Re: backup snapshot

[Diogene Laerce]

Hi Mr Heinz,

Thanks for stopping by. ;)


Le 11/08/2015 20:15, Heinz Diehl a écrit :
 On 11.08.2015, Diogene Laerce wrote: 

 Is there a trick I don't see here ? Because if the backup of those
 directories is enough for a full restoration of a system state, this method 
 is far more
 efficient than the others, isn't it ?
 If you backup all your partitions with rsync, all you have to do in an
 emergency case is to boot from e.g. CD or a memory stick [1] and
 reverse the rsync command.

 Example:

 rsync -avxHSAX /home/ /backup/home   -- backup
 rsync -avxHSAX --delete /backup/home/ /home   -- restore

 The --delete parameter will take care of all the files that are not
 in the same state as when they were saved.

 So in case of a complete disaster, just restore your data as
 described. If your boot sector is damaged, you'll however have to
 restore it by hand (which isn't all too difficult).

 [1] http://www.sysresccd.org

I actually modified Rick's script with those rsync parameters you provided
which seem more accurate for my system.

Kind regards,

-- 
“One original thought is worth a thousand mindless quotings.”
“Le vrai n'est pas plus sûr que le probable.”

  Diogene Laerce




signature.asc
Description: OpenPGP digital signature
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Diogene Laerce]

Le 11/08/2015 22:54, Rick Stevens a écrit :
 On 08/11/2015 10:52 AM, Diogene Laerce wrote:
 Hi Rick,

 Le 11/08/2015 19:18, Rick Stevens a écrit :
 On 08/11/2015 09:18 AM, Chris Murphy wrote:
 On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr
 wrote:

 https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync

 Is there a trick I don't see here ? Because if the backup of those
 directories
 is enough for a full restoration of a system state, this method is
 far more
 efficient than the others, isn't it ? As one does not need to reboot
 and
 just have
 to make a script or/and a cron job to make his own snapshot.

 I don't consider a backup and snapshot to be the same thing. Copying
 (or rsyncing) some directories to another volume is a backup. A
 snapshot is a deduplicated copy of something at a particular moment in
 time, on the same volume or storage pool. It's not a backup, in that
 if the pool implodes both the original and snapshot are lost. A
 snapshot can be used as a source for a backup, since you can make a
 snapshot that doesn't change while the backup is happening.

 I agree with Chris. Snapshots are ways of restoring data that has been
 perhaps corrupted or deleted or going back in time to some earlier
 point in the filesystems' life. They aren't backups.

 Backup philosophies and techniques vary depending on what you need for
 your unique situation. I'm not claiming what I do is the best, but this
 is what works for me:

 On the last Friday of a month, I plug in a big ESATA or USB3 drive and
 use it to store the output of Mondo Rescue. I have the backup in the
 form of DVD-sized ISO images I can burn DVDs from and there is a
 recovery DVD you can boot from. That gives me a backup usable to
 restore to bare metal.

 Once a week (or more often if there's been significant changes), I use
 a _different_ ESATA or USB3 drive and run an rsync that backs up
 everything except a few things (/proc, /sys, /dev, /media,
 /var/log/journal, various caches, etc.) to a directory on that external
 drive based on the hostname and date I ran the backup. That permits me
 to restore data that's a bit more recent than the MondoRescue stuff.

 I'd be happy to share the MondoArchive and rsync scripts if you wish.
 Tweak to suit your needs.

 Thank you for the offer.

 I'm going to stick with RedoBackup but I'd really like to have a look at
 your rsync scripts, and maybe at what various caches you think of.
 
 The rsync backup is done via this script:
 
 -- CUT HERE --
 #!/bin/bash
 # Back up system to a specific directory given on the command line.
 Excludes
 # the /proc, /sys, /dev and /media directories
 MYHOST=`hostname`
 TODAY=`date +%d-%b-%Y`
 if [ $# -lt 1 ]; then
 TGT=/media/500GB-Drive/Backups/$MYHOST-BackUp-$TODAY
 OPT=
 else
 TGT=$1/$MYHOST-BackUp-$TODAY
 OPT=--exclude=$1/***
 fi
 
 nice -n 19 /bin/rsync -avXA --exclude-from=/etc/skipdirs.rsync $OPT / $TGT
 -- CUT HERE --
 
 And the /etc/skipdirs.rsync file is:
 
 -- CUT HERE --
 /proc/*
 /sys/*
 /dev/*
 /media/**
 /var/log/journal/*
 /BACKUPS/***
 **/.cache/google-chrome/***
 **/.ccache/***
 /run/media/**
 var/lib/docker/*
 -- CUT HERE --

Thank you for those.

Caches are roughly what I was thinking of but I would not have
thought of nice.. Nice.

Kind regards,
-- 
“One original thought is worth a thousand mindless quotings.”
“Le vrai n'est pas plus sûr que le probable.”

  Diogene Laerce



signature.asc
Description: OpenPGP digital signature
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Paul Cartwright]

On 08/11/2015 02:15 PM, Heinz Diehl wrote:
 So in case of a complete disaster, just restore your data as
 described. If your boot sector is damaged, you'll however have to
 restore it by hand (which isn't all too difficult).

 [1] http://www.sysresccd.org
whenever I mess up my MBR, I boot the boot rescue CD.. it is automatic,
rebuilds the MBR so grub works again..

http://sourceforge.net/p/boot-repair-cd/home/Home/



-- 
Paul Cartwright
Registered Linux User #367800 and new counter #561587

-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Rick Stevens]

On 08/11/2015 10:52 AM, Diogene Laerce wrote:

Hi Rick,

Le 11/08/2015 19:18, Rick Stevens a écrit :

On 08/11/2015 09:18 AM, Chris Murphy wrote:

On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr
wrote:


https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync


Is there a trick I don't see here ? Because if the backup of those
directories
is enough for a full restoration of a system state, this method is
far more
efficient than the others, isn't it ? As one does not need to reboot
and
just have
to make a script or/and a cron job to make his own snapshot.


I don't consider a backup and snapshot to be the same thing. Copying
(or rsyncing) some directories to another volume is a backup. A
snapshot is a deduplicated copy of something at a particular moment in
time, on the same volume or storage pool. It's not a backup, in that
if the pool implodes both the original and snapshot are lost. A
snapshot can be used as a source for a backup, since you can make a
snapshot that doesn't change while the backup is happening.


I agree with Chris. Snapshots are ways of restoring data that has been
perhaps corrupted or deleted or going back in time to some earlier
point in the filesystems' life. They aren't backups.

Backup philosophies and techniques vary depending on what you need for
your unique situation. I'm not claiming what I do is the best, but this
is what works for me:

On the last Friday of a month, I plug in a big ESATA or USB3 drive and
use it to store the output of Mondo Rescue. I have the backup in the
form of DVD-sized ISO images I can burn DVDs from and there is a
recovery DVD you can boot from. That gives me a backup usable to
restore to bare metal.

Once a week (or more often if there's been significant changes), I use
a _different_ ESATA or USB3 drive and run an rsync that backs up
everything except a few things (/proc, /sys, /dev, /media,
/var/log/journal, various caches, etc.) to a directory on that external
drive based on the hostname and date I ran the backup. That permits me
to restore data that's a bit more recent than the MondoRescue stuff.

I'd be happy to share the MondoArchive and rsync scripts if you wish.
Tweak to suit your needs.


Thank you for the offer.

I'm going to stick with RedoBackup but I'd really like to have a look at
your rsync scripts, and maybe at what various caches you think of.


The rsync backup is done via this script:

-- CUT HERE --
#!/bin/bash
# Back up system to a specific directory given on the command line. 
Excludes

# the /proc, /sys, /dev and /media directories
MYHOST=`hostname`
TODAY=`date +%d-%b-%Y`
if [ $# -lt 1 ]; then
TGT=/media/500GB-Drive/Backups/$MYHOST-BackUp-$TODAY
OPT=
else
TGT=$1/$MYHOST-BackUp-$TODAY
OPT=--exclude=$1/***
fi

nice -n 19 /bin/rsync -avXA --exclude-from=/etc/skipdirs.rsync $OPT / $TGT
-- CUT HERE --

And the /etc/skipdirs.rsync file is:

-- CUT HERE --
/proc/*
/sys/*
/dev/*
/media/**
/var/log/journal/*
/BACKUPS/***
**/.cache/google-chrome/***
**/.ccache/***
/run/media/**
var/lib/docker/*
-- CUT HERE --

Obviously, change the TGT bit to reflect where your drive is
located and modify the skipdirs.rsync file to reflect your needs.

As far as the mondoarchive stuff:

-- CUT HERE --
#!/bin/bash
# runmondoarchive.sh - Run mondoarchive with standard options

# Only allow this script to be run as the root user...
if [ $EUID -ne 0 ]; then
echo This script must be run as the root user. Exiting.
exit 1
fi

# Grab this host's name...
THISHOST=`hostname -s`

# Get the date as mm-dd-...
DT=`date +%m-%d-%Y`

# Say where the images are to be saved...
DEST=/media/1TB-Drive/Backups/MondoImages

# Scratch directory...
SCR=/tmp/mondoscratch
if [ ! -d $SCR ]; then
mkdir -p $SCR
fi

# Temp directory...
TMP=/tmp/mondotemp
if [ ! -d $TMP ]; then
mkdir -p $TMP
fi

# Set how big the images can be...
#SZ=640m# CD/CD-R
SZ=4480m# DVD/DVD-R

# Call MondoArchive. Arguments are:
#
#   -O  Back up the host
#   (you can add -V to verify images
#   against the current filesystem)
#   -i  Create ISO files
#   -p $THISHOST-$DTSpecify the prefix each ISO image will
#   have (hostname-dd-mm-)
#   -I /Back up EVERYTHING from the root on down
#   -E /media   Do NOT back up anything in /media
#   -d $DESTDestination where ISO images will go
#   -s $SZ  Set the maximum size of each ISO image
#   -S $SCR Set scratch directory (where ISOs are
#   

Re: backup snapshot

[Diogene Laerce]

Le 11/08/2015 03:31, Erik Grun a écrit :
 Am 11.08.2015 um 03:25 schrieb Erik Grun:
  Am 08.08.2015 um 15:57 schrieb Heinz Diehl:
  On 07.08.2015, Diogene Laerce wrote:

  After a sad experience with a system update, I would like to
  ask if there is a software on Fedora or more generally on Linux
  which would allow me to make a complete snapshot of the system
  ?

  rsync -avxHSAX /source/ /target


  What Heinz Diel is saying.


 What I forgot to add:
 Have a look at this page

 https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync

Is there a trick I don't see here ? Because if the backup of those
directories
is enough for a full restoration of a system state, this method is far more
efficient than the others, isn't it ? As one does not need to reboot and
just have
to make a script or/and a cron job to make his own snapshot.

Thank you,

-- 
“One original thought is worth a thousand mindless quotings.”
“Le vrai n'est pas plus sûr que le probable.”

  Diogene Laerce





signature.asc
Description: OpenPGP digital signature
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Chris Murphy]

On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr wrote:

 https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync

 Is there a trick I don't see here ? Because if the backup of those
 directories
 is enough for a full restoration of a system state, this method is far more
 efficient than the others, isn't it ? As one does not need to reboot and
 just have
 to make a script or/and a cron job to make his own snapshot.

I don't consider a backup and snapshot to be the same thing. Copying
(or rsyncing) some directories to another volume is a backup. A
snapshot is a deduplicated copy of something at a particular moment in
time, on the same volume or storage pool. It's not a backup, in that
if the pool implodes both the original and snapshot are lost. A
snapshot can be used as a source for a backup, since you can make a
snapshot that doesn't change while the backup is happening.

-- 
Chris Murphy
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Diogene Laerce]

Le 11/08/2015 18:18, Chris Murphy a écrit :
 On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr wrote:

 https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync
 Is there a trick I don't see here ? Because if the backup of those
 directories
 is enough for a full restoration of a system state, this method is far more
 efficient than the others, isn't it ? As one does not need to reboot and
 just have
 to make a script or/and a cron job to make his own snapshot.
 I don't consider a backup and snapshot to be the same thing. Copying
 (or rsyncing) some directories to another volume is a backup. A
 snapshot is a deduplicated copy of something at a particular moment in
 time, on the same volume or storage pool. It's not a backup, in that
 if the pool implodes both the original and snapshot are lost. A
 snapshot can be used as a source for a backup, since you can make a
 snapshot that doesn't change while the backup is happening.

Actually, until now, I didn't get the subtle difference between those.

As they don't serve the same purpose : I will use snapshots with bfrts
for the Fedora update system issue, and backups for more general
system failure. Kind of the same conclusion I found here :

   
http://www.esg-global.com/blogs/snapshots-vs-backups-a-great-debate-no-longer/

Thanks again to all of you, kind regards,

-- 
“One original thought is worth a thousand mindless quotings.”
“Le vrai n'est pas plus sûr que le probable.”

  Diogene Laerce




signature.asc
Description: OpenPGP digital signature
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Andrew R Paterson]

Hi,
this stupid argument rears its head again - snapshots vs backups.
At least now someone advocating time-consuming backups has come up with a 
logical argument - You usually put the system into a static state before 
performing a backup.
Trouble is - the backup usually means your system is down (e.g. in single 
user mode) for as long as it takes to perform the backup.
A snapshot on the other hand can be validly performed (on a mirrorred file-
system by shutting apps down (going to single user mode/rebooting) detaching a 
mirror, bringing your system back online immediately (resyncing to a new 
mirror)  then (if you so wish!) performing a normal backup of the detached 
mirror - or (if you simply don't run database apps or such like) simply 
stashng the detached mirror away as todays backup.
Expect lots of flack from keen backup! fans here.

KInd Regards
Andy

On Tuesday 11 August 2015 19:52:02 Diogene Laerce wrote:
 Hi Rick,
 
 Le 11/08/2015 19:18, Rick Stevens a écrit :
  On 08/11/2015 09:18 AM, Chris Murphy wrote:
  On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr
  
  wrote:
  https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync
  
  Is there a trick I don't see here ? Because if the backup of those
  directories
  is enough for a full restoration of a system state, this method is
  far more
  efficient than the others, isn't it ? As one does not need to reboot
  and
  just have
  to make a script or/and a cron job to make his own snapshot.
  
  I don't consider a backup and snapshot to be the same thing. Copying
  (or rsyncing) some directories to another volume is a backup. A
  snapshot is a deduplicated copy of something at a particular moment in
  time, on the same volume or storage pool. It's not a backup, in that
  if the pool implodes both the original and snapshot are lost. A
  snapshot can be used as a source for a backup, since you can make a
  snapshot that doesn't change while the backup is happening.
  
  I agree with Chris. Snapshots are ways of restoring data that has been
  perhaps corrupted or deleted or going back in time to some earlier
  point in the filesystems' life. They aren't backups.
  
  Backup philosophies and techniques vary depending on what you need for
  your unique situation. I'm not claiming what I do is the best, but this
  is what works for me:
  
  On the last Friday of a month, I plug in a big ESATA or USB3 drive and
  use it to store the output of Mondo Rescue. I have the backup in the
  form of DVD-sized ISO images I can burn DVDs from and there is a
  recovery DVD you can boot from. That gives me a backup usable to
  restore to bare metal.
  
  Once a week (or more often if there's been significant changes), I use
  a _different_ ESATA or USB3 drive and run an rsync that backs up
  everything except a few things (/proc, /sys, /dev, /media,
  /var/log/journal, various caches, etc.) to a directory on that external
  drive based on the hostname and date I ran the backup. That permits me
  to restore data that's a bit more recent than the MondoRescue stuff.
  
  I'd be happy to share the MondoArchive and rsync scripts if you wish.
  Tweak to suit your needs.
 
 Thank you for the offer.
 
 I'm going to stick with RedoBackup but I'd really like to have a look at
 your rsync scripts, and maybe at what various caches you think of.
 
 Kind regards,

-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Heinz Diehl]

On 11.08.2015, Diogene Laerce wrote: 

 Is there a trick I don't see here ? Because if the backup of those
 directories is enough for a full restoration of a system state, this method 
 is far more
 efficient than the others, isn't it ?

If you backup all your partitions with rsync, all you have to do in an
emergency case is to boot from e.g. CD or a memory stick [1] and
reverse the rsync command.

Example:

rsync -avxHSAX /home/ /backup/home   -- backup
rsync -avxHSAX --delete /backup/home/ /home   -- restore

The --delete parameter will take care of all the files that are not
in the same state as when they were saved.

So in case of a complete disaster, just restore your data as
described. If your boot sector is damaged, you'll however have to
restore it by hand (which isn't all too difficult).

[1] http://www.sysresccd.org

-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Rick Stevens]

On 08/11/2015 09:18 AM, Chris Murphy wrote:

On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr wrote:


https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync


Is there a trick I don't see here ? Because if the backup of those
directories
is enough for a full restoration of a system state, this method is far more
efficient than the others, isn't it ? As one does not need to reboot and
just have
to make a script or/and a cron job to make his own snapshot.


I don't consider a backup and snapshot to be the same thing. Copying
(or rsyncing) some directories to another volume is a backup. A
snapshot is a deduplicated copy of something at a particular moment in
time, on the same volume or storage pool. It's not a backup, in that
if the pool implodes both the original and snapshot are lost. A
snapshot can be used as a source for a backup, since you can make a
snapshot that doesn't change while the backup is happening.


I agree with Chris. Snapshots are ways of restoring data that has been
perhaps corrupted or deleted or going back in time to some earlier
point in the filesystems' life. They aren't backups.

Backup philosophies and techniques vary depending on what you need for
your unique situation. I'm not claiming what I do is the best, but this
is what works for me:

On the last Friday of a month, I plug in a big ESATA or USB3 drive and
use it to store the output of Mondo Rescue. I have the backup in the
form of DVD-sized ISO images I can burn DVDs from and there is a
recovery DVD you can boot from. That gives me a backup usable to
restore to bare metal.

Once a week (or more often if there's been significant changes), I use
a _different_ ESATA or USB3 drive and run an rsync that backs up
everything except a few things (/proc, /sys, /dev, /media,
/var/log/journal, various caches, etc.) to a directory on that external
drive based on the hostname and date I ran the backup. That permits me
to restore data that's a bit more recent than the MondoRescue stuff.

I'd be happy to share the MondoArchive and rsync scripts if you wish.
Tweak to suit your needs.
--
- Rick Stevens, Systems Engineer, AllDigitalri...@alldigital.com -
- AIM/Skype: therps2ICQ: 226437340   Yahoo: origrps2 -
--
-  Whoever said Money can't buy happiness obviously never had any  -
- money! -
--
--
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Diogene Laerce]

Hi Rick,

Le 11/08/2015 19:18, Rick Stevens a écrit :
 On 08/11/2015 09:18 AM, Chris Murphy wrote:
 On Tue, Aug 11, 2015 at 3:05 AM, Diogene Laerce me_buss...@yahoo.fr
 wrote:

 https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync

 Is there a trick I don't see here ? Because if the backup of those
 directories
 is enough for a full restoration of a system state, this method is
 far more
 efficient than the others, isn't it ? As one does not need to reboot
 and
 just have
 to make a script or/and a cron job to make his own snapshot.

 I don't consider a backup and snapshot to be the same thing. Copying
 (or rsyncing) some directories to another volume is a backup. A
 snapshot is a deduplicated copy of something at a particular moment in
 time, on the same volume or storage pool. It's not a backup, in that
 if the pool implodes both the original and snapshot are lost. A
 snapshot can be used as a source for a backup, since you can make a
 snapshot that doesn't change while the backup is happening.

 I agree with Chris. Snapshots are ways of restoring data that has been
 perhaps corrupted or deleted or going back in time to some earlier
 point in the filesystems' life. They aren't backups.

 Backup philosophies and techniques vary depending on what you need for
 your unique situation. I'm not claiming what I do is the best, but this
 is what works for me:

 On the last Friday of a month, I plug in a big ESATA or USB3 drive and
 use it to store the output of Mondo Rescue. I have the backup in the
 form of DVD-sized ISO images I can burn DVDs from and there is a
 recovery DVD you can boot from. That gives me a backup usable to
 restore to bare metal.

 Once a week (or more often if there's been significant changes), I use
 a _different_ ESATA or USB3 drive and run an rsync that backs up
 everything except a few things (/proc, /sys, /dev, /media,
 /var/log/journal, various caches, etc.) to a directory on that external
 drive based on the hostname and date I ran the backup. That permits me
 to restore data that's a bit more recent than the MondoRescue stuff.

 I'd be happy to share the MondoArchive and rsync scripts if you wish.
 Tweak to suit your needs.

Thank you for the offer.

I'm going to stick with RedoBackup but I'd really like to have a look at
your rsync scripts, and maybe at what various caches you think of.

Kind regards,

-- 
“One original thought is worth a thousand mindless quotings.”
“Le vrai n'est pas plus sûr que le probable.”

  Diogene Laerce




signature.asc
Description: OpenPGP digital signature
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Chris Murphy]

The other context for snapshots are system rollbacks, which is on a
sliding continuum between stateless vs stateful systems. So you can
get certain aspects of statelessness with snapshots, with an otherwise
stateful system. This is how Windows has done updates for a long time
now, and snapper, and Fedora Atomic (rpm-ostree) work. The underlying
technical details of how the snapshot is achieved are dissimilar, but
the basic idea is the same which is you have multiple trees and can
revert to previous states.

So maybe it's better to call these rollbacks in terms of user
selectable stateful states haha. Whereas statelessness is like a
system reset: such as what we find on mobile devices, and since
Windows 8.

Is restoring an rsync backup to a currently running system a rollback?
It's not atomic, and unless you first backup the current state you
can't then do a rollforward after you've done the rollback because
you've overwritten the current state with the backup. And since the
overwrite happens with in-use files, it's not atomic. Any mistakes and
it can easily implode the system in a way that you can't go forward or
backward to get to a bootable system and you're in diagnose and repair
mode.

NTFS shadow copy, snapper+Btrfs (or LVM thinp), and rpm-ostree are all
atomic rollbacks. I think it can be argued that rollbacks imply the
expectation of atomicity. Otherwise you'd just say restoring from
backup or doing a system restore/rebuild from backup.


-- 
Chris Murphy
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Chris Murphy]

On Tue, Aug 11, 2015 at 1:56 PM, jd1008 jd1...@gmail.com wrote:


 On 08/11/2015 01:27 PM, Chris Murphy wrote:

 The other context for snapshots are system rollbacks, which is on a
 sliding continuum between stateless vs stateful systems. So you can
 get certain aspects of statelessness with snapshots, with an otherwise
 stateful system. This is how Windows has done updates for a long time
 now, and snapper, and Fedora Atomic (rpm-ostree) work. The underlying
 technical details of how the snapshot is achieved are dissimilar, but
 the basic idea is the same which is you have multiple trees and can
 revert to previous states.

 So maybe it's better to call these rollbacks in terms of user
 selectable stateful states haha. Whereas statelessness is like a
 system reset: such as what we find on mobile devices, and since
 Windows 8.

 Is restoring an rsync backup to a currently running system a rollback?
 It's not atomic, and unless you first backup the current state you
 can't then do a rollforward after you've done the rollback because
 you've overwritten the current state with the backup. And since the
 overwrite happens with in-use files, it's not atomic. Any mistakes and
 it can easily implode the system in a way that you can't go forward or
 backward to get to a bootable system and you're in diagnose and repair
 mode.

 NTFS shadow copy, snapper+Btrfs (or LVM thinp), and rpm-ostree are all
 atomic rollbacks. I think it can be argued that rollbacks imply the
 expectation of atomicity. Otherwise you'd just say restoring from
 backup or doing a system restore/rebuild from backup.


 Thanx Chris for this elucidation.
 However, it is still not clear how one can do an atomic rollback
 when the currently booted system and perhaps logged in user
 has (have) already modified some files. How does one proceed then?
 Save the files modified since last backup or snapshot, do the restoration/
 rollback, and then go through the tedious process of comparing
 the just saved files against the ones brought by rollback?
 How does one proceed in such a situation?

There's the big picture point of view:

It depends on the snapshot+rollback strategy and underlying storage
technology involved.

It's possible that the snapshot taken is used for rollback; or it can
be the thing updates are applied to.

For snapper+btrfs+yast (on opensuse 13.2) the behavior is to take a
snapshot before any configuration changes including adding or deleting
packages. So the current state is snapshot, then the current state is
modified. A rollback means reverting to a snapshot, which has two
kinds: temporary via GRUB menu, and persistent via a yast or snapper
command.

For ostree and rpm-ostree (Fedora Atomic), the behavior isn't a
filesystem snapshot in a Btrfs sense because this works on plain ext4
and XFS by using uniquely named directory trees and hardlinks. So a
new tree faux-snapshot is setup with a bunch of hardlinks, and then
that new directory tree is what gets the updates. The currently active
tree is not modified. There is no need for a rollback if the update
fails, because the current tree was never modified.

And then there's the detailed view:

The different strategies above have different layouts to separate out
what things are rolled back and not. For example there's never a
snapshot or rollback of /home. Your /home files are always progressing
unidirectionally through time, no matter what. And the same thing
applies to /etc and much of /var. So if you make a mistake with some
system settings, those don't get rolled back with the above
strategies.

Could you have more granularity so you could snapshot and rollback
such things? Sure. But someone would have to do that implementation
work and test it. And probably it'd need to be something fairly
agnostic file system wise so it works on anything; which could just be
a simple config file rename and versioning scheme. If it were Btrfs
based you could very cheaply snapshot these things separate from other
directories and for selective rollback of one file use cp --reflink
behind the scenes.


-- 
Chris Murphy
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[jd1008]

On 08/11/2015 01:27 PM, Chris Murphy wrote:

The other context for snapshots are system rollbacks, which is on a
sliding continuum between stateless vs stateful systems. So you can
get certain aspects of statelessness with snapshots, with an otherwise
stateful system. This is how Windows has done updates for a long time
now, and snapper, and Fedora Atomic (rpm-ostree) work. The underlying
technical details of how the snapshot is achieved are dissimilar, but
the basic idea is the same which is you have multiple trees and can
revert to previous states.

So maybe it's better to call these rollbacks in terms of user
selectable stateful states haha. Whereas statelessness is like a
system reset: such as what we find on mobile devices, and since
Windows 8.

Is restoring an rsync backup to a currently running system a rollback?
It's not atomic, and unless you first backup the current state you
can't then do a rollforward after you've done the rollback because
you've overwritten the current state with the backup. And since the
overwrite happens with in-use files, it's not atomic. Any mistakes and
it can easily implode the system in a way that you can't go forward or
backward to get to a bootable system and you're in diagnose and repair
mode.

NTFS shadow copy, snapper+Btrfs (or LVM thinp), and rpm-ostree are all
atomic rollbacks. I think it can be argued that rollbacks imply the
expectation of atomicity. Otherwise you'd just say restoring from
backup or doing a system restore/rebuild from backup.



Thanx Chris for this elucidation.
However, it is still not clear how one can do an atomic rollback
when the currently booted system and perhaps logged in user
has (have) already modified some files. How does one proceed then?
Save the files modified since last backup or snapshot, do the restoration/
rollback, and then go through the tedious process of comparing
the just saved files against the ones brought by rollback?
How does one proceed in such a situation?

--
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Chris Murphy]

On Sun, Aug 9, 2015 at 1:43 PM, Diogene Laerce me_buss...@yahoo.fr wrote:

 http://snapper.io/
 http://rpm-software-management.github.io/dnf-plugins-extras/snapper.html

 The documentation says that snapper supports ext4 but only
 experimentally, do
 you have any feedback on the matter ?

It requires LVM thinly provisioned volumes. Proper configuration for
snapshots is necessary and I don't know whether the defaults from the
installer do that.





 Le 08/08/2015 02:52, Chris Murphy a écrit :
 The closest it comes is choosing Btrfs for installation (you can use
 something else for /home if you want). The root subvolume on Btrfs can
 then be snapshot before you do a dnf update, update the snapshot's
 copy of fstab, and if things go bad you can change the
 rootflags=subvol=subvolname boot parameter to that of the snapshot
 name. But this assumes some Btrfs knowledge, which at least is not
 nearly as esoteric and complicated like the rabbit hole that is LVM
 thin volume snapshots. But that's also an option if you're at least
 semi-comfortable with LVM.

 I even saw there :


 https://wiki.archlinux.org/index.php/Btrfs_-_Tips_and_tricks#Automatic_snapshots_on_each_boot

 that one can configure automatic snapshots at boot which is not far from
 what you propose.

 I guess this is reproducible on Fedora and that would definitely be my
 preference.

Seems like it's adaptable for Fedora.

So long as the top level of the file system is not mounted (where all
subvolumes and snapshots are located), it's OK to do a conventional
backup of the whole system. But if the top level is mounted anywhere
not excluded from that backup, the backup will see each snapshot as a
completely separate instance of all of that data. So if you have root,
and 5 snapshots, a backup will see six root fs's that need to be
backed up unless you exclude those snapshots somehow (by not making
them available in the mounted hierarchy is easiest).



-- 
Chris Murphy
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Erik Grun]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Am 08.08.2015 um 15:57 schrieb Heinz Diehl:
 On 07.08.2015, Diogene Laerce wrote:
 
 After a sad experience with a system update, I would like to ask 
 if there is a software on Fedora or more generally on Linux which
 would allow me to make a complete snapshot of the system ?
 
 rsync -avxHSAX /source/ /target
 

What Heinz Diel is saying.
-BEGIN PGP SIGNATURE-
Version: GnuPG v2
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=cl/G
-END PGP SIGNATURE-
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Erik Grun]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Am 11.08.2015 um 03:25 schrieb Erik Grun:
 Am 08.08.2015 um 15:57 schrieb Heinz Diehl:
 On 07.08.2015, Diogene Laerce wrote:
 
 After a sad experience with a system update, I would like to
 ask if there is a software on Fedora or more generally on Linux
 which would allow me to make a complete snapshot of the system
 ?
 
 rsync -avxHSAX /source/ /target
 
 
 What Heinz Diel is saying.
 

What I forgot to add:
Have a look at this page

https://wiki.archlinux.org/index.php/Full_System_Backup_with_rsync
-BEGIN PGP SIGNATURE-
Version: GnuPG v2
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=KAnO
-END PGP SIGNATURE-
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot posting

[Paul Livingston]

Excellent post!  I'm not the OP but this scratches and itch for me particularly 
since it is a bare metal solution

Thanks

On Aug 8, 2015, at 8:00 AM, users-requ...@lists.fedoraproject.org wrote:

 From: Michael D. Setzer II mi...@kuentos.guam.net
 Subject: Re: backup snapshot
 Date: August 7, 2015 10:58:25 PM EDT
 To: Diogene Laerce me_buss...@yahoo.fr, users@lists.fedoraproject.org
 
 
 Not a snapshot, but there are various programs that can do a bare image of
 the system and it works with Windows and LInux. They can be done at the
 partition level or the entire disk.
 
 I am the current maintainer of the G4L project, and there is also GNU and
 Clonezilla that can do similar things.
 
 With my classroom lab that has systems with 500G disks with windows 7 and
 Fedora. I have a 160G W7 partition, and make an image of it to another
 partition that is about 24G in size. Takes about 12 minutes to make image,
 and about 10 minutes to restore. Have an option on the grub menu that can
 automatically, restore it, so if students mess up windows, it can be quickly
 restored to the previous image. Use NTFSCLONE option for the windows.
 
 Similar process can be done with Linux, but since it has multiple partitions,
 one needs to do an image of each one, or one can do a full disk image, but it
 has to be made to another device like external disk or ftp server.
 
 Another recommendation, unlike NTFSCLONE, which only backs up used
 data, the raw method will be much more effictive if the unused space on each
 partition is cleared (Nulls written to sectors). Program has options to do 
 this,
 and then make images of each partition or the whole disk.Does take time
 since it has to read every sector, but image is much smaller with
 compression.
 
 I have gotten even better speeds by using USB3 128G  flash. Using the USB
 3 flash, the same windows partition can be reimaged in about 4 1/2 minutes
 using USB 3 port. Takes about 8 minutes in a USB 2 port. Single hard disk
 takes longer, since it has to read and write from same device. The time to
 create the image is about the same, since the compression process seems to
 be the bottle neck there.
 
 I generally always, make images of critical machines, and home machine, so
 that if something goes wrong, I can quickly get a machine back and running
 to a known state. One could just backup the /boot, and / and maybe /home
 partitions depending on setup, and restore them.
 
 The G4L also, has a program calles fsarchiver that is a filelevel backup
 program that works with Linux, but I included it as a request of a user, and
 have time limited testing, in which it worked fine, but prefer the bare metal
 options.
 
 So, not sure if that is the solution you are looking for.

-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Diogene Laerce]

Hi,

Thanks for your answers.


Le 07/08/2015 22:12, Michael Morgan a écrit :
 Btrfs root, snapper, and python-dnf-plugins-extras-snapper have been a  
 godsend for me in this regard. I also run snapper-timeline on my
/home  subvol which has saved me from frustration a few times already.
btrfs sounds good but I will need to wait for the next installation from
scratch as I run on ext4, on multiple partitions /boot (ext2), /etc,
/opt, etc.. And conversion from ext4 does not seem worth the risk, for
the moment.
 http://snapper.io/
 http://rpm-software-management.github.io/dnf-plugins-extras/snapper.html

The documentation says that snapper supports ext4 but only
experimentally, do
you have any feedback on the matter ?


Le 08/08/2015 02:52, Chris Murphy a écrit :
 The closest it comes is choosing Btrfs for installation (you can use
 something else for /home if you want). The root subvolume on Btrfs can
 then be snapshot before you do a dnf update, update the snapshot's
 copy of fstab, and if things go bad you can change the
 rootflags=subvol=subvolname boot parameter to that of the snapshot
 name. But this assumes some Btrfs knowledge, which at least is not
 nearly as esoteric and complicated like the rabbit hole that is LVM
 thin volume snapshots. But that's also an option if you're at least
 semi-comfortable with LVM.

I even saw there :

   
https://wiki.archlinux.org/index.php/Btrfs_-_Tips_and_tricks#Automatic_snapshots_on_each_boot

that one can configure automatic snapshots at boot which is not far from
what you propose.

I guess this is reproducible on Fedora and that would definitely be my
preference.


 Another option, which is still maturing and really intended now only
 as a platform for deploying containers, is Fedora Atomic (rpm-ostree).
 This is really where rollbacks are at, because updates are atomic, and
 it deals with all the gory bootloader details. The tea leaves suggests
 this is the direction for a future Workstation product that can do
 what you describe.

That, seems to be way over my head and what I want to achieve as I just
have one computer to manage.


Le 08/08/2015 04:58, Michael D. Setzer II a écrit :
 Not a snapshot, but there are various programs that can do a bare image of  
 the system and it works with Windows and Linux. They can be done at
the  partition level or the entire disk.   I am the current
maintainer of the G4L project, and there is also GNU and  Clonezilla
that can do similar things. I actually went for redobackup :
http://redobackup.org/ It has a web access as it runs on ubuntu, that's
what convince me. So I will go for this for now, and I will update my
system to btrfs next time I install.
Le 08/08/2015 15:57, Heinz Diehl a écrit :
 rsync -avxHSAX /source/ /target 

Heinz.. Try to catch up please.. Sorry ! I couldn't resist.. :)

Seriously, I'm not sure that rsync, even executed with root privileges, will
have access to all files and will be able to copy the entire /dev/sd[n]
identically. Will it ?

Thanks again to all ! ;)

-- 
One original thought is worth a thousand mindless quotings.
Le vrai n'est pas plus sûr que le probable.
  Diogene Laerce







signature.asc
Description: OpenPGP digital signature
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Heinz Diehl]

On 07.08.2015, Diogene Laerce wrote: 

 After a sad experience with a system update, I would like to ask
 if there is a software on Fedora or more generally on Linux
 which would allow me to make a complete snapshot of the system ?

rsync -avxHSAX /source/ /target

-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Chris Murphy]

On Fri, Aug 7, 2015 at 1:56 PM, Diogene Laerce me_buss...@yahoo.fr wrote:
 Hi,

 After a sad experience with a system update, I would like to ask
 if there is a software on Fedora or more generally on Linux
 which would allow me to make a complete snapshot of the system ?

 Something as easy of use as the snapshot feature in VirtualBox
 would be great.. But near would do as well.

Well, VirtualBox snapshots have an equivalent in virt-manager that are
as easy to use. But if you're talking about bare metal, no there is
nothing even remotely approaching that level of ease for rollbacks
yet.

The closest it comes is choosing Btrfs for installation (you can use
something else for /home if you want). The root subvolume on Btrfs can
then be snapshot before you do a dnf update, update the snapshot's
copy of fstab, and if things go bad you can change the
rootflags=subvol=subvolname boot parameter to that of the snapshot
name. But this assumes some Btrfs knowledge, which at least is not
nearly as esoteric and complicated like the rabbit hole that is LVM
thin volume snapshots. But that's also an option if you're at least
semi-comfortable with LVM.

Another option, which is still maturing and really intended now only
as a platform for deploying containers, is Fedora Atomic (rpm-ostree).
This is really where rollbacks are at, because updates are atomic, and
it deals with all the gory bootloader details. The tea leaves suggests
this is the direction for a future Workstation product that can do
what you describe.

-- 
Chris Murphy
-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Michael D. Setzer II]

Not a snapshot, but there are various programs that can do a bare image of 
the system and it works with Windows and LInux. They can be done at the 
partition level or the entire disk.

I am the current maintainer of the G4L project, and there is also GNU and 
Clonezilla that can do similar things. 

With my classroom lab that has systems with 500G disks with windows 7 and 
Fedora. I have a 160G W7 partition, and make an image of it to another 
partition that is about 24G in size. Takes about 12 minutes to make image, 
and about 10 minutes to restore. Have an option on the grub menu that can 
automatically, restore it, so if students mess up windows, it can be quickly 
restored to the previous image. Use NTFSCLONE option for the windows.

Similar process can be done with Linux, but since it has multiple partitions, 
one needs to do an image of each one, or one can do a full disk image, but it 
has to be made to another device like external disk or ftp server.

Another recommendation, unlike NTFSCLONE, which only backs up used 
data, the raw method will be much more effictive if the unused space on each 
partition is cleared (Nulls written to sectors). Program has options to do 
this, 
and then make images of each partition or the whole disk.Does take time 
since it has to read every sector, but image is much smaller with 
compression.

I have gotten even better speeds by using USB3 128G  flash. Using the USB 
3 flash, the same windows partition can be reimaged in about 4 1/2 minutes 
using USB 3 port. Takes about 8 minutes in a USB 2 port. Single hard disk 
takes longer, since it has to read and write from same device. The time to 
create the image is about the same, since the compression process seems to 
be the bottle neck there. 

I generally always, make images of critical machines, and home machine, so 
that if something goes wrong, I can quickly get a machine back and running 
to a known state. One could just backup the /boot, and / and maybe /home 
partitions depending on setup, and restore them. 

The G4L also, has a program calles fsarchiver that is a filelevel backup 
program that works with Linux, but I included it as a request of a user, and 
have time limited testing, in which it worked fine, but prefer the bare metal 
options.

So, not sure if that is the solution you are looking for. 

On 7 Aug 2015 at 21:56, Diogene Laerce wrote:

To: users@lists.fedoraproject.org
From:   Diogene Laerce me_buss...@yahoo.fr
Subject:backup snapshot
Date sent:  Fri, 7 Aug 2015 21:56:09 +0200
Send reply to:  me_buss...@yahoo.fr,
Community support for Fedora users users@lists.fedoraproject.org

 Hi,
 
 After a sad experience with a system update, I would like to ask
 if there is a software on Fedora or more generally on Linux
 which would allow me to make a complete snapshot of the system ?
 
 Something as easy of use as the snapshot feature in VirtualBox
 would be great.. But near would do as well.
 
 Kind regards,
 
 -- 
 “One original thought is worth a thousand mindless quotings.”
 “Le vrai n'est pas plus sûr que le probable.”
 
   Diogene Laerce
 
 
 


+--+
  Michael D. Setzer II -  Computer Science Instructor  
  Guam Community College  Computer Center  
  mailto:mi...@kuentos.guam.net
  mailto:msetze...@gmail.com
  http://www.guam.net/home/mikes
  Guam - Where America's Day Begins
  G4L Disk Imaging Project maintainer 
  http://sourceforge.net/projects/g4l/
+--+

http://setiathome.berkeley.edu (Original)
Number of Seti Units Returned:  19,471
Processing time:  32 years, 290 days, 12 hours, 58 minutes
(Total Hours: 287,489)

BOINC@HOME CREDITS
ROSETTA 32682316.626719   |   SETI58701396.531536
ABC 16613838.513356   |   EINSTEIN67783109.164895

-- 
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org

Re: backup snapshot

[Gordon Messmer]

On 08/07/2015 12:56 PM, Diogene Laerce wrote:

Hi,

After a sad experience with a system update, I would like to ask
if there is a software on Fedora or more generally on Linux
which would allow me to make a complete snapshot of the system ?

Something as easy of use as the snapshot feature in VirtualBox
would be great.. But near would do as well.



http://snapper.io/
http://rpm-software-management.github.io/dnf-plugins-extras/snapper.html

--
users mailing list
users@lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org