path: root/rdiff-backup/FAQ-body.html

<h3>Table of contents</h3>

<ol><li><a href="#__future__">When I try to run rdiff-backup it says
"ImportError: No module named __future__" or "SyntaxError: invalid
syntax".  What's happening?</a></li>

<li><a href="#verbosity">What do the different verbosity levels mean?</a></li>

<li><a href="#windows">Does rdiff-backup run under Windows?</a></li>

<li><a href="#OSX">Does rdiff-backup run under Mac OS X?</a></li>

<li><a href="#remove_dir">My backup set contains some files that I just realized I don't want/need backed up.  How do I remove them from the backup volume to save space?</li>

<li><a href="#solaris">Does rdiff-backup work under Solaris?</a></li>

<li><a href="#speed">How fast is rdiff-backup?  Can it be run on large
data sets?</a></li>

<li><a href="#statistics">What do the various fields mean in the
session statistics and directory statistics files?</a></li>

<li><a href="#bwlimit">Is there some way to limit rdiff-backup's
bandwidth usage, as in rsync's --bwlimit option?</a></li>

<li><a href="#leak">How much memory should rdiff-backup use?  Is there a
memory leak?</a></li>

<li><a href="#dir_not_empty">I use NFS and keep getting some error that includes "OSError: [Errno 39] Directory not empty"</a></li>

<li><a href="#regress_failure">For some reason rdiff-backup failed
while backing up.  Now every time it runs it says "regressing
destination" and then fails again.  What should I do?</a></li>

</ol>


<h3>Questions and Answers</h3>

<ol>

<a name="__future__">
<li><strong>When I try to run rdiff-backup it says "ImportError: No
module named __future__" or "SyntaxError: invalid syntax".  What's
happening?</strong>

<P>rdiff-backup versions 0.2.x require Python version 2.1 or later,
and versions 0.3.x and later require Python version 2.2 or later.  If
you don't know what version of python you are running, type in "python
-V" from the shell.  I'm sorry if this is inconvenient, but
rdiff-backup uses generators, iterators, nested scoping, and
static/class methods extensively, and these were only added in version
2.2.

<P>If you have two versions of python installed, and running "python"
defaults to an early version, you'll probably have to change the first
line of the rdiff-backup script.  For instance, you could set it to:

<pre>#!/usr/bin/env python2.2</pre>
</li>

<a name="verbosity">
<li><strong>What do the different verbosity levels mean?</strong>

<P>There is no formal specification, but here is a rough description
(settings are always cumulative, so 5 displays everything 4 does):

<P>
<table cellspacing="10">
<tr><td>0</td><td>No information given</td></tr>
<tr><td>1</td><td>Fatal Errors displayed</td></tr>
<tr><td>2</td><td>Warnings</td></tr>
<tr><td>3</td><td>Important messages, and maybe later some global statistics (default)</td></tr>
<tr><td>4</td><td>Some global settings, miscellaneous messages</td></tr>
<tr><td>5</td><td>Mentions which files were changed</td></tr>
<tr><td>6</td><td>More information on each file processed</td></tr>
<tr><td>7</td><td>More information on various things</td></tr>
<tr><td>8</td><td>All logging is dated</td></tr>
<tr><td>9</td><td>Details on which objects are moving across the connection</td></tr>
</table>

<a name="windows">
<li><strong>Does rdiff-backup run under Windows?</strong>

<P>Yes, apparently it is possible.  First, follow Jason Piterak's
instructions:

<pre>
Subject: Cygwin rdiff-backup
From: Jason  Piterak &lt;Jason_Piterak@c-i-s.com&gt;
Date: Mon, 4 Feb 2002 16:54:24 -0500 (13:54 PST)
To: rdiff-backup@keywest.Stanford.EDU

Hello all,
  On a lark, I thought I would attempt to get rdiff-backup to work under
Windows98 under Cygwin. We have a number of NT/Win2K servers in the field
that I'd love to be backing up via rdiff-backup, and this was the start of
getting that working. 

SUMMARY: 
  o You can get all the pieces for rdiff-backup working under Cygwin.
  o The backup process works up to the point of writing any files with
timestamps.
      ... This is because the ':' character is reserved for Alternate Data
Stream (ADS) file designations under NTFS.

HOW TO GET IT WORKING (to a point, anyway):
  o Install Cygwin
  o Download the Python 2.2 update through the Cygwin installer and install.
  o Download the librsync libraries from the usual place, but before
compiling...
  o Cygwin does not use/provide glibc. Because of this, you have to repoint
some header files in the Makefile:

   -- Make sure that you have /usr/include/inttypes.h
      redirected to /usr/include/sys/types.h. Do this by:

      create a file /usr/include/inttypes.h with the contents:
      <protect>#include &lt;sys/types.h&gt;</protect>
  o Put rdiff-backup in your PATH, as you normally would.

</pre>

Then, whenever you use rdiff-backup to back up or restore from a unix
system to Windows, use the <strong>--windows-mode</strong> switch.
This compensates for some windows file systems' inability to store
hard links, symlinks, device files, sockets, fifos, case sensitive
filenames, and filenames with colons (":") in them.  No information
will be lost however, because rdiff-backup stores metadata in a
separate file.

<P>
<a name="OSX">
<li><strong>Does rdiff-backup run under Mac OS X?</strong>

<p>
Yes, but there may be some issues installing librsync.  See this
message from Gerd Knops:

<pre>
From: Gerd Knops <gerti@bitart.com>
Date: Thu, 3 Oct 2002 03:56:47 -0500 (01:56 PDT)

[parts of original message deleted]
these instructions build it fine with all tests running OK
(librsync-0.9.5.1 on OS X 10.2.1):

	aclocal
	autoconf
	automake --foreign --add-missing
	env CFLAGS=-no-cpp-precomp ./configure
	make
	make install
</pre>

Also, if you are backing up to a file system that is not case
sensitive you may need to use "--chars-to-quote A-Z".  If you do use
--chars-to-quote, remember to use it with the same arguments when
restoring or listing incrementes.
</li>

<P>
<a name="remove_dir">
<li><strong>My backup set contains some files that I just realized I
don't want/need backed up.  How do I remove them from the backup
volume to save space?</strong>

<P>Let's take an example.  Suppose you ran
<pre>rdiff-backup /usr /backup</pre>
and now realize that you don't want /usr/local backed up on /backup.
Next time you back up, you run
<pre>rdiff-backup --exclude /usr/local /usr /backup</pre>
so that /usr/local is no longer copied to /backup/usr/local.

However, old information about /usr/local is still present in
/backup/rdiff-backup-data/increments/usr/local.  You can try to
manually remove this old information, but it's safer to let it be
removed by rdiff-backup when you run it with the --remove-older-than
option.
</li>

<P>
<a name="solaris">
<li><strong>Does rdiff-backup work under Solaris?</strong>

<P>There may be a problem with rdiff-backup and Solaris' libthread.
Adding "ulimit -n unlimited" may fix the problem though.  Here is a
post by Kevin Spicer on the subject:

<pre>
Subject: RE: Crash report....still not^H^H^H working
From: "Spicer, Kevin" <Kevin.Spicer@bmrb.co.uk>
Date: Sat, 11 May 2002 23:36:42 +0100
To: rdiff-backup@keywest.Stanford.EDU

Quick mail to follow up on this.. 
My rdiff backup (on Solaris 2.6 if you remember) has now worked
reliably for nearly two weeks after I added...

    ulimit -n unlimited 

to the start of my cron job and created a wrapper script on the remote
machine which looked like this...

    #!/bin/sh 
    ulimit -n unlimited 
    rdiff-backup --server 
    exit 

And changed the remote schema on the command line of rdiff-backup to
call the wrapper script rather than rdiff-backup itself on the remote
machine.  As for the /dev/zero thing I've done a bit of Googleing and
it seems that /dev/zero is used internally by libthread on Solaris
(which doesn't really explain why its opening more than 64 files - but
at least I think I've now got round it).
</pre>
</li>

<P>
<a name="speed">
<li><strong>How fast is rdiff-backup?  Can it be run on large
data sets?</strong>

<P>rdiff-backup can be limited by the CPU, disk IO, or available
bandwidth, and the length of a session can be affected by the amount
of data, how much the data changed, and how many files are present.
That said, in the typical case the number/size of changed files is
relatively small compared to that of unchanged files, and rdiff-backup
is often either CPU or bandwidth bound, and takes time proportional to
the total number of files.  Initial mirrorings will usually be
bandwidth or disk bound, and will take much longer than subsequent
updates.

<P>To give one arbitrary data point, when I back up my personal HD
locally (about 36GB, 530000 files, maybe 500 MB turnover, athlon 2000,
7200 IDE disks, version 0.12.2) rdiff-backup takes about 15 minutes
and is usually CPU bound.
</li>

<p>
<a name="statistics">
<li><strong>What do the various fields mean in the
session statistics and directory statistics files?</strong>

<P>Let's examine an example session statistics file:

<pre>
StartTime 1028200920.44 (Thu Aug  1 04:22:00 2002)
EndTime 1028203082.77 (Thu Aug  1 04:58:02 2002)
ElapsedTime 2162.33 (36 minutes 2.33 seconds)
SourceFiles 494619
SourceFileSize 8535991560 (7.95 GB)
MirrorFiles 493797
MirrorFileSize 8521756994 (7.94 GB)
NewFiles 1053
NewFileSize 23601632 (22.5 MB)
DeletedFiles 231
DeletedFileSize 10346238 (9.87 MB)
ChangedFiles 572
ChangedSourceSize 86207321 (82.2 MB)
ChangedMirrorSize 85228149 (81.3 MB)
IncrementFiles 1857
IncrementFileSize 13799799 (13.2 MB)
TotalDestinationSizeChange 28034365 (26.7 MB)
Errors 0
</pre>

<P>StartTime and EndTime are measured in seconds since the epoch.
ElapsedTime is just EndTime - StartTime, the length of the
rdiff-backup session.

<P>SourceFiles are the number of files found in the source directory,
and SourceFileSize is the total size of those files.  MirrorFiles are
the number of files found in the mirror directory (not including the
rdiff-backup-data directory) and MirrorFileSize is the total size of
those files.  All sizes are in bytes.  If the source directory hasn't
changed since the last backup, MirrorFiles == SourceFiles and
SourceFileSize == MirrorFileSize.

<P>NewFiles and NewFileSize are the total number and size of the files
found in the source directory but not in the mirror directory.  They
are new as of the last backup.

<P>DeletedFiles and DeletedFileSize are the total number and size of
the files found in the mirror directory but not the source directory.
They have been deleted since the last backup.

<P>ChangedFiles are the number of files that exist both on the mirror
and on the source directories and have changed since the previous
backup.  ChangedSourceSize is their total size on the source
directory, and ChangedMirrorSize is their total size on the mirror
directory.

<P>IncrementFiles is the number of increment files written to the
rdiff-backup-data directory, and IncrementFileSize is their total
size.  Generally one increment file will be written for every new,
deleted, and changed file.

<P>TotalDestinationSizeChange is the number of bytes the destination
directory as a whole (mirror portion and rdiff-backup-data directory)
has grown during the given rdiff-backup session.  This is usually
close to IncrementFileSize + NewFileSize - DeletedFileSize +
ChangedSourceSize - ChangedMirrorSize, but it also includes the space
taken up by the hardlink_data file to record hard links.
</li>

<a name="bwlimit">
<li><strong>Is there some way to limit rdiff-backup's
bandwidth usage, as in rsync's --bwlimit option?</strong>

<P>There is no internal rdiff-backup option to do this.  However, the
--sleep-ratio option can limit overall resource usage, including
bandwidth.  Also, external utilities such as <a href="http://www.cons.org/cracauer/cstream.html">cstream</a> can be
used to monitor bandwidth explicitly.  trevor@tecnopolis.ca writes:

<pre>
rdiff-backup --remote-schema
  'cstream -v 1 -t 10000 | ssh %s '\''rdiff-backup --server'\'' | cstream -t 20000'
  'netbak@foo.bar.com::/mnt/backup' localbakdir

(must run from a bsh-type shell, not a csh type)

That would apply a limit in both directions [10000 bytes/sec outgoing,
20000 bytes/sec incoming].  I don't think you'd ever really want to do
this though as really you just want to limit it in one direction.
Also, note how I only -v 1 in one direction.  You probably don't want
to output stats for both directions as it will confuse whatever script
you have parsing the output.  I guess it wouldn't hurt for manual runs
however.
</pre>

To only limit bandwidth in one directory, simply remove one of the
cstream commands.  Two cstream caveats may be worth mentioning:

<ol> <li>Because cstream is limiting the uncompressed data heading
into or out of ssh, if ssh compression is turned on, cstream may be
overly restrictive.</li>

<li>cstream may be "bursty", limiting average bandwidth but allowing
rdiff-backup to exceed it for significant periods.</li>
</ol>

<p>
Another option is to limit bandwidth at a lower (and perhaps more
appropriate) level.  Adam Lazur mentions <a
href="http://lartc.org/wondershaper/">The Wonder Shaper</a>.</p>
</li>

<a name="leak">
<li><strong>How much memory should rdiff-backup use?  Is there a
memory leak?</strong>

<p>The amount of memory rdiff-backup uses should not depend much on
the size of directories being processed.  Keeping track of hard links
may use up memory, so if you have, say, hundreds of thousands of files
hard linked together, rdiff-backup may need tens of MB.

<p>If rdiff-backup seems to be leaking memory, it is probably because
it is using an early version of librsync.  <strong>librsync 0.9.5
leaks lots of memory.</strong> Version 0.9.5.1 should not leak and is
available from the rdiff-backup homepage.
</li>

<a name="dir_not_empty">
<li><strong>I use NFS and keep getting some error that includes "OSError: [Errno 39] Directory not empty"</strong>

<P>Several users have reported seeing errors that contain lines like
this:

<pre>
File "/usr/lib/python2.2/site-packages/rdiff_backup/rpath.py",
    line 661, in rmdir
OSError: [Errno 39] Directory not empty:
    '/nfs/backup/redfish/win/Program Files/Common Files/GMT/Banners/11132'
Exception exceptions.TypeError: "'NoneType' object is not callable"
     in &lt;bound method GzipFile.__del__ of
</pre>

<p> All of these users were backing up onto NFS (Network File System).
I think this is probably a bug in NFS, although tell me if you know
how to make rdiff-backup more NFS-friendly.  To avoid this problem,
run rdiff-backup locally on both ends instead of over NFS.  This
should be faster anyway.
</li>

<a name="regress_failure">
<li><strong>For some reason rdiff-backup failed
while backing up.  Now every time it runs it says "regressing
destination" and then fails again.  What should I do?</strong>

<P>Firstly, this shouldn't happen.  If it does, it indicates a
corrupted destination directory, a bug in rdiff-backup, or some other
serious recurring problem.</p>

<p>However, here is a workaround that you might want to use, even
though it probably won't solve the underlying problem:  In the
destination's rdiff-backup-data directory, there should be two
"current_mirror" files, for instance
current_mirror.2003-09-08T04:22:01-07:00.data and
current_mirror.2003-09-07T16:43:00-07:00.data.  Just delete the one
with the earlier date.  The next time rdiff-backup runs it won't try
regressing the destination.
</li>

</ol>