[gentoo-user] backup hardware setup

[gentoo-user] backup hardware setup

[meino.cramer]

Hi,

I bought two identical external harddrives, USB 3.0, with 1 TByte each
(no SSD - the good ole mechanical ones...;).

The intended use is for backup of longer files. The drives will
contain the same contents.

Currently there are still "clean metal" (no partitioning, no fs).

Data integrity and recoverability (Uhhh...that words looks wrong...) in
case of an desaster is more important than speed.

What is the recommended way of partitioning ?
What filesystem to choose?


Thank you very much in advance for any help!
Best regards,
mcc


PS: Running vanilla kernel 3.15.1....

Re: [gentoo-user] backup hardware setup

[thegeezer]

On 06/24/2014 03:43 PM, meino.cramer@gmx.de wrote:
> Hi,
>
> I bought two identical external harddrives, USB 3.0, with 1 TByte each
> (no SSD - the good ole mechanical ones...;).
>
> The intended use is for backup of longer files. The drives will
> contain the same contents.
>
> Currently there are still "clean metal" (no partitioning, no fs).
>
> Data integrity and recoverability (Uhhh...that words looks wrong...) in
> case of an desaster is more important than speed.
>
> What is the recommended way of partitioning ?
> What filesystem to choose?
>
>
> Thank you very much in advance for any help!
> Best regards,
> mcc
>
>
> PS: Running vanilla kernel 3.15.1....
>
>
>
>

I do this using hard links and rsync to only copy changed data.
this creates a dated folder structure that i can then rsync / cp using a
livecd to baremetal and basically allows best recoverability, imho.
so long as the filesystem supports hard links you are golden.
you might want btrfs for this for long term storage to help in case of
bitrot, but rsync should refresh the file if it is suddenly unreadable
(meaning any other hard lnked versoins are also up the swanny)
ymmv depending on what it is you are backing up

#!/bin/bash
echo 'preparing..'
date=`date "+%Y-%m-%d_%H.%M.%S"`
workingfolder="/mnt/usb/backupsyncs/myhost1"
fromfolder="root@myhost1:/* --exclude=/var/tmp --exclude=/dev
--exclude=/mnt --exclude=/opt --exclude=/proc --exclude=/sys
--exclude=/usr/portage --exclude=/usr/src"

echo "Date " $date
echo "From " $fromfolder
echo "To   " $workingfolder

echo "move current to be dated"
mv $workingfolder/current $workingfolder/backup-$date

echo "now syncing into dated folder"
rsync -vz --partial --modify-window 5 -W --delete -a $fromfolder
$workingfolder/backup-$date

echo "cleaning up..linkcopying dated folder to <current>"
cp -al $workingfolder/backup-$date $workingfolder/current

Re: [gentoo-user] backup hardware setup

[meino.cramer]

thegeezer <thegeezer@thegeezer.net> [14-06-24 17:16]:
> On 06/24/2014 03:43 PM, meino.cramer@gmx.de wrote:
> > Hi,
> >
> > I bought two identical external harddrives, USB 3.0, with 1 TByte each
> > (no SSD - the good ole mechanical ones...;).
> >
> > The intended use is for backup of longer files. The drives will
> > contain the same contents.
> >
> > Currently there are still "clean metal" (no partitioning, no fs).
> >
> > Data integrity and recoverability (Uhhh...that words looks wrong...) in
> > case of an desaster is more important than speed.
> >
> > What is the recommended way of partitioning ?
> > What filesystem to choose?
> >
> >
> > Thank you very much in advance for any help!
> > Best regards,
> > mcc
> >
> >
> > PS: Running vanilla kernel 3.15.1....
> >
> >
> >
> >
> 
> I do this using hard links and rsync to only copy changed data.
> this creates a dated folder structure that i can then rsync / cp using a
> livecd to baremetal and basically allows best recoverability, imho.
> so long as the filesystem supports hard links you are golden.
> you might want btrfs for this for long term storage to help in case of
> bitrot, but rsync should refresh the file if it is suddenly unreadable
> (meaning any other hard lnked versoins are also up the swanny)
> ymmv depending on what it is you are backing up
> 
> #!/bin/bash
> echo 'preparing..'
> date=`date "+%Y-%m-%d_%H.%M.%S"`
> workingfolder="/mnt/usb/backupsyncs/myhost1"
> fromfolder="root@myhost1:/* --exclude=/var/tmp --exclude=/dev
> --exclude=/mnt --exclude=/opt --exclude=/proc --exclude=/sys
> --exclude=/usr/portage --exclude=/usr/src"
> 
> echo "Date " $date
> echo "From " $fromfolder
> echo "To   " $workingfolder
> 
> echo "move current to be dated"
> mv $workingfolder/current $workingfolder/backup-$date
> 
> echo "now syncing into dated folder"
> rsync -vz --partial --modify-window 5 -W --delete -a $fromfolder
> $workingfolder/backup-$date
> 
> echo "cleaning up..linkcopying dated folder to <current>"
> cp -al $workingfolder/backup-$date $workingfolder/current
> 
> 
> 

Hi,

thank you for your reply! :)

...I am sure, whether I want btrfs. On the net I found
for example this:
http://www.phoronix.com/scan.php?page=news_item&px=MTY1MDU

with sentences like:
"The Btrfs file-system changes for the Linux 3.15 kernel mostly deal
with bug fixes and performance fixes while some corruption fixes are
also expected to come."

...sounds a little different to "stable" I think...

What do you think?

Re: [gentoo-user] backup hardware setup

[thegeezer]

On 06/24/2014 04:28 PM, meino.cramer@gmx.de wrote:
> http://www.phoronix.com/scan.php?page=news_item&px=MTY1MDU with
> sentences like: "The Btrfs file-system changes for the Linux 3.15
> kernel mostly deal with bug fixes and performance fixes while some
> corruption fixes are also expected to come." ...sounds a little
> different to "stable" I think... What do you think? 
i hear ya, but if all you are doing is something like code i submitted,
you have no concerns.
even ext4 had an odd corruption bug not too long ago, and that was after
it was stable.
more importantly there is a wide group of folks using btrfs and active
development.
I wouldn't bother with ext2 and ext3 -- fsck will take _forever_ and
heavens help you if you unplug the usb without unmounting

if you are that concerned you might want to make one drive ext4, one
drive btrfs (or another pairing of your choice)
and you can then guard against the filesystem choice by cycling the
disks daily/weekly

Re: [gentoo-user] backup hardware setup

[Rich Freeman]

On Tue, Jun 24, 2014 at 11:28 AM,  <meino.cramer@gmx.de> wrote:
> ...I am sure, whether I want btrfs. On the net I found
> for example this:
> http://www.phoronix.com/scan.php?page=news_item&px=MTY1MDU
>
> with sentences like:
> "The Btrfs file-system changes for the Linux 3.15 kernel mostly deal
> with bug fixes and performance fixes while some corruption fixes are
> also expected to come."
>
> ...sounds a little different to "stable" I think...

I have mixed feelings regarding btrfs.

For a disk that is going to be mostly offline, low-activity, long-term
storage btrfs has some plusses and minuses, especially since you
intend to mirror it (manually or otherwise).

The only real minus is that btrfs is still fairly experimental.  You
could run into problems.  However, what you're doing is a very simple
use case and the write loads are certainly going to be quite low.
Most of the scenarios that cause issues with btrfs are unlikely to
come up.

Btrfs has a few advantages.  I'd say the biggest one is that it
checksums everything and can detect silent corruption.  For a disk
that is just going to sit around for a long time offline that is a big
plus - any other filesystem option isn't going to detect any
corruption to your archive other than unreadable clusters (or disks).
You could get around this by only storing data in a format that
already can detect corruption, and then writing scripts to check
everything, but you'll be manually copying data across your mirror if
you find issues, and that will be really tricky if you're using
something like mdadm since there is no easy way to pick which copy it
gives you.

Btrfs can also mirror your data which guarantees that all of it is
replicated.  Sure, rsync can do this also, but if for whatever reason
something gets changed without updating mtime/ctime/size it won't spot
it unless you have it set to hash everything (which is VERY slow so
nobody does this).  Mdadm would be a better choice, but as I pointed
out it can't detect silent corruption, and is hard to recover if you
discover it yourself.  With btrfs everything is always mirrored (if
you set that up) and a simple scrub command periodically will test to
make sure everything is fine and restore anything that isn't.

I have some old hard drives that I'm using for storage and I use btrfs
on them.  However, ultimately it isn't anything I can't afford to lose
either.

There is nothing wrong with ext4+rsync and maybe an occasional
recursive diff.  You just need to do a bit more with scripting/etc and
make sure you stay on top of it, and watch out for gotchas like having
the wrong disk mounted/etc.  That will be less efficient than mdadm or
btrfs since rsync has no idea what has changed without walking the
whole tree, but it probably isn't a big deal and minus the automation
is probably the most bulletproof option.  Just keep in mind it is only
bulletproof if you don't miss anything when you set it up - this is
relying on you to manage the mirroring/etc and if you drop the ball
then data will be at risk.

So, I won't enthusiastically recommend btrfs, but it may be worth
consideration...

Rich

Re: [gentoo-user] backup hardware setup

[Alan McKinnon]

On 24/06/2014 16:43, meino.cramer@gmx.de wrote:
> Hi,
> 
> I bought two identical external harddrives, USB 3.0, with 1 TByte each
> (no SSD - the good ole mechanical ones...;).
> 
> The intended use is for backup of longer files. The drives will
> contain the same contents.
> 
> Currently there are still "clean metal" (no partitioning, no fs).
> 
> Data integrity and recoverability (Uhhh...that words looks wrong...) in
> case of an desaster is more important than speed.
> 
> What is the recommended way of partitioning ?
> What filesystem to choose?
> 
> 
> Thank you very much in advance for any help!
> Best regards,
> mcc
> 
> 
> PS: Running vanilla kernel 3.15.1....

You haven't given much in the way of detail, so I assume you have
regular needs, nothing fancy, and it's all a bunch of files right?

In that case, partitioning and filesystem type are largely irrelevant as
long as you don't have corruption. With one caveat:

You must always make sure the source drive is intact and ok. If not, and
you back it up anyway, then you are already toast (you will overwrite
your last backup with new faulty data).

There's several approaches to how to do the transfer:

If you have say a general fileserver with lots of files that don't
change much or often, just rsync everything in one go. There is no
optimization you can do that will perform much faster than rsync.

If you have a big busy filesystem that changes often and lots, then use
lvm (or anything that can make snapshots) and rsync that.

If you have a huge database where everything is changing all the time,
don't do filesystem copies, use the tools provided by the db vendor. I
doubt this is your need as you would have said so, but it's worth
mentioning.


-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] backup hardware setup

[meino.cramer]

Alan McKinnon <alan.mckinnon@gmail.com> [14-06-24 19:12]:
> On 24/06/2014 16:43, meino.cramer@gmx.de wrote:
> > Hi,
> > 
> > I bought two identical external harddrives, USB 3.0, with 1 TByte each
> > (no SSD - the good ole mechanical ones...;).
> > 
> > The intended use is for backup of longer files. The drives will
> > contain the same contents.
> > 
> > Currently there are still "clean metal" (no partitioning, no fs).
> > 
> > Data integrity and recoverability (Uhhh...that words looks wrong...) in
> > case of an desaster is more important than speed.
> > 
> > What is the recommended way of partitioning ?
> > What filesystem to choose?
> > 
> > 
> > Thank you very much in advance for any help!
> > Best regards,
> > mcc
> > 
> > 
> > PS: Running vanilla kernel 3.15.1....
> 
> You haven't given much in the way of detail, so I assume you have
> regular needs, nothing fancy, and it's all a bunch of files right?
> 
> In that case, partitioning and filesystem type are largely irrelevant as
> long as you don't have corruption. With one caveat:
> 
> You must always make sure the source drive is intact and ok. If not, and
> you back it up anyway, then you are already toast (you will overwrite
> your last backup with new faulty data).
> 
> There's several approaches to how to do the transfer:
> 
> If you have say a general fileserver with lots of files that don't
> change much or often, just rsync everything in one go. There is no
> optimization you can do that will perform much faster than rsync.
> 
> If you have a big busy filesystem that changes often and lots, then use
> lvm (or anything that can make snapshots) and rsync that.
> 
> If you have a huge database where everything is changing all the time,
> don't do filesystem copies, use the tools provided by the db vendor. I
> doubt this is your need as you would have said so, but it's worth
> mentioning.
> 
> 
> -- 
> Alan McKinnon
> alan.mckinnon@gmail.com
> 
> 

Hi Alan,

thanks for your reply! :)

Yes...your are right. I have a lot static (=not changing) data on my
harddisk...mostly things like video tutorials (blender), videos of 
birds I filmed, dokuments and such...

They are eating up the space on my systems harddisk.

Do I decided to put them on a extern hd and an identical copy on
another identical external harddisk.

Its mainly a task of updateing the data on the external drives with
that what is new (and static and big and falls under what I described
above) on my systems harddisk.

I will check rsync for that!

Re: [gentoo-user] backup hardware setup

[Alan McKinnon]

On 24/06/2014 19:32, meino.cramer@gmx.de wrote:
> Alan McKinnon <alan.mckinnon@gmail.com> [14-06-24 19:12]:
>> On 24/06/2014 16:43, meino.cramer@gmx.de wrote:
>>> Hi,
>>>
>>> I bought two identical external harddrives, USB 3.0, with 1 TByte each
>>> (no SSD - the good ole mechanical ones...;).
>>>
>>> The intended use is for backup of longer files. The drives will
>>> contain the same contents.
>>>
>>> Currently there are still "clean metal" (no partitioning, no fs).
>>>
>>> Data integrity and recoverability (Uhhh...that words looks wrong...) in
>>> case of an desaster is more important than speed.
>>>
>>> What is the recommended way of partitioning ?
>>> What filesystem to choose?
>>>
>>>
>>> Thank you very much in advance for any help!
>>> Best regards,
>>> mcc
>>>
>>>
>>> PS: Running vanilla kernel 3.15.1....
>>
>> You haven't given much in the way of detail, so I assume you have
>> regular needs, nothing fancy, and it's all a bunch of files right?
>>
>> In that case, partitioning and filesystem type are largely irrelevant as
>> long as you don't have corruption. With one caveat:
>>
>> You must always make sure the source drive is intact and ok. If not, and
>> you back it up anyway, then you are already toast (you will overwrite
>> your last backup with new faulty data).
>>
>> There's several approaches to how to do the transfer:
>>
>> If you have say a general fileserver with lots of files that don't
>> change much or often, just rsync everything in one go. There is no
>> optimization you can do that will perform much faster than rsync.
>>
>> If you have a big busy filesystem that changes often and lots, then use
>> lvm (or anything that can make snapshots) and rsync that.
>>
>> If you have a huge database where everything is changing all the time,
>> don't do filesystem copies, use the tools provided by the db vendor. I
>> doubt this is your need as you would have said so, but it's worth
>> mentioning.
>>
>>
>> -- 
>> Alan McKinnon
>> alan.mckinnon@gmail.com
>>
>>
> 
> Hi Alan,
> 
> thanks for your reply! :)
> 
> Yes...your are right. I have a lot static (=not changing) data on my
> harddisk...mostly things like video tutorials (blender), videos of 
> birds I filmed, dokuments and such...
> 
> They are eating up the space on my systems harddisk.
> 
> Do I decided to put them on a extern hd and an identical copy on
> another identical external harddisk.
> 
> Its mainly a task of updateing the data on the external drives with
> that what is new (and static and big and falls under what I described
> above) on my systems harddisk.
> 
> I will check rsync for that!


That changes things just a little bit - I thought your two drives were
going to be one for live and one for backup. Do you intend to move these
files off your main drive onto the identical externals, or just copy the
files?

I would have those two external drives using different filesystems, just
in case as they are your only copy and external drives are fragile in
use and in storage. Exact fs type doesn't really matter - ext4 and xfs,
or ext* and btrfs, it's all good.

Just do make sure you don't use rsync with --delete for this :-)



-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] backup hardware setup

[meino.cramer]

Alan McKinnon <alan.mckinnon@gmail.com> [14-06-24 20:00]:
> On 24/06/2014 19:32, meino.cramer@gmx.de wrote:
> > Alan McKinnon <alan.mckinnon@gmail.com> [14-06-24 19:12]:
> >> On 24/06/2014 16:43, meino.cramer@gmx.de wrote:
> >>> Hi,
> >>>
> >>> I bought two identical external harddrives, USB 3.0, with 1 TByte each
> >>> (no SSD - the good ole mechanical ones...;).
> >>>
> >>> The intended use is for backup of longer files. The drives will
> >>> contain the same contents.
> >>>
> >>> Currently there are still "clean metal" (no partitioning, no fs).
> >>>
> >>> Data integrity and recoverability (Uhhh...that words looks wrong...) in
> >>> case of an desaster is more important than speed.
> >>>
> >>> What is the recommended way of partitioning ?
> >>> What filesystem to choose?
> >>>
> >>>
> >>> Thank you very much in advance for any help!
> >>> Best regards,
> >>> mcc
> >>>
> >>>
> >>> PS: Running vanilla kernel 3.15.1....
> >>
> >> You haven't given much in the way of detail, so I assume you have
> >> regular needs, nothing fancy, and it's all a bunch of files right?
> >>
> >> In that case, partitioning and filesystem type are largely irrelevant as
> >> long as you don't have corruption. With one caveat:
> >>
> >> You must always make sure the source drive is intact and ok. If not, and
> >> you back it up anyway, then you are already toast (you will overwrite
> >> your last backup with new faulty data).
> >>
> >> There's several approaches to how to do the transfer:
> >>
> >> If you have say a general fileserver with lots of files that don't
> >> change much or often, just rsync everything in one go. There is no
> >> optimization you can do that will perform much faster than rsync.
> >>
> >> If you have a big busy filesystem that changes often and lots, then use
> >> lvm (or anything that can make snapshots) and rsync that.
> >>
> >> If you have a huge database where everything is changing all the time,
> >> don't do filesystem copies, use the tools provided by the db vendor. I
> >> doubt this is your need as you would have said so, but it's worth
> >> mentioning.
> >>
> >>
> >> -- 
> >> Alan McKinnon
> >> alan.mckinnon@gmail.com
> >>
> >>
> > 
> > Hi Alan,
> > 
> > thanks for your reply! :)
> > 
> > Yes...your are right. I have a lot static (=not changing) data on my
> > harddisk...mostly things like video tutorials (blender), videos of 
> > birds I filmed, dokuments and such...
> > 
> > They are eating up the space on my systems harddisk.
> > 
> > Do I decided to put them on a extern hd and an identical copy on
> > another identical external harddisk.
> > 
> > Its mainly a task of updateing the data on the external drives with
> > that what is new (and static and big and falls under what I described
> > above) on my systems harddisk.
> > 
> > I will check rsync for that!
> 
> 
> That changes things just a little bit - I thought your two drives were
> going to be one for live and one for backup. Do you intend to move these
> files off your main drive onto the identical externals, or just copy the
> files?
> 
> I would have those two external drives using different filesystems, just
> in case as they are your only copy and external drives are fragile in
> use and in storage. Exact fs type doesn't really matter - ext4 and xfs,
> or ext* and btrfs, it's all good.
> 
> Just do make sure you don't use rsync with --delete for this :-)
> 
> 
> 
> -- 
> Alan McKinnon
> alan.mckinnon@gmail.com
> 
> 

Yes, I will delete the data from my systems drive...

You wrote:
"I would have those two external drives using different filesystems"

Different to what? Different to the fs on the system drive? Both
external drives use different filesystems? All three use different
filesystems?

And how can this help, if the drives are fragile? (I understand
"fragile" as "mechanical not robust" (sorry I am no native english
speaker))

I will use this "mobile disks" not really as the word "mobile" implies. They
will only "travel" manually between a secure place and my PC.
When in use, they will rest on the floor of the room (so they can not
be dropped) and _under_ the case of my PC (ole school big tower metal
case with a gap between the bottom of the case and the floor of the
room.)

Re: [gentoo-user] backup hardware setup

[Alan McKinnon]

On 24/06/2014 20:34, meino.cramer@gmx.de wrote:
> Alan McKinnon <alan.mckinnon@gmail.com> [14-06-24 20:00]:
>> On 24/06/2014 19:32, meino.cramer@gmx.de wrote:
>>> Alan McKinnon <alan.mckinnon@gmail.com> [14-06-24 19:12]:
>>>> On 24/06/2014 16:43, meino.cramer@gmx.de wrote:
>>>>> Hi,
>>>>>
>>>>> I bought two identical external harddrives, USB 3.0, with 1 TByte each
>>>>> (no SSD - the good ole mechanical ones...;).
>>>>>
>>>>> The intended use is for backup of longer files. The drives will
>>>>> contain the same contents.
>>>>>
>>>>> Currently there are still "clean metal" (no partitioning, no fs).
>>>>>
>>>>> Data integrity and recoverability (Uhhh...that words looks wrong...) in
>>>>> case of an desaster is more important than speed.
>>>>>
>>>>> What is the recommended way of partitioning ?
>>>>> What filesystem to choose?
>>>>>
>>>>>
>>>>> Thank you very much in advance for any help!
>>>>> Best regards,
>>>>> mcc
>>>>>
>>>>>
>>>>> PS: Running vanilla kernel 3.15.1....
>>>>
>>>> You haven't given much in the way of detail, so I assume you have
>>>> regular needs, nothing fancy, and it's all a bunch of files right?
>>>>
>>>> In that case, partitioning and filesystem type are largely irrelevant as
>>>> long as you don't have corruption. With one caveat:
>>>>
>>>> You must always make sure the source drive is intact and ok. If not, and
>>>> you back it up anyway, then you are already toast (you will overwrite
>>>> your last backup with new faulty data).
>>>>
>>>> There's several approaches to how to do the transfer:
>>>>
>>>> If you have say a general fileserver with lots of files that don't
>>>> change much or often, just rsync everything in one go. There is no
>>>> optimization you can do that will perform much faster than rsync.
>>>>
>>>> If you have a big busy filesystem that changes often and lots, then use
>>>> lvm (or anything that can make snapshots) and rsync that.
>>>>
>>>> If you have a huge database where everything is changing all the time,
>>>> don't do filesystem copies, use the tools provided by the db vendor. I
>>>> doubt this is your need as you would have said so, but it's worth
>>>> mentioning.
>>>>
>>>>
>>>> -- 
>>>> Alan McKinnon
>>>> alan.mckinnon@gmail.com
>>>>
>>>>
>>>
>>> Hi Alan,
>>>
>>> thanks for your reply! :)
>>>
>>> Yes...your are right. I have a lot static (=not changing) data on my
>>> harddisk...mostly things like video tutorials (blender), videos of 
>>> birds I filmed, dokuments and such...
>>>
>>> They are eating up the space on my systems harddisk.
>>>
>>> Do I decided to put them on a extern hd and an identical copy on
>>> another identical external harddisk.
>>>
>>> Its mainly a task of updateing the data on the external drives with
>>> that what is new (and static and big and falls under what I described
>>> above) on my systems harddisk.
>>>
>>> I will check rsync for that!
>>
>>
>> That changes things just a little bit - I thought your two drives were
>> going to be one for live and one for backup. Do you intend to move these
>> files off your main drive onto the identical externals, or just copy the
>> files?
>>
>> I would have those two external drives using different filesystems, just
>> in case as they are your only copy and external drives are fragile in
>> use and in storage. Exact fs type doesn't really matter - ext4 and xfs,
>> or ext* and btrfs, it's all good.
>>
>> Just do make sure you don't use rsync with --delete for this :-)
>>
>>
>>
>> -- 
>> Alan McKinnon
>> alan.mckinnon@gmail.com
>>
>>
> 
> Yes, I will delete the data from my systems drive...
> 
> You wrote:
> "I would have those two external drives using different filesystems"
> 
> Different to what? Different to the fs on the system drive? Both
> external drives use different filesystems? All three use different
> filesystems?


Different to each other


> And how can this help, if the drives are fragile? (I understand
> "fragile" as "mechanical not robust" (sorry I am no native english
> speaker))

If one drive is say btrfs and the other say ext4 and you hit a
corruption bug in btrfs, then you still have an uncorrupted ext4 copy

> 
> I will use this "mobile disks" not really as the word "mobile" implies. They
> will only "travel" manually between a secure place and my PC.
> When in use, they will rest on the floor of the room (so they can not
> be dropped) and _under_ the case of my PC (ole school big tower metal
> case with a gap between the bottom of the case and the floor of the
> room.)


External drives have a much higher failure rate than internal drives.
people don't expect them to fail or be dropped or accidentally plugged
in in the wrong order and the wrong one to be mkfs'ed (until it does
happen). These are real risks that you can't ignore whereas with a good
internal drive you can often get away with it.

So it only make sense to take sensible precautions that cost you very
little, especially considering these two drives will be your only copy.



-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] backup hardware setup

[Rich Freeman]

On Tue, Jun 24, 2014 at 4:02 PM, Alan McKinnon <alan.mckinnon@gmail.com> wrote:
>
> External drives have a much higher failure rate than internal drives.
> people don't expect them to fail or be dropped or accidentally plugged
> in in the wrong order and the wrong one to be mkfs'ed (until it does
> happen). These are real risks that you can't ignore whereas with a good
> internal drive you can often get away with it.
>

++

Don't ignore the potential for logical errors.  If you have some
script that magically rsyncs stuff then don't make the mistake of
moving data over and rsyncing the old copy over the new, or mounting
the devices in a manner that isn't robust when udev changes all your
device labels, and so on.  That seems like the most likely way your
data is going to get scrambled, unless you have them both in your car
and end up in a crash.

That was one of the reasons I went with btrfs for my offline copy.  If
it unmounts, then I know I have two copies of everything.  If it
mounts, I know it found both mirrors.  If I scrub and there are no
errors, then I know both copies are good.  You can do that in other
ways, but make sure you actually catch the failure modes.

Rich

Re: [gentoo-user] backup hardware setup

[thegeezer]

On 06/24/2014 09:52 PM, Rich Freeman wrote:
> Don't ignore the potential for logical errors. If you have some script
> that magically rsyncs stuff then don't make the mistake of moving data
> over and rsyncing the old copy over the new, or mounting the devices
> in a manner that isn't robust when udev changes all your device
> labels, and so on. That seems like the most likely way your data is
> going to get scrambled, unless you have them both in your car and end
> up in a crash. That was one of the reasons I went with btrfs for my
> offline copy. 
cunning, i like it.
i like it so much i think i will do this myself
> If it unmounts, then I know I have two copies of everything. If it
> mounts, I know it found both mirrors. If I scrub and there are no
> errors, then I know both copies are good. You can do that in other
> ways, but make sure you actually catch the failure modes. Rich