diff mbox

[06/37] debugfs: force logdump to display (old) journal contents

Message ID 20140501231302.31890.45397.stgit@birch.djwong.org
State Superseded, archived
Headers show

Commit Message

Darrick Wong May 1, 2014, 11:13 p.m. UTC 
If the user passes -a more than once to logdump, try to dump old log
contents.  This can be used to try to track down journal problems even
after recovery.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 debugfs/logdump.c |   11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)



--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Lukas Czerner May 2, 2014, 11:49 a.m. UTC | #1 
On Thu, 1 May 2014, Darrick J. Wong wrote:

> Date: Thu, 01 May 2014 16:13:02 -0700
> From: Darrick J. Wong <darrick.wong@oracle.com>
> To: tytso@mit.edu, darrick.wong@oracle.com
> Cc: linux-ext4@vger.kernel.org
> Subject: [PATCH 06/37] debugfs: force logdump to display (old) journal
>     contents
> 
> If the user passes -a more than once to logdump, try to dump old log
> contents.  This can be used to try to track down journal problems even
> after recovery.

You need to update man page as well for this. Also I wonder what's
the behaviour if '-a' and '-b' or '-c' are specified simultaneously
and '-a' is specified multiple times ?

Thanks!
-Lukas

> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>  debugfs/logdump.c |   11 ++++++++---
>  1 file changed, 8 insertions(+), 3 deletions(-)
> 
> 
> diff --git a/debugfs/logdump.c b/debugfs/logdump.c
> index 8b9dc5b..bf4bef5 100644
> --- a/debugfs/logdump.c
> +++ b/debugfs/logdump.c
> @@ -393,9 +393,13 @@ static void dump_journal(char *cmdname, FILE *out_file,
>  	fprintf(out_file, "Journal starts at block %u, transaction %u\n",
>  		blocknr, transaction);
>  
> -	if (!blocknr)
> +	if (!blocknr) {
>  		/* Empty journal, nothing to do. */
> -		return;
> +		if (dump_all < 2)
> +			return;
> +		else
> +			blocknr = 1;
> +	}
>  
>  	while (1) {
>  		retval = read_journal_block(cmdname, source,
> @@ -420,7 +424,8 @@ static void dump_journal(char *cmdname, FILE *out_file,
>  			fprintf (out_file, "Found sequence %u (not %u) at "
>  				 "block %u: end of journal.\n",
>  				 sequence, transaction, blocknr);
> -			return;
> +			if (dump_all < 2)
> +				return;
>  		}
>  
>  		if (dump_descriptors) {
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick Wong May 6, 2014, 12:24 a.m. UTC | #2 
On Fri, May 02, 2014 at 01:49:37PM +0200, Lukáš Czerner wrote:
> On Thu, 1 May 2014, Darrick J. Wong wrote:
> 
> > Date: Thu, 01 May 2014 16:13:02 -0700
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > To: tytso@mit.edu, darrick.wong@oracle.com
> > Cc: linux-ext4@vger.kernel.org
> > Subject: [PATCH 06/37] debugfs: force logdump to display (old) journal
> >     contents
> > 
> > If the user passes -a more than once to logdump, try to dump old log
> > contents.  This can be used to try to track down journal problems even
> > after recovery.
> 
> You need to update man page as well for this. Also I wonder what's
> the behaviour if '-a' and '-b' or '-c' are specified simultaneously
> and '-a' is specified multiple times ?

I'll update the manpage.  -c seems to hexdump the contents of any block that we
find while iterating the journal.  -b would seem to allow you to dump an
arbitrary block #, but I could never get it to do that.

In any case, specifying -a even once will make logdump dump every block and
ignore -b.

--D
> 
> Thanks!
> -Lukas
> 
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >  debugfs/logdump.c |   11 ++++++++---
> >  1 file changed, 8 insertions(+), 3 deletions(-)
> > 
> > 
> > diff --git a/debugfs/logdump.c b/debugfs/logdump.c
> > index 8b9dc5b..bf4bef5 100644
> > --- a/debugfs/logdump.c
> > +++ b/debugfs/logdump.c
> > @@ -393,9 +393,13 @@ static void dump_journal(char *cmdname, FILE *out_file,
> >  	fprintf(out_file, "Journal starts at block %u, transaction %u\n",
> >  		blocknr, transaction);
> >  
> > -	if (!blocknr)
> > +	if (!blocknr) {
> >  		/* Empty journal, nothing to do. */
> > -		return;
> > +		if (dump_all < 2)
> > +			return;
> > +		else
> > +			blocknr = 1;
> > +	}
> >  
> >  	while (1) {
> >  		retval = read_journal_block(cmdname, source,
> > @@ -420,7 +424,8 @@ static void dump_journal(char *cmdname, FILE *out_file,
> >  			fprintf (out_file, "Found sequence %u (not %u) at "
> >  				 "block %u: end of journal.\n",
> >  				 sequence, transaction, blocknr);
> > -			return;
> > +			if (dump_all < 2)
> > +				return;
> >  		}
> >  
> >  		if (dump_descriptors) {
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Theodore Ts'o May 12, 2014, 1:41 a.m. UTC | #3 
On Mon, May 05, 2014 at 05:24:53PM -0700, Darrick J. Wong wrote:
> I'll update the manpage.  -c seems to hexdump the contents of any block that we
> find while iterating the journal.  -b would seem to allow you to dump an
> arbitrary block #, but I could never get it to do that.

It's used to dump information _about_ an arbitrary block.  Here's an
example of some of the cool things you can do with logdump:

<tytso@closure> {/usr/projects/e2fsprogs/e2fsprogs}   (next)
1742% gunzip <  tests/f_jnl_32bit/image.gz  > /tmp/image
<tytso@closure> {/usr/projects/e2fsprogs/e2fsprogs}   (next)
1743% debugfs /tmp/image
debugfs 1.42.9 (4-Feb-2014)
debugfs:  logdump -b 680
Journal starts at block 1, transaction 2
  FS block 66 logged at sequence 3, journal block 8 (flags 0x2)
    (block bitmap for block 680: block is SET)
  FS block 680 logged at sequence 3, journal block 205 (flags 0x2)
  FS block 66 logged at sequence 4, journal block 231 (flags 0x2)
    (block bitmap for block 680: block is SET)
  FS block 680 logged at sequence 4, journal block 234 (flags 0x2)
  FS block 66 logged at sequence 5, journal block 339 (flags 0x2)
    (block bitmap for block 680: block is SET)
  FS block 680 logged at sequence 5, journal block 450 (flags 0x2)
No magic number at block 464: end of journal.
debugfs: icheck 680
Block	 Inode number
680	 2132
debugfs:  logdump -i <2132>
Inode 2132 is at group 1, block 364, offset 384
Journal starts at block 1, transaction 2
  FS block 364 logged at sequence 3, journal block 197 (flags 0x2)
    (inode block for inode 2132):
    Inode: 2132   Type: directory        Mode:  0755   Flags: 0x80000
    Generation: 3167953082    Version: 0x00000008
    User:     0   Group:     0   Size: 1024
    File ACL: 0    Directory ACL: 0
    Links: 9   Blockcount: 2
    Fragment:  Address: 0    Number: 0    Size: 0
    ctime: 0x4fa1639e -- Wed May  2 12:41:02 2012
    atime: 0x4fa1639e -- Wed May  2 12:41:02 2012
    mtime: 0x4fa1639e -- Wed May  2 12:41:02 2012
    Blocks:  (0+1): 127754 (1+1): 4 (5+1): 680 
  FS block 364 logged at sequence 4, journal block 233 (flags 0x2)
    (inode block for inode 2132):
    Inode: 2132   Type: directory        Mode:  0755   Flags: 0x80000
    Generation: 3167953082    Version: 0x0000000c
    User:     0   Group:     0   Size: 1024
    File ACL: 0    Directory ACL: 0
    Links: 13   Blockcount: 2
    Fragment:  Address: 0    Number: 0    Size: 0
    ctime: 0x4fa1639e -- Wed May  2 12:41:02 2012
    atime: 0x4fa1639e -- Wed May  2 12:41:02 2012
    mtime: 0x4fa1639e -- Wed May  2 12:41:02 2012
    Blocks:  (0+1): 127754 (1+1): 4 (5+1): 680 
  FS block 364 logged at sequence 5, journal block 434 (flags 0x2)
    (inode block for inode 2132):
    Inode: 2132   Type: directory        Mode:  0755   Flags: 0x80000
    Generation: 3167953082    Version: 0x00000015
    User:     0   Group:     0   Size: 1024
    File ACL: 0    Directory ACL: 0
    Links: 4   Blockcount: 2
    Fragment:  Address: 0    Number: 0    Size: 0
    ctime: 0x4fa163a7 -- Wed May  2 12:41:11 2012
    atime: 0x4fa163a7 -- Wed May  2 12:41:11 2012
    mtime: 0x4fa163a7 -- Wed May  2 12:41:11 2012
    Blocks:  (0+1): 127754 (1+1): 4 (5+1): 680 
No magic number at block 464: end of journal.
debugfs: quit

The idea is that this can be useful when debugging a potentially
corrupted journal, or for advanced file system recovery.

Note that logdump -c is most useful in combination with -b, for
example: "logdump -b 680 -c".

	  	      	  	   	- Ted
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick Wong May 14, 2014, 12:05 a.m. UTC | #4 
On Sun, May 11, 2014 at 09:41:19PM -0400, Theodore Ts'o wrote:
> On Mon, May 05, 2014 at 05:24:53PM -0700, Darrick J. Wong wrote:
> > I'll update the manpage.  -c seems to hexdump the contents of any block that we
> > find while iterating the journal.  -b would seem to allow you to dump an
> > arbitrary block #, but I could never get it to do that.
> 
> It's used to dump information _about_ an arbitrary block.  Here's an
> example of some of the cool things you can do with logdump:

Oh, -b is for FS physical blocks, not for logical blocks in the journal itself,
I get it!  Thanks for pointing that out! :)

The patch (in the other email) looks fine.

--D
> 
> <tytso@closure> {/usr/projects/e2fsprogs/e2fsprogs}   (next)
> 1742% gunzip <  tests/f_jnl_32bit/image.gz  > /tmp/image
> <tytso@closure> {/usr/projects/e2fsprogs/e2fsprogs}   (next)
> 1743% debugfs /tmp/image
> debugfs 1.42.9 (4-Feb-2014)
> debugfs:  logdump -b 680
> Journal starts at block 1, transaction 2
>   FS block 66 logged at sequence 3, journal block 8 (flags 0x2)
>     (block bitmap for block 680: block is SET)
>   FS block 680 logged at sequence 3, journal block 205 (flags 0x2)
>   FS block 66 logged at sequence 4, journal block 231 (flags 0x2)
>     (block bitmap for block 680: block is SET)
>   FS block 680 logged at sequence 4, journal block 234 (flags 0x2)
>   FS block 66 logged at sequence 5, journal block 339 (flags 0x2)
>     (block bitmap for block 680: block is SET)
>   FS block 680 logged at sequence 5, journal block 450 (flags 0x2)
> No magic number at block 464: end of journal.
> debugfs: icheck 680
> Block	 Inode number
> 680	 2132
> debugfs:  logdump -i <2132>
> Inode 2132 is at group 1, block 364, offset 384
> Journal starts at block 1, transaction 2
>   FS block 364 logged at sequence 3, journal block 197 (flags 0x2)
>     (inode block for inode 2132):
>     Inode: 2132   Type: directory        Mode:  0755   Flags: 0x80000
>     Generation: 3167953082    Version: 0x00000008
>     User:     0   Group:     0   Size: 1024
>     File ACL: 0    Directory ACL: 0
>     Links: 9   Blockcount: 2
>     Fragment:  Address: 0    Number: 0    Size: 0
>     ctime: 0x4fa1639e -- Wed May  2 12:41:02 2012
>     atime: 0x4fa1639e -- Wed May  2 12:41:02 2012
>     mtime: 0x4fa1639e -- Wed May  2 12:41:02 2012
>     Blocks:  (0+1): 127754 (1+1): 4 (5+1): 680 
>   FS block 364 logged at sequence 4, journal block 233 (flags 0x2)
>     (inode block for inode 2132):
>     Inode: 2132   Type: directory        Mode:  0755   Flags: 0x80000
>     Generation: 3167953082    Version: 0x0000000c
>     User:     0   Group:     0   Size: 1024
>     File ACL: 0    Directory ACL: 0
>     Links: 13   Blockcount: 2
>     Fragment:  Address: 0    Number: 0    Size: 0
>     ctime: 0x4fa1639e -- Wed May  2 12:41:02 2012
>     atime: 0x4fa1639e -- Wed May  2 12:41:02 2012
>     mtime: 0x4fa1639e -- Wed May  2 12:41:02 2012
>     Blocks:  (0+1): 127754 (1+1): 4 (5+1): 680 
>   FS block 364 logged at sequence 5, journal block 434 (flags 0x2)
>     (inode block for inode 2132):
>     Inode: 2132   Type: directory        Mode:  0755   Flags: 0x80000
>     Generation: 3167953082    Version: 0x00000015
>     User:     0   Group:     0   Size: 1024
>     File ACL: 0    Directory ACL: 0
>     Links: 4   Blockcount: 2
>     Fragment:  Address: 0    Number: 0    Size: 0
>     ctime: 0x4fa163a7 -- Wed May  2 12:41:11 2012
>     atime: 0x4fa163a7 -- Wed May  2 12:41:11 2012
>     mtime: 0x4fa163a7 -- Wed May  2 12:41:11 2012
>     Blocks:  (0+1): 127754 (1+1): 4 (5+1): 680 
> No magic number at block 464: end of journal.
> debugfs: quit
> 
> The idea is that this can be useful when debugging a potentially
> corrupted journal, or for advanced file system recovery.
> 
> Note that logdump -c is most useful in combination with -b, for
> example: "logdump -b 680 -c".
> 
> 	  	      	  	   	- Ted
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/debugfs/logdump.c b/debugfs/logdump.c
index 8b9dc5b..bf4bef5 100644
--- a/debugfs/logdump.c
+++ b/debugfs/logdump.c
@@ -393,9 +393,13 @@  static void dump_journal(char *cmdname, FILE *out_file,
 	fprintf(out_file, "Journal starts at block %u, transaction %u\n",
 		blocknr, transaction);
 
-	if (!blocknr)
+	if (!blocknr) {
 		/* Empty journal, nothing to do. */
-		return;
+		if (dump_all < 2)
+			return;
+		else
+			blocknr = 1;
+	}
 
 	while (1) {
 		retval = read_journal_block(cmdname, source,
@@ -420,7 +424,8 @@  static void dump_journal(char *cmdname, FILE *out_file,
 			fprintf (out_file, "Found sequence %u (not %u) at "
 				 "block %u: end of journal.\n",
 				 sequence, transaction, blocknr);
-			return;
+			if (dump_all < 2)
+				return;
 		}
 
 		if (dump_descriptors) {