune2fs -U random test.img" >> $OUT
-echo 'n' | TUNE2FS_FORCE_PROMPT=1 $TUNE2FS -U random $TMPFILE >> $OUT 2>&1
-status=$?
-echo Exit status is $status >> $OUT
-
 # check
 $FSCK -yD -N test_filesys $TMPFILE >> $OUT 2>&1
 
@@ -111,6 +117,21 @@ $FSCK $FSCK_OPT -N test_filesys $TMPFILE >> $OUT 2>&1
 status=$?
 echo Exit status is $status >> $OUT
 
+echo " " >> $OUT
+echo "Testing with metadata checksum enabled" >> $OUT
+
+MKE2FS_CONFIG=$CONF $MKE2FS -F -T ext4m -U 6fc3daa4-180d-4f2b-a6f2-f7a5efb79bcf $TMPFILE 524288 >> $OUT 2>&1
+$DUMPE2FS $TMPFILE 2> /dev/null | grep '^Group 0:' -B99 -A20 | sed -f $cmd_dir/filter.sed > $OUT.before
+$FSCK $FSCK_OPT -f -N test_filesys $TMPFILE >> $OUT 2>&1
+status=$?
+echo Exit status is $status >> $OUT
+
+# change uuid
+echo "tune2fs -U random test.img" >> $OUT
+echo 'n' | TUNE2FS_FORCE_PROMPT=1 $TUNE2FS -U random $TMPFILE >> $OUT 2>&1
+status=$?
+echo Exit status is $status >> $OUT
+
 rm $TMPFILE $OUT.before $OUT.after $CONF
 
 #
-- 
2.19.1

Re: Metadata_csum is still enabled after trying to disable it.

[Darrick J. Wong]

On Wed, Nov 21, 2018 at 06:48:18PM +0100, Jean-Christophe Guillain wrote:
> Hi,
> 
> I see a strange behaviour when playing with the metadata_csum parameter of ext4.
> (I use a Debian Stretch, with a kernel 4.9.0-8-amd64 #1 SMP Debian 4.9.110-3+deb9u6 (2018-10-08) x86_64 GNU/Linux)
> 
> How to reproduce :
> 
> * I enable metada checksum : OK
>     root@test# tune2fs -O metadata_csum /dev/my_vol
>     tune2fs 1.43.4 (31-Jan-2017)
>     Enabling checksums could take some time.
>     Proceed anyway (or wait 5 seconds) ? (y,N) y
> 
> * I check : Checksum enabled
>     root@test# tune2fs -l /dev/my_vol
>     tune2fs 1.43.4 (31-Jan-2017)
>     Filesystem volume name:   <none>
>     [...]
>     Journal backup:           inode blocks
>     Checksum type:            crc32c
>     Checksum:                 0xb0e3a180
> 
> * I change filesystem UUID : OK
>     root@test# tune2fs -U random /dev/my_vol
>     tune2fs 1.43.4 (31-Jan-2017)
>     Setting UUID on a checksummed filesystem could take some time.
>     Proceed anyway (or wait 5 seconds) ? (y,N) y
> 
> * I disable metadata checksum : OK
>     root@test# tune2fs -O ^metadata_csum /dev/my_vol
>     tune2fs 1.43.4 (31-Jan-2017)
>     Disabling checksums could take some time.
>     Proceed anyway (or wait 5 seconds) ? (y,N) y
> 
> * I change filesystem UUID : It looks like metada checksum still is enabled
>     root@test# tune2fs -U random /dev/my_vol
>     tune2fs 1.43.4 (31-Jan-2017)
>     Setting UUID on a checksummed filesystem could take some time.
>     Proceed anyway (or wait 5 seconds) ? (y,N) y
> 
> * I fsck the filesystem to see if it could fix this.
>     root@test# e2fsck -f /dev/my_vol
>     2fsck 1.43.4 (31-Jan-2017)
>     Pass 1: Checking inodes, blocks, and sizes
>     Pass 2: Checking directory structure
>     Pass 3: Checking directory connectivity
>     Pass 4: Checking reference counts
>     Pass 5: Checking group summary information
>     /dev/my_vol: 16303/3932160 files (0.9% non-contiguous), 1829950/15728640 blocks
> 
> * I check : metadata checksum looks like disabled
>     root@test# tune2fs -l /dev/my_vol tune2fs 1.43.4 (31-Jan-2017)
>     Filesystem volume name:   <none>
>     [...]
>     Journal backup:           inode blocks
> 
> * I change filesystem UUID : It looks like metada checksum still is enabled
>     root@test# tune2fs -U random /dev/my_vol
>     tune2fs 1.43.4 (31-Jan-2017)
>     Setting UUID on a checksummed filesystem could take some time.
>     Proceed anyway (or wait 5 seconds) ? (y,N) y
> 
> Is it a bug ?

Yes.

> How can I actually disable this parameter ?

I think you have to turn off the metadata_csum_seed feature, but tbh
tune2fs really shouldn't pester about !metadata_csum,metadata_csum_seed
filesystems...

--D

> Best regards,
> jC Guillain
> 
> 
> -- IMPORTANT NOTICE: 
> 
> The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
> 

Metadata_csum is still enabled after trying to disable it.

[Jean-Christophe Guillain]

Hi,

I see a strange behaviour when playing with the metadata_csum parameter of ext4.
(I use a Debian Stretch, with a kernel 4.9.0-8-amd64 #1 SMP Debian 4.9.110-3+deb9u6 (2018-10-08) x86_64 GNU/Linux)

How to reproduce :

* I enable metada checksum : OK
    root@test# tune2fs -O metadata_csum /dev/my_vol
    tune2fs 1.43.4 (31-Jan-2017)
    Enabling checksums could take some time.
    Proceed anyway (or wait 5 seconds) ? (y,N) y

* I check : Checksum enabled
    root@test# tune2fs -l /dev/my_vol
    tune2fs 1.43.4 (31-Jan-2017)
    Filesystem volume name:   <none>
    [...]
    Journal backup:           inode blocks
    Checksum type:            crc32c
    Checksum:                 0xb0e3a180

* I change filesystem UUID : OK
    root@test# tune2fs -U random /dev/my_vol
    tune2fs 1.43.4 (31-Jan-2017)
    Setting UUID on a checksummed filesystem could take some time.
    Proceed anyway (or wait 5 seconds) ? (y,N) y

* I disable metadata checksum : OK
    root@test# tune2fs -O ^metadata_csum /dev/my_vol
    tune2fs 1.43.4 (31-Jan-2017)
    Disabling checksums could take some time.
    Proceed anyway (or wait 5 seconds) ? (y,N) y

* I change filesystem UUID : It looks like metada checksum still is enabled
    root@test# tune2fs -U random /dev/my_vol
    tune2fs 1.43.4 (31-Jan-2017)
    Setting UUID on a checksummed filesystem could take some time.
    Proceed anyway (or wait 5 seconds) ? (y,N) y

* I fsck the filesystem to see if it could fix this.
    root@test# e2fsck -f /dev/my_vol
    2fsck 1.43.4 (31-Jan-2017)
    Pass 1: Checking inodes, blocks, and sizes
    Pass 2: Checking directory structure
    Pass 3: Checking directory connectivity
    Pass 4: Checking reference counts
    Pass 5: Checking group summary information
    /dev/my_vol: 16303/3932160 files (0.9% non-contiguous), 1829950/15728640 blocks

* I check : metadata checksum looks like disabled
    root@test# tune2fs -l /dev/my_vol tune2fs 1.43.4 (31-Jan-2017)
    Filesystem volume name:   <none>
    [...]
    Journal backup:           inode blocks

* I change filesystem UUID : It looks like metada checksum still is enabled
    root@test# tune2fs -U random /dev/my_vol
    tune2fs 1.43.4 (31-Jan-2017)
    Setting UUID on a checksummed filesystem could take some time.
    Proceed anyway (or wait 5 seconds) ? (y,N) y

Is it a bug ?
How can I actually disable this parameter ?

Best regards,
jC Guillain


-- IMPORTANT NOTICE: 

The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.