(pseudo) incremental backup with different exclude lists using hardlinks and rsync

1. Introduction

ccollect is a backup utility written in the sh-scripting language. It does not depend on a specific shell, only /bin/sh needs to be bourne shell compatible (like dash, ksh, zsh, bash, …).

1.1. Supported and tested operating systems and architectures

ccollect was successfully tested on the following platforms:

  • FreeBSD on amd64/i386

  • GNU/Linux on amd64/arm/hppa/i386/ppc

  • Mac OS X 10.5

  • NetBSD on alpha/amd64/i386/sparc/sparc64

  • OpenBSD on amd64

It should run on any Unix that supports rsync and has a POSIX-compatible bourne shell. If your platform is not listed above and you have it successfully running, please drop me a mail.

1.2. Why you COULD only backup from remote hosts, not to them

While considering the design of ccollect, I thought about enabling backup to remote hosts. Though this sounds like a nice feature ("Backup my notebook to the server now."), in my opinion it is a bad idea to backup to a remote host.

But as more and more people requested this feature, it was implemented, so you have the choice whether you want to use it or not.

1.2.1. Reason

If you want to backup TO a remote host, you have to loosen security on it.

Imagine the following situation: You backup your farm of webservers TO a backup host somewhere else. Now one of your webservers which has access to your backup host gets compromised.

Your backup server will be compromised, too.

And the attacker will have access to all data on the other webservers.

1.2.2. Doing it securely

Think of it the other way round: The backup server (now behind a firewall, not accessable from outside) connects to the webservers and pulls the data from them. If someone gets access to one of the webservers, this person will perhaps not even see your machine. If the attacker sees connections from a host to the compromised machine, she will not be able to log in on the backup machine. All other backups are still secure.

1.3. Incompatibilities and changes

1.3.1. Versions 0.9 and 1.0

  • Added "Error: " prefix in _exit_err()

1.3.2. Versions 0.8 and 0.9

  • Renamed script to ccollect (.sh is not needed)

  • Removed feature to backup to a host via ccollect, added new tool (FIXME: insert name here) that takes care of this via tunnel

  • Perhaps creating subdirectory of source name (idea from Stefan Schl√∂rholz)

1.3.3. Versions 0.7 and 0.8

The argument order changed:
  • Old: "<interval name> [args] <sources to backup>"

  • New: "[args] <interval name> <sources to backup>"

If you did not use arguments (most people do not), nothing will change for you.

Deletion of incomplete backups using the delete_incomplete option
  • Old: Only incomplete backups from the current interval have been removed

  • New: All incomplete backups are deleted

Support for standard values
  • Old: no support

  • New: Options in $CCOLLECT_CONF/defaults are used as defaults (see below)

1.3.4. Versions 0.6 and 0.7

The format of destination changed:
  • Before 0.7 it was a (link to a) directory

  • As of 0.7 it is a textfile containing the destination

You can update your configuration using tools/config-pre-0.7-to-0.7.sh.

Added remote_host
  • As of 0.7 it is possible to backup to hosts (see section remote_host below).

1.3.5. Versions 0.5 and 0.6

The format of rsync_options changed:
  • Before 0.6 it was whitespace delimeted

  • As of 0.6 it is newline seperated (so you can pass whitespaces to rsync)

You can update your configuration using tools/config-pre-0.6-to-0.6.sh.

The name of the backup directories changed:
  • Before 0.6: "date +%Y-%m-%d-%H%M"

  • As of 0.6: "date +%Y%m%d-%H%M" (better readable, date is closer together)

For the second change there is no updated needed, as XXXX- is always before XXXXX (- comes before digit).

1.3.6. Versions 0.4 and 0.5

Not a real incompatibilty, but seems to fit in this section:

0.5 does NOT require
  • PaX

  • bc


1.3.7. Versions < 0.4 and 0.4

Since ccollect 0.4 there are several incompatibilities with earlier versions:

List of incompatibilities
  • pax (Posix) is now required, cp -al (GNU specific) is removed

  • "interval" was written with two l (ell), which is wrong in English

  • Changed the name of backup directories, removed the colon in the interval

  • ccollect will now exit when preexec returns non-zero

  • ccollect now reports when postexec returns non-zero

You can convert your old configuration directory using config-pre-0.4-to-0.4.sh, which can be found in the tools/ subdirectory:

[10:05] hydrogenium:ccollect-0.4# ./tools/config-pre-0.4-to-0.4.sh /etc/ccollect

2. Quick start

For those who do not want to read the whole long document:

# get latest ccollect tarball from http://www.nico.schottelius.org/software/ccollect/
# replace value for CCV with the current version
export CCV=0.8.1

# replace 'wget' with 'fetch' on bsd
"$holen" http://www.nico.schottelius.org/software/ccollect/ccollect-${CCV}.tar.bz2

# extract the tarball, change to the newly created directory
tar -xvjf ccollect-${CCV}.tar.bz2
cd ccollect-${CCV}

# create mini-configuration
# first create directory structure
mkdir -p miniconfig/defaults/intervals
mkdir    miniconfig/sources

# create sample intervals
echo 2 > miniconfig/defaults/intervals/testinterval
echo 3 > miniconfig/defaults/intervals/testinterval2

# create destination directory, where the backups will be kept
mkdir ~/DASI

# create sample source, which will be saved
mkdir miniconfig/sources/testsource

# We will save '/bin' to the directory '~/DASI'
echo '/bin' > miniconfig/sources/testsource/source

# configure ccollect to use ~/DASI as destination
echo ~/DASI > miniconfig/sources/testsource/destination

# We want to see what happens and also a small summary at the end
touch miniconfig/sources/testsource/verbose
touch miniconfig/sources/testsource/summary

echo "do the backup, twice"
CCOLLECT_CONF=./miniconfig ./ccollect.sh testinterval testsource
CCOLLECT_CONF=./miniconfig ./ccollect.sh testinterval testsource

echo "the third time ccollect begins to remove old backups"
echo -n "Hit enter to see it"
CCOLLECT_CONF=./miniconfig ./ccollect.sh testinterval testsource

echo "Now we add another interval, ccollect should clone from existent ones"
echo -n "Hit enter to see it"
CCOLLECT_CONF=./miniconfig ./ccollect.sh testinterval2 testsource

echo "Let's see how much space we used with two backups and compare it to /bin"
du -s ~/DASI /bin

# report success
echo "Please report success using ./tools/report_success.sh"

Cutting and pasting the complete section above to your shell will result in the download of ccollect, the creation of a sample configuration and the execution of some backups.

3. Requirements

3.1. Installing ccollect

For the installation you need at least

3.2. Using ccollect

Running ccollect requires the following tools to be installed:
  • date

  • rsync

  • ssh (if you want to use rsync over ssh, which is recommened for security)

4. Installing

Either type make install or simply copy it to a directory in your $PATH and execute chmod 0755 /path/to/ccollect.sh. If you like to use the new management scripts (available since 0.6), copy the following scripts to a directory in $PATH:

  • tools/ccollect_add_source.sh

  • tools/ccollect_analyse_logs.sh.sh

  • tools/ccollect_delete_source.sh

  • tools/ccollect_list_intervals.sh

  • tools/ccollect_logwrapper.sh

After having installed and used ccollect, report success using ./tools/report_success.sh.

5. Configuring

For configuration aid have a look at the above mentioned tools, which can assist you quite well. When you are successfully using ccollect, I would be happy if you add a link to your website, stating "I backup with ccollect", which points to the ccollect homepage. So more people now about ccollect, use it and improve it. You can also report success using tools/report_success.sh.

5.1. Runtime options

ccollect looks for its configuration in /etc/ccollect or, if set, in the directory specified by the variable $CCOLLECT_CONF:

# sh-compatible (dash, zsh, mksh, ksh, bash, ...)
$ CCOLLECT_CONF=/your/config/dir ccollect.sh ...

# csh
$ ( setenv CCOLLECT_CONF /your/config/dir ; ccollect.sh ... )

When you start ccollect, you have to specify in which interval to backup (daily, weekly, yearly; you can specify the names yourself, see below) and which sources to backup (or -a to backup all sources).

The interval specifies how many backups are kept.

There are also some self-explanatory parameters you can pass to ccollect, simply use ccollect.sh --help for info.

5.2. General configuration

The general configuration can be found in $CCOLLECT_CONF/defaults or /etc/ccollect/defaults. All options specified there are generally valid for all source definitions, although the values can be overwritten in the source configuration.

All configuration entries are plain-text files (use UTF-8 for non-ascii characters).

5.2.1. Interval definition

The interval definition can be found in $CCOLLECT_CONF/defaults/intervals/ or /etc/ccollect/defaults/intervals. Each file in this directory specifies an interval. The name of the file is the same as the name of the interval: intervals/'<interval name>'.

The content of this file should be a single line containing a number. This number defines how many versions of this interval are kept.


   [10:23] zaphodbeeblebrox:ccollect-0.2% ls -l conf/defaults/intervals/
   insgesamt 12
   -rw-r--r--  1 nico users 3 2005-12-08 10:24 daily
   -rw-r--r--  1 nico users 3 2005-12-08 11:36 monthly
   -rw-r--r--  1 nico users 2 2005-12-08 11:36 weekly
   [10:23] zaphodbeeblebrox:ccollect-0.2% cat conf/defaults/intervals/*

This means to keep 28 daily backups, 12 monthly backups and 4 weekly.

5.2.2. General pre- and post-execution

If you add $CCOLLECT_CONF/defaults/pre_exec or /etc/ccollect/defaults/pre_exec (same with post_exec), ccollect will start pre_exec before the whole backup process and post_exec after backup of all sources is done.

If pre_exec exits with a non-zero return code, the whole backup process will be aborted.

The pre_exec and post_exec script can access the following exported variables:

  • INTERVAL: the interval selected (daily)

  • no_sources: number of sources to backup (2)

  • source_$no: name of the source, $no starts at 0 ($source_0)

The following example describes how to report free disk space in human readable format before and after the whole backup process:

[13:00] hydrogenium:~# mkdir -p /etc/ccollect/defaults/
[13:00] hydrogenium:~# echo '#!/bin/sh' >  /etc/ccollect/defaults/pre_exec
[13:01] hydrogenium:~# echo ''          >> /etc/ccollect/defaults/pre_exec
[13:01] hydrogenium:~# echo 'df -h'     >> /etc/ccollect/defaults/pre_exec
[13:01] hydrogenium:~# chmod 0755 /etc/ccollect/defaults/pre_exec
[13:01] hydrogenium:~# ln -s /etc/ccollect/defaults/pre_exec /etc/ccollect/defaults/post_exec

5.3. Source configuration

Each source configuration exists in $CCOLLECT_CONF/sources/$name or /etc/ccollect/sources/$name.

The name you choose for the subdirectory describes the source.

Each source contains at least the following files:

  • source (a text file containing the rsync compatible path to backup)

  • destination (a text file containing the directory we should backup to)

Additionally a source may have the following files:

  • pre_exec program to execute before backing up this source

  • post_exec program to execute after backing up this source

  • verbose whether to be verbose (passes -v to rsync)

  • very_verbose be very verbose (mkdir -v, rm -v and rsync -vv)

  • summary create a transfer summary when rsync finished

  • exclude exclude list for rsync. newline seperated list.

  • rsync_options extra options for rsync. newline seperated list.

  • delete_incomplete delete incomplete backups

  • remote_host host to backup to

  • rsync_failure_codes list of rsync exit codes that indicate complete failure

  • mtime Sort backup directories based on their modification time

  • quiet_if_down Suppress error messages if source is not connectable


   [10:47] zaphodbeeblebrox:ccollect-0.2% ls -l  conf/sources/testsource2
   insgesamt 12
   lrwxrwxrwx  1 nico users   20 2005-11-17 16:44 destination
   -rw-r--r--  1 nico users   62 2005-12-07 17:43 exclude
   drwxr-xr-x  2 nico users 4096 2005-12-07 17:38 intervals
   -rw-r--r--  1 nico users   15 2005-11-17 16:44 source
   [10:47] zaphodbeeblebrox:ccollect-0.2% cat conf/sources/testsource2/exclude
   etwas mit leerzeichenli
   [10:47] zaphodbeeblebrox:ccollect-0.2% ls -l  conf/sources/testsource2/intervals
   insgesamt 4
   -rw-r--r--  1 nico users 2 2005-12-07 17:38 daily
   [10:48] zaphodbeeblebrox:ccollect-0.2% cat conf/sources/testsource2/intervals/daily
   [10:48] zaphodbeeblebrox:ccollect-0.2% cat conf/sources/testsource2/source

5.3.1. Default options

If you add $CCOLLECT_CONF/defaults/option_name, the value will be used in abscence of the option in a source. If you want to prevent the default value to be used in a source, you can create the file $CCOLLECT_CONF/sources/$name/no_option_name (i.e. prefix it with no_.


   [9:04] ikn2:ccollect% touch conf/defaults/verbose
   [9:04] ikn2:ccollect% touch conf/sources/local/no_verbose

This enables the verbose option for all sources, but disables it for the source local.

If an option is specified in the defaults folder and in the source, the source specific version overrides the default one:


   [9:05] ikn2:ccollect% echo "backup-host" > conf/defaults/remote_host
   [9:05] ikn2:ccollect% echo "different-host" > conf/sources/local/remote_host

You can use all source options as defaults, with the exception of

  • source

  • destination

  • pre_exec

  • post_exec

5.3.2. Detailed description of "source"

source describes a rsync compatible source (one line only).

For instance backup_user@foreign_host:/home/server/video. To use the rsync protocol without the ssh-tunnel, use rsync::USER@HOST/SRC. For more information have a look at the manpage of rsync(1).

5.3.3. Detailed description of "destination"

destination must be a text file containing the destination directory. destination USED to be a link to the destination directory in earlier versions, so do not be confused if you see such examples.


   [11:36] zaphodbeeblebrox:ccollect-0.2% cat conf/sources/testsource2/destination

5.3.4. Detailed description of "remote_host"

remote_host must be a text file containing the destination host. If this file is existing, you are backing up your data TO this host and not to you local host.

Warning: You need to have ssh access to the remote host. rsync and ccollect will connect to that host via ssh. ccollect needs the shell access, because it needs to find out how many backups exist on the remote host and to be able to delete them.


   [10:17] denkbrett:ccollect-0.7.0% cat conf/sources/remote1/remote_host

It may contain all the ssh-specific values like myuser@yourhost.ch.

5.3.5. Detailed description of "verbose"

verbose tells ccollect that the log should contain verbose messages.

If this file exists in the source specification -v will be passed to rsync.


   [11:35] zaphodbeeblebrox:ccollect-0.2% touch conf/sources/testsource1/verbose

5.3.6. Detailed description of "very_verbose"

very_verbose tells ccollect that it should log very verbosely.

If this file exists in the source specification -v will be passed to rsync, rm and mkdir.


   [23:67] nohost:~% touch conf/sources/testsource1/very_verbose

5.3.7. Detailed description of "summary"

If you create the file summary in the source definition, ccollect will present you a nice summary at the end.

backup:~# touch /etc/ccollect/sources/root/summary
backup:~# ccollect.sh werktags root
==> ccollect.sh: Beginning backup using interval werktags <==
[root] Beginning to backup this source ...
[root] Currently 3 backup(s) exist, total keeping 50 backup(s).
[root] Beginning to backup, this may take some time...
[root] Hard linking...
[root] Transferring files...
[root] Number of files: 84183
[root] Number of files transferred: 32
[root] Total file size: 26234080536 bytes
[root] Total transferred file size: 9988252 bytes
[root] Literal data: 9988252 bytes
[root] Matched data: 0 bytes
[root] File list size: 3016771
[root] File list generation time: 1.786 seconds
[root] File list transfer time: 0.000 seconds
[root] Total bytes sent: 13009119
[root] Total bytes received: 2152
[root] sent 13009119 bytes  received 2152 bytes  2891393.56 bytes/sec
[root] total size is 26234080536  speedup is 2016.26
[root] Successfully finished backup.
==> Finished ccollect.sh <==

You could also combine it with verbose or very_verbose, but these already print some statistics (though not all / the same as presented by summary).

5.3.8. Detailed description of "exclude"

exclude specifies a list of paths to exclude. The entries are seperated by a newline (\n).


   [11:35] zaphodbeeblebrox:ccollect-0.2% cat conf/sources/testsource2/exclude
   etwas mit leerzeichenli
   something with spaces is not a problem

5.3.9. Detailed description of "intervals/"

When you create the subdirectory intervals/ in your source configuration directory, you can specify individiual intervals for this specific source. Each file in this directory describes an interval.


   [11:37] zaphodbeeblebrox:ccollect-0.2% ls -l conf/sources/testsource2/intervals/
   insgesamt 8
   -rw-r--r--  1 nico users 2 2005-12-07 17:38 daily
   -rw-r--r--  1 nico users 3 2005-12-14 11:33 yearly
   [11:37] zaphodbeeblebrox:ccollect-0.2% cat  conf/sources/testsource2/intervals/*

5.3.10. Detailled description of "rsync_options"

When you create the file rsync_options in your source configuration, all the parameters in this file will be passed to rsync. This way you can pass additional options to rsync. For instance you can tell rsync to show progress ("--progress"), or which -password-file ("--password-file") to use for automatic backup over the rsync-protocol.


   [23:42] hydrogenium:ccollect-0.2% cat conf/sources/test_rsync/rsync_options

5.3.11. Detailled description of "pre_exec" and "post_exec"

When you create pre_exec and / or post_exec in your source configuration, ccollect will execute this command before and respectively after doing the backup for this specific source. If you want to have pre-/post-exec before and after all backups, see above for general configuration.

If pre_exec exits with a non-zero return code, the backup process of this source will be aborted (i.e. backup skipped).

The post_exec script can access the following exported variables from ccollect:

  • name: name of the source that is being backed up

  • destination_name: contains the base directory name (daily.20091031-1013.24496)

  • destination_dir: full path (/tmp/ccollect/daily.20091031-1013.24496)

  • destination_full: like destination_dir, but prepended with the remote_host, if set (host:/tmp/ccollect/daily.20091031-1013.24496 or /tmp/ccollect/daily.20091031-1013.24496)


[13:09] hydrogenium:ccollect-0.3% cat conf/sources/with_exec/pre_exec

# Show whats free before
df -h
[13:09] hydrogenium:ccollect-0.3% cat conf/sources/with_exec/post_exec

# Show whats free after
df -h

5.3.12. Detailed description of "delete_incomplete"

If you create the file delete_incomplete in a source specification directory, ccollect will look for incomplete backups (when the whole ccollect process was interrupted) and remove them. Without this file ccollect will only warn the user.

5.3.13. Detailed description of "rsync_failure_codes"

If you have the file rsync_failure_codes in your source configuration directory, it should contain a newline-separated list of numbers representing rsync exit codes. If rsync exits with any code in this list, a marker will be left in the destination directory indicating failure of this backup. If you have enabled delete_incomplete, then this backup will be deleted during the next ccollect run on the same interval.

5.3.14. Detailed description of "mtime"

By default, ccollect.sh chooses the most recent backup directory for cloning or the oldest for deletion based on the directory’s last change time (ctime). With this option, the sorting is done based on modification time (mtime). With this version of ccollect, the ctime and mtime of your backups will normally be the same and this option has no effect. However, if you, for example, move your backups to another hard disk using cp -a or rsync -a, you should use this option because the ctimes are not preserved during such operations.

If you have any backups in your repository made with ccollect version 0.7.1 or earlier, do not use this option.

5.3.15. Detailed description of "quiet_if_down"

By default, ccollect.sh emits a series of error messages if a source is not connectable. With this option enabled, ccollect still reports that the source is not connectable but the associated error messages generated by rsync or ssh are suppressed. You may want to use this option for sources, like notebook PCs, that are often disconnected.

6. Hints

6.1. Smart logging

Since ccollect-0.6.1 you can use the ccollect-logwrapper.sh(1) for logging. You call it the same way you call ccollect.sh and it will create a logfile containing the output of ccollect.sh. For more information look at the manpage ccollect-logwrapper. The following is an example running ccollect-logwrapper.sh:

u0219 ~ # ~chdscni9/ccollect-logwrapper.sh daily u0160.nshq.ch.netstream.com
ccollect-logwrapper.sh (11722): Starting with arguments: daily u0160.nshq.ch.netstream.com
ccollect-logwrapper.sh (11722): Finished.

6.2. Using a different ssh port

Mostly easy is to use your ~/.ssh/config file:

host mx2.schottelius.org
   Port 2342

If you only use that port for backup only and normally want to use another port, you can add HostName and "HostKeyAlias" (if you also have different keys on the different ports):

Host hhydrogenium
   Hostname bruehe.schottelius.org
   Port 666
   HostKeyAlias hydrogenium

Host bruehe
   Hostname bruehe.schottelius.org
   Port 22
   HostKeyAlias bruehe.schottelius.org

6.3. Using source names or interval in pre_/post_exec scripts

The pre-/post_exec scripts can access some internal variables from ccollect:

  • INTERVAL: The interval specified on the command line

  • no_sources: number of sources

  • source_$NUM: the name of the source

  • name: the name of the currently being backuped source (not available for generic pre_exec script)

Only available for post_exec:

  • remote_host: name of host we backup to (empty if unused)

6.4. Using rsync protocol without ssh

When you have a computer with little computing power, it may be useful to use rsync without ssh, directly using the rsync protocol (specify user@host::share in source). You may wish to use rsync_options to specify a password file to use for automatic backup.


backup:~# cat /etc/ccollect/sources/sample.backup.host.org/source

backup:~# cat /etc/ccollect/sources/sample.backup.host.org/rsync_options

backup:~# cat /etc/ccollect/sources/sample.backup.host.org/rsync_password

This hint was reported by Daniel Aubry.

6.5. Not excluding top-level directories

When you exclude "/proc" or "/mnt" from your backup, you may run into trouble when you restore your backup. When you use "/proc/*" or "/mnt/\*" instead, ccollect will backup empty directories.


When those directories contain hidden files (those beginning with a dot (.)), they will still be transferred!

This hint was reported by Marcus Wagner.

6.6. Re-using already created rsync-backups

If you used rsync directly before you use ccollect, you can use this old backup as initial backup for ccollect: You simply move it into a directory below the destination directory and name it "interval.0".


backup:/home/backup/web1# ls
bin   dev  etc   initrd  lost+found  mnt  root  srv  usr  vmlinuz
boot  doc  home  lib     media       opt  sbin  tmp  var  vmlinuz.old

backup:/home/backup/web1# mkdir daily.0

# ignore error about copying to itself
backup:/home/backup/web1# mv * daily.0 2>/dev/null

backup:/home/backup/web1# ls

Now you can use /home/backup/web1 as the destination for the backup.


It does not matter anymore how you name your directory, as ccollect uses the -c option from ls to find out which directory to clone from.


Older versions (pre 0.6, iirc) had a problem, if you named the first backup something like "daily.initial". It was needed to use the "0" (or some number that is lower than the current year) as extension. ccollect used sort to find the latest backup. ccollect itself uses interval.YEARMONTHDAY-HOURMINUTE.PID. This notation was always before "daily.initial", as numbers are earlier in the list which is produced by sort. So, if you had a directory named "daily.initial", ccollect always diffed against this backup and transfered and deleted files which where deleted in previous backups. This means you simply wasted resources, but your backup had beer complete anyway.

6.7. Using pre_/post_exec

Your pre_/post_exec script does not need to be a script, you can also use a link to

  • an existing program

  • an already written script

The only requirement is that it is executable.

6.8. Using source specific interval definitions

When you are backing up multiple hosts via cron each night, it may be a problem that host "big_server" may only have 4 daily backups, because otherwise its backup device will be full. But for all other hosts you want to keep 20 daily backups. In this case you would create /etc/ccollect/default/intervals/daily containing "20" and /etc/ccollect/sources/big_server/intervals/daily containing "4".

Source specific intervals always overwrite the default values. If you have to specify it individually for every host, because of different requirements, you can even omit creating /etc/ccollect/default/intervals/daily.

6.9. Comparing backups

If you want to see what changed between two backups, you can use rsync directly:

[12:00] u0255:ddba034.netstream.ch#  rsync -n -a --delete --stats --progress daily.20080324-0313.17841/ daily.20080325-0313.31148/

This results in a listing of changes. Because we pass -n to rsync no transfer is made (i.e. report only mode).

This hint was reported by Daniel Aubry.

6.10. Testing for host reachabilty

If you want to test whether the host you try to backup is reachable, you can use the following script as source specific pre-exec:

# ping -c1 -q `cat "/etc/ccollect/sources/$name/source" | cut -d"@" -f2 | cut -d":" -f1`

This prevents the deletion of old backups, if the host is not reachable.

This hint was reported by Daniel Aubry.

6.11. Easy check for errors

If you want to see whether there have been any errors while doing the backup, you can run ccollect together with ccollect_analyse_logs.sh:

$ ccollect | ccollect_analyse_logs.sh e

7. F.A.Q.

7.1. What happens if one backup is broken or empty?

Let us assume that one backup failed (connection broke or the source hard disk had some failures). Therefore we’ve got one incomplete backup in our history.

ccollect will transfer the missing files the next time you use it. This leads to

  • more transferred files

  • much greater disk space usage, as no hardlinks can be used

If the whole ccollect process was interrupted, ccollect (since 0.6) can detect that and remove the incomplete backups, so you can clone from a complete backup instead

7.2. When backing up from localhost the destination is also included. Is this a bug?

No. ccollect passes your source definition directly to rsync. It does not try to analyze it. So it actually does not know if a source comes from local harddisk or from a remote server. And it does not want to. When you backup from the local harddisk (which is perhaps not even a good idea when thinking of security), add the destination to source/exclude. (Daniel Aubry reported this problem)

7.3. Why does ccollect say "Permission denied" with my pre-/postexec script?

The most common error is that you have not given your script the correct permissions. Try chmod 0755 /etc/ccollect/sources/'yoursource'/*_exec`.

When a part of your path you specified in the source is a (symbolic, hard links are not possible for directories) link, the backup must fail.

First of all, let us have a look at how it looks like:

==> ccollect 0.4: Beginning backup using interval taeglich <==
[testsource] Sa Apr 29 00:01:55 CEST 2006 Beginning to backup
[testsource] Currently 0 backup(s) exist(s), total keeping 10 backup(s).
[testsource] Beginning to backup, this may take some time...
[testsource] Creating /etc/ccollect/sources/testsource/destination/taeglich.2006-04-29-0001.3874 ...
[testsource] Sa Apr 29 00:01:55 CEST 2006 Transferring files...
[testsource] rsync: recv_generator: mkdir "/etc/ccollect/sources/testsource/destination/taeglich.2006-04-29-0001.3874/home/user/nico/projekte/ccollect" failed: No such file or directory (2)
[testsource] rsync: stat "/etc/ccollect/sources/testsource/destination/taeglich.2006-04-29-0001.3874/home/user/nico/projekte/ccollect" failed: No such file or directory (2)

So what is the problem? It is very obvious when you look deeper into it:

% cat /etc/ccollect/sources/testsource/source
% ls -l /home/user/nico/projekte
lrwxrwxrwx 1 nico nico 29 2005-12-02 23:28 /home/user/nico/projekte -> oeffentlich/computer/projekte
% ls -l /etc/ccollect/sources/testsource/destination/taeglich.2006-04-29-0001.3874/home/user/nico
lrwxrwxrwx 1 nico nico 29 2006-04-29 00:01 projekte -> oeffentlich/computer/projekte

rsync creates the directory structure before it creates the symbolic link. This link now links to something not reachable (dead link). It is impossible to create subdirectories under the broken link.

In conclusion you cannot use paths with a linked part.

However, you can backup directories containing symbolic links (in this case you could backup /home/user/nico, which contains /home/user/nico/projekte and oeffentlich/computer/projekte).

7.5. How can I prevent missing the right time to enter my password?

As ccollect first deletes the old backups, it may take some time until rsync requests the password for the ssh session from you.

The easiest way not to miss that point is running ccollect in screen, which has the ability to monitor the output for activity. So as soon as your screen beeps, after ccollect began to remove the last directory, you can enter your password (have a look at screen(1), especially "C-a M" and "C-a _", for more information).

7.6. Backup fails, if autofs is running, but sources not reachable

If you are trying to backup a system containing paths that are managed by autofs, you may run into this error:

2009-12-01-23:14:15: ccollect 0.8.1: Beginning backup using interval monatlich
[ikn] 2009-12-01-23:14:15: Beginning to backup
[ikn] 2009-12-01-23:14:15: Executing /home/users/nico/ethz/ccollect/sources/ikn/pre_exec ...
Enter LUKS passphrase:
[ikn] Command successful.
[ikn] key slot 0 unlocked.
[ikn] 2009-12-01-23:14:23: Finished /home/users/nico/ethz/ccollect/sources/ikn/pre_exec (return code 0). [ikn] directory has vanished: "/home/users/nico/privat/firmen/ethz/autofs/projects"
[ikn] directory has vanished: "/home/users/nico/privat/firmen/ethz/autofs/scratch"
[ikn] directory has vanished: "/home/users/nico/privat/firmen/ethz/autofs/sgscratch"
[ikn] directory has vanished: "/home/users/nico/privat/firmen/ethz/autofs/supp"
[ikn] directory has vanished: "/home/users/nico/privat/firmen/ethz/autofs/sysadmin"
[ikn] rsync warning: some files vanished before they could be transferred (code 24) at main.c(1057) [sender=3.0.6]
[ikn] 2009-12-01-23:44:23: Source / is not readable. Skipping.

Thus, if you are unsure whether autofs paths can be mounted during backup, stop autofs in pre_exec and reenable it in post_exec.

8. Examples

8.1. A backup host configuration from scratch

srwali01:~# mkdir /etc/ccollect
srwali01:~# mkdir -p /etc/ccollect/defaults/intervals/
srwali01:~# echo 28 > /etc/ccollect/defaults/intervals/taeglich
srwali01:~# echo 52 > /etc/ccollect/defaults/intervals/woechentlich
srwali01:~# cd /etc/ccollect/
srwali01:/etc/ccollect# mkdir sources
srwali01:/etc/ccollect# cd sources/
srwali01:/etc/ccollect/sources# ls
srwali01:/etc/ccollect/sources# mkdir local-root
srwali01:/etc/ccollect/sources# cd local-root/
srwali01:/etc/ccollect/sources/local-root# echo / > source
srwali01:/etc/ccollect/sources/local-root# cat > exclude << EOF
> /proc
> /sys
> /mnt
srwali01:/etc/ccollect/sources/local-root# echo /mnt/hdbackup/local-root > destination
srwali01:/etc/ccollect/sources/local-root# mkdir /mnt/hdbackup/local-root
srwali01:/etc/ccollect/sources/local-root# ccollect.sh taeglich local-root
/o> ccollect.sh: Beginning backup using interval taeglich
/=> Beginning to backup "local-root" ...
|-> 0 backup(s) already exist, keeping 28 backup(s).

After that, I added some more sources:

srwali01:~# cd /etc/ccollect/sources
srwali01:/etc/ccollect/sources# mkdir windos-wl6
srwali01:/etc/ccollect/sources# cd windos-wl6/
srwali01:/etc/ccollect/sources/windos-wl6# echo /mnt/win/SYS/WL6 > source
srwali01:/etc/ccollect/sources/windos-wl6# echo /mnt/hdbackup/wl6 > destination
srwali01:/etc/ccollect/sources/windos-wl6# mkdir /mnt/hdbackup/wl6
srwali01:/etc/ccollect/sources/windos-wl6# cd ..
srwali01:/etc/ccollect/sources# mkdir windos-daten
srwali01:/etc/ccollect/sources/windos-daten# echo /mnt/win/Daten > source
srwali01:/etc/ccollect/sources/windos-daten# echo /mnt/hdbackup/windos-daten > destination
srwali01:/etc/ccollect/sources/windos-daten# mkdir /mnt/hdbackup/windos-daten

# Now add some remote source
srwali01:/etc/ccollect/sources/windos-daten# cd ..
srwali01:/etc/ccollect/sources# mkdir srwali03
srwali01:/etc/ccollect/sources# cd srwali03/
srwali01:/etc/ccollect/sources/srwali03# cat > exclude << EOF
> /proc
> /sys
> /mnt
> /home
srwali01:/etc/ccollect/sources/srwali03# echo 'root@' > source
srwali01:/etc/ccollect/sources/srwali03# echo /mnt/hdbackup/srwali03 > destination
srwali01:/etc/ccollect/sources/srwali03# mkdir /mnt/hdbackup/srwali03
# du (coreutils) 5.2.1
[10:53] srsyg01:sources% du -sh ~/backupdir
4.6M    /home/nico/backupdir
[10:53] srsyg01:sources% du -sh ~/backupdir/*
4.1M    /home/nico/backupdir/daily.2005-12-08-10:52.28456
4.1M    /home/nico/backupdir/daily.2005-12-08-10:53.28484
4.1M    /home/nico/backupdir/daily.2005-12-08-10:53.28507
4.1M    /home/nico/backupdir/daily.2005-12-08-10:53.28531
4.1M    /home/nico/backupdir/daily.2005-12-08-10:53.28554
4.1M    /home/nico/backupdir/daily.2005-12-08-10:53.28577

srwali01:/etc/ccollect/sources# du -sh /mnt/hdbackup/wl6/
186M    /mnt/hdbackup/wl6/
srwali01:/etc/ccollect/sources# du -sh /mnt/hdbackup/wl6/*
147M    /mnt/hdbackup/wl6/taeglich.2005-12-08-14:42.312
147M    /mnt/hdbackup/wl6/taeglich.2005-12-08-14:45.588

The backup of our main fileserver:

backup:~# df -h /home/backup/srsyg01/
Filesystem            Size  Used Avail Use% Mounted on
                      591G  451G  111G  81% /home/backup/srsyg01
backup:~# du -sh /home/backup/srsyg01/*
432G    /home/backup/srsyg01/daily.2006-01-24-01:00.15990
432G    /home/backup/srsyg01/daily.2006-01-26-01:00.30152
434G    /home/backup/srsyg01/daily.2006-01-27-01:00.4596
435G    /home/backup/srsyg01/daily.2006-01-28-01:00.11998
437G    /home/backup/srsyg01/daily.2006-01-29-01:00.19115
437G    /home/backup/srsyg01/daily.2006-01-30-01:00.26405
438G    /home/backup/srsyg01/daily.2006-01-31-01:00.1148
439G    /home/backup/srsyg01/daily.2006-02-01-01:00.8321
439G    /home/backup/srsyg01/daily.2006-02-02-01:00.15383
439G    /home/backup/srsyg01/daily.2006-02-03-01:00.22567
16K     /home/backup/srsyg01/lost+found
backup:~# du --version | head -n1
du (coreutils) 5.2.1

Newer versions of du also detect the hardlinks, so we can even compare the sizes directly with du:

[8:16] eiche:~# du --version | head -n 1
du (GNU coreutils) 5.93
[8:17] eiche:schwarzesloch# du -slh hydrogenium/*
19G     hydrogenium/durcheinander.0
18G     hydrogenium/durcheinander.2006-01-17-00:27.13820
19G     hydrogenium/durcheinander.2006-01-25-23:18.31328
19G     hydrogenium/durcheinander.2006-01-26-00:11.3332
[8:22] eiche:schwarzesloch# du -sh hydrogenium/*
19G     hydrogenium/durcheinander.0
12G     hydrogenium/durcheinander.2006-01-17-00:27.13820
1.5G    hydrogenium/durcheinander.2006-01-25-23:18.31328
200M    hydrogenium/durcheinander.2006-01-26-00:11.3332

In the second report (without -l) the sizes include the space the inodes of the hardlinks allocate.

8.3. A collection of backups on the backup server

All the data of my important hosts is backuped to eiche into /mnt/schwarzesloch/backup:

[9:24] eiche:backup# ls *
woechentlich.2006-01-26-22:22.4153   woechentlich.2006-02-12-11:48.2461
woechentlich.2006-01-26-22:23.4180   woechentlich.2006-02-18-23:00.7898
woechentlich.2006-02-05-02:43.14281  woechentlich.2006-02-25-23:00.13480
woechentlich.2006-02-06-00:24.15509  woechentlich.2006-03-04-23:00.25439

durcheinander.2006-01-27-11:16.6391   durcheinander.2006-02-13-01:07.2895
durcheinander.2006-01-30-19:29.9505   durcheinander.2006-02-17-08:20.6707
durcheinander.2006-01-30-22:27.9623   durcheinander.2006-02-24-16:24.12461
durcheinander.2006-02-03-09:52.12885  durcheinander.2006-03-03-19:17.18075
durcheinander.2006-02-05-23:00.15068  durcheinander.2006-03-17-22:41.5007

woechentlich.2006-02-04-10:32.13766  woechentlich.2006-02-16-23:00.6185
woechentlich.2006-02-05-23:02.15093  woechentlich.2006-02-23-23:00.11783
woechentlich.2006-02-06-08:22.15994  woechentlich.2006-03-02-23:00.17346
woechentlich.2006-02-06-19:40.16321  woechentlich.2006-03-09-23:00.29317
woechentlich.2006-02-12-11:51.2514   woechentlich.2006-03-16-23:00.4218

And this incremental backup and the archive are copied to an external usb harddisk (attention: you should really use -H to backup the backup):

[9:23] eiche:backup# df -h
Filesystem            Size  Used Avail Use% Mounted on
rootfs                 14G  8.2G  4.9G  63% /
/dev/root              14G  8.2G  4.9G  63% /
/dev/root              14G  8.2G  4.9G  63% /dev/.static/dev
tmpfs                  10M  444K  9.6M   5% /dev
/dev/hdh               29G  3.7M   29G   1% /mnt/datenklo
tmpfs                 110M  4.0K  110M   1% /dev/shm
/dev/mapper/nirvana   112G   90G   23G  81% /mnt/datennirvana
                      230G  144G   86G  63% /mnt/schwarzesloch
/dev/mapper/archiv     38G   20G   19G  52% /mnt/archiv
                      280G   36M  280G   1% /mnt/usb/backup
[9:24] eiche:backup# cat ~/bin/sync-to-usb

rsync -av -H --delete /mnt/schwarzesloch/ "$DDIR/schwarzesloch/"

rsync -av -H --delete /mnt/archiv/ "$DDIR/archiv/"

8.4. Processes running when doing ccollect -j

Truncated output from ps axuwwwf:

   S+   11:40   0:00  |   |   |   \_ /bin/sh /usr/local/bin/ccollect.sh daily -j ddba034 ddba045 ddba046 ddba047 ddba049 ddna010 ddna011
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba034
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba034
   R+   11:40  23:40  |   |   |       |   |   \_ rsync -a --delete --numeric-ids --relative --delete-excluded --link-dest=/home/server/backup/ddba034
   S+   11:40   0:00  |   |   |       |   |       \_ ssh -l root ddba034.netstream.ch rsync --server --sender -vlogDtprR --numeric-ids . /
   S+   11:41   0:11  |   |   |       |   |       \_ rsync -a --delete --numeric-ids --relative --delete-excluded --link-dest=/home/server/backup/ddb
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba034
   S+   11:40   0:00  |   |   |       |       \_ sed s:^:\[ddba034\] :
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba045
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba045
   R+   11:40   0:02  |   |   |       |   |   \_ rm -rf /etc/ccollect/sources/ddba045/destination/daily.2006-10-19-1807.6934
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba045
   S+   11:40   0:00  |   |   |       |       \_ sed s:^:\[ddba045\] :
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba046
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba046
   R+   11:40   0:02  |   |   |       |   |   \_ rm -rf /etc/ccollect/sources/ddba046/destination/daily.2006-10-19-1810.7072
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba046
   S+   11:40   0:00  |   |   |       |       \_ sed s:^:\[ddba046\] :
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba047
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba047
   R+   11:40   0:03  |   |   |       |   |   \_ rm -rf /etc/ccollect/sources/ddba047/destination/daily.2006-10-19-1816.7268
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba047
   S+   11:40   0:00  |   |   |       |       \_ sed s:^:\[ddba047\] :
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba049
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba049
   D+   11:40   0:03  |   |   |       |   |   \_ rm -rf /etc/ccollect/sources/ddba049/destination/daily.2006-10-19-1821.7504
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddba049
   S+   11:40   0:00  |   |   |       |       \_ sed s:^:\[ddba049\] :
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddna010
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddna010
   R+   11:40   0:03  |   |   |       |   |   \_ rm -rf /etc/ccollect/sources/ddna010/destination/daily.2006-10-19-1805.6849
   S+   11:40   0:00  |   |   |       |   \_ /bin/sh /usr/local/bin/ccollect.sh daily ddna010
   S+   11:40   0:00  |   |   |       |       \_ sed s:^:\[ddna010\] :
   S+   11:40   0:00  |   |   |       \_ /bin/sh /usr/local/bin/ccollect.sh daily ddna011
   S+   11:40   0:00  |   |   |           \_ /bin/sh /usr/local/bin/ccollect.sh daily ddna011
   R+   12:08   0:00  |   |   |           |   \_ rm -rf /etc/ccollect/sources/ddna011/destination/daily.2006-10-20-1502.7824
   S+   11:40   0:00  |   |   |           \_ /bin/sh /usr/local/bin/ccollect.sh daily ddna011
   S+   11:40   0:00  |   |   |               \_ sed s:^:\[ddna011\] :

As you can see, six processes are deleting old backups, while one backup (ddba034) is already copying data.