xfs
[Top] [All Lists]

Re: xfsrestore: "is a directory: discarding ino"

To: Mandy Kirkconnell <alkirkco@xxxxxxx>
Subject: Re: xfsrestore: "is a directory: discarding ino"
From: Tim Shimmin <tes@xxxxxxx>
Date: Tue, 23 Mar 2004 12:11:16 +1100
Cc: Jason Parker-Burlingham <jasonp@xxxxxxxxx>, Nathan Scott <nathans@xxxxxxx>, linux-xfs@xxxxxxxxxxx
In-reply-to: <405F6B77.70308@xxxxxxx>; from alkirkco@xxxxxxx on Mon, Mar 22, 2004 at 04:40:55PM -0600
References: <878yhtpmxi.fsf@xxxxxxxxxxxxxxx> <20040322212635.GA699@frodo> <405F6B77.70308@xxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.2.5.1i
Hi,

Another thing that may be helpful is some debugging output for restore.
The verbose option of -v5 gives useful stuff.
However, on large restores it may produce annoyingly large output.
If the output is too large then it may be helpful to turn on debugging
at a particular point in the code using mlog_override_level().
For example,
to turn on debugging just prior to the "directory post-processing":
    $ p_rdiff -u content.c
    --- xfsdump/restore/content.c_1.32      2004-03-23 12:06:43.000000000 +1100
    +++ xfsdump/restore/content.c   2004-03-23 12:00:15.000000000 +1100
    @@ -2343,6 +2343,7 @@
                    mlog( MLOG_VERBOSE, _(
                          "directory post-processing\n") );
                    win_locks_off(); /* we are single threaded here */
    +mlog_override_level(5);
                    rv = treepost( path1, path2 );
                    win_locks_on();
                    switch ( rv ) {
 
--Tim


On Mon, Mar 22, 2004 at 04:40:55PM -0600, Mandy Kirkconnell wrote:
> Nathan Scott wrote:
> > On Sun, Mar 21, 2004 at 10:42:33PM -0500, Jason Parker-Burlingham wrote:
> > 
> >>I'm running a 2.4.23 kernel from CVS with LVM1 patches applied.  I'm
> >>trying to move an XFS filesystem from one logical volume to another
> >>using xfsrestore 2.2.18 and not having a lot of success:
> >>
> >>   # xfsdump -J - /mnt | xfsrestore -J - /pivot
> >>   [...]
> >>   xfsrestore: reading directories
> >>   xfsdump: dumping non-directory files
> >>   xfsrestore: 2694 directories and 69903 entries processed
> >>   xfsrestore: directory post-processing
> >>   xfsrestore: WARNING: mkdir  failed: No such file or directory
> >>   xfsrestore: WARNING: mkdir  failed: No such file or directory
> >>   xfsrestore: WARNING: mkdir  failed: No such file or directory
> >>   xfsrestore: WARNING: mkdir  failed: No such file or directory
> >>   xfsrestore: WARNING: unable to set non-root extended attribute for : No 
> >> such file or directory (2)
> >>   xfsrestore: WARNING: unable to set secure extended attribute for : No 
> >> such file or directory (2)
> >>   xfsrestore: restoring non-directory files
> >>   xfsrestore: WARNING: open of ///gconf.xml.defaults/schemas///UI/ failed: 
> >> No such file or directory: discarding ino 135
> >>   [...]
> >>   xfsrestore: WARNING: open of /Kodak Pictures// failed: Is a directory: 
> >> discarding ino 10811
> >>
> >>The point at which the errors occur does not appear to be repeatable.
> >>
> >>If I let the dump/restore run to completion, xfsrestore reports
> >>success and the filesystem appears to be complete, but directory
> >>permissions are set to 000.
> >>
> >>Is there some other option I should be passing to xfsrestore?
> > 
> > 
> > Don't think so, looks like a bug.  Almost looks like the filename
> > string is null for some reason, somewhere in restore (those mkdir
> > errors, etc, above - look like they're trying to print filenames).
> > 
> > Does this ring any bells Mandy?  
> 
> No, this is new to me.  I agree that it looks like there are a bunch of 
> null filename strings.  Jason, you mentioned that the point of error 
> doesn't seem to be consistent but it sounds like it's reproducible in 
> the sense that you've seen this behaviour more than once.  Is there 
> anything unique or special about the data you're restoring?  It looks 
> like the FS may have extended attributes (non-root, secure, or both ??). 
>   I haven't been able to repro this yet so any additional info would be 
> helpful.
> 
> Mandy
> 


<Prev in Thread] Current Thread [Next in Thread>