It looks like I have been using this box with mint 17 for about 5 years
and now need to update.

I only use the desktop for browsing and a few documents, spreadsheets
etc so by no means a power user.

In the past I simply took a copy of home and installed a complete new
version is this still the safest way or is there an upgrade path that
retains my current home and installed programs?

If I re install I am thinking of having home on a separate partition on
a new SSD, is it worth keeping the 5 year old hard drive in the machine
for storing bigger files or is there likely to be a reliability problem
due to age?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 10 Nov 2021 12:28:33 +0100, Marco Moock wrote:

If I re install I am thinking of having home on a separate partition on
a new SSD, is it worth keeping the 5 year old hard drive in the machine
for storing bigger files or is there likely to be a reliability problem
due to age?

No, 5 years should be ok, but check SMART parameters. If they are bad, replace the disk.

IME running time, i.e. spun-up hours, are more important than the
calendar age of the drive, so its a good idea to be running smartd
configured to generate a weekly report on your drives current state. This assumes that you're running logwatch and do at least glance at its
reports.

My last disk failures were a couple of drives that both failed at around
50,000 hours. One was a 3.5" WD blue and IIRC the other was a Fujitsu
2.5" drive in a Lenovo laptop.


--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Am Wed, 10 Nov 2021 10:59:28 +0000
schrieb AJH <news@loampitsfarm.co.uk>:

In the past I simply took a copy of home and installed a complete new
version is this still the safest way or is there an upgrade path that
retains my current home and installed programs?

I don't recommend that, only copy the files you know what they are for. Sometimes old configurations files for the desktop environment cause
problems.

If I re install I am thinking of having home on a separate partition
on a new SSD, is it worth keeping the 5 year old hard drive in the
machine for storing bigger files or is there likely to be a
reliability problem due to age?

No, 5 years should be ok, but check SMART parameters. If they are bad,
replace the disk.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Martin Gregorie <martin@mydomain.invalid> wrote:

On Wed, 10 Nov 2021 12:28:33 +0100, Marco Moock wrote:

If I re install I am thinking of having home on a separate partition on
a new SSD, is it worth keeping the 5 year old hard drive in the machine
for storing bigger files or is there likely to be a reliability problem
due to age?

No, 5 years should be ok, but check SMART parameters. If they are bad, replace the disk.

IME running time, i.e. spun-up hours, are more important than the
calendar age of the drive, so its a good idea to be running smartd
configured to generate a weekly report on your drives current state. This assumes that you're running logwatch and do at least glance at its
reports.

I'm not sure about that, isn't the power-up/power-down sequence more
hard work for the drive than just spinning around? I have several
drives that are powered up permanently (except for the occasional
power failure) for ten years and more.

I have had drives fail but not particularly old ones.

--
Chris Green
·

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Am Wed, 10 Nov 2021 13:06:25 +0000
schrieb Chris Green <cl@isbd.net>:

I'm not sure about that, isn't the power-up/power-down sequence more
hard work for the drive than just spinning around? I have several
drives that are powered up permanently (except for the occasional
power failure) for ten years and more.

I have had drives fail but not particularly old ones.

Multiple times of powering up and down also abrades the HDD, but
spinning around is bad for the bearing.
HDDs fail, but some fails can be predicted by bad SMART values.
Especially server operators replace them more often, because the
failure rate increases over the years and small HDDs are cheap
meanwhile.
Drives that were powered up for years (mostly in servers) sometimes
fail after they were powered down for a short time.

HDDs and SSDs can fail anytime, so always have a backup.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 10 Nov 2021 13:06:25 +0000, Chris Green wrote:

Martin Gregorie <martin@mydomain.invalid> wrote:

On Wed, 10 Nov 2021 12:28:33 +0100, Marco Moock wrote:

If I re install I am thinking of having home on a separate partition
on a new SSD, is it worth keeping the 5 year old hard drive in the
machine for storing bigger files or is there likely to be a
reliability problem due to age?

No, 5 years should be ok, but check SMART parameters. If they are
bad,
replace the disk.

IME running time, i.e. spun-up hours, are more important than the
calendar age of the drive, so its a good idea to be running smartd
configured to generate a weekly report on your drives current state.
This assumes that you're running logwatch and do at least glance at its
reports.

I'm not sure about that, isn't the power-up/power-down sequence more
hard work for the drive than just spinning around? I have several
drives that are powered up permanently (except for the occasional power failure) for ten years and more.

Pass, but for sure a drive that's parked isn't wearing out its bearings.

I have had drives fail but not particularly old ones.

So have I - remember the Maxstor drives PC World used to sell? I had
their 20 and 40 MB units back around 2000 and don't remember having any
that lasted more than a year of fairly light duty. Their only saving
grace was that they slowly lost the ability to read data, so fortunately
I always managed to get a final back-up off then - even it it did take
well over an hour, with all the retries, to make that final backup.

After that I switched to WD and, apart from a single bad batch, when I
bought a new disk that failed almost immediately and so did its
replacement, but at least the two replacements turned up fast, direct
from WD, and the last of those was a good 'un that ran 24/7 for years.

However, the point I really wanted to make was to install smartd if you
haven't already, and read its weekly disk status report: when the age-
related performance metrics start to climb, its time to buy a replacement
disk and keep making those backups so, when the drive finally fails your downtime is only the time needed to install the new drive and restore
from your last backup if a dd from old to new drive fails - IIRC that's
always worked for me so far, but thats not going to stop me making
backups!


--
--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 10.11.2021 at 13:06, Chris Green scribbled:

Martin Gregorie <martin@mydomain.invalid> wrote:

On Wed, 10 Nov 2021 12:28:33 +0100, Marco Moock wrote:

If I re install I am thinking of having home on a separate
partition on a new SSD, is it worth keeping the 5 year old hard
drive in the machine for storing bigger files or is there likely
to be a reliability problem due to age?

No, 5 years should be ok, but check SMART parameters. If they are
bad, replace the disk.

IME running time, i.e. spun-up hours, are more important than the
calendar age of the drive, so its a good idea to be running smartd configured to generate a weekly report on your drives current
state. This assumes that you're running logwatch and do at least
glance at its reports.

I'm not sure about that, isn't the power-up/power-down sequence more
hard work for the drive than just spinning around?

Yes, it is. And this machine here also has an old SATA2 HDD in it
which is going on 5 years — knocking on wood — but I do keep the
machine running 24/7, and so the drive rarely ever needs to spin up or
down.

For that matter, said drive was actually a replacement under warranty
because the original HDD in the machine that it came out of broke down
after only 6 weeks.

I have several drives that are powered up permanently (except for the occasional power failure) for ten years and more.

I have had drives fail but not particularly old ones.

Same thing here. The only old drive that ever failed on me was still of
the PATA variety — it might have been a Quantum but I'm not sure
anymore. All other drive failures I've had were with new drives, and
they all went belly up within one or two months after acquiring the
machine.

What you definitely want to stay away from are the WD Green drives, if
they're still making those. They spin down and back up all of the
time, and they don't last for very long. The WD Blue and Black drives
are okay, though.

--
With respect,
= Aragorn =

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 10 Nov 2021 22:10:15 +0100, Aragorn wrote:

What you definitely want to stay away from are the WD Green drives, if they're still making those. They spin down and back up all of the time,
and they don't last for very long. The WD Blue and Black drives are
okay, though.

I've, fortunately or so it seems, never looked at a Green drive, but I
have wondered about getting a WD Red for my house server, which runs
24/7. Have you tried that colour?

Currently all my systems are on WD Blue, apart from an old Lenovo R61i,
which has a Sanyo 128GB SSD fitted. Before you ask, the R61i's disk
interface is unable to talk to a disk of more than 200GB. Its original
disk was 128GB and when that died the smallest HDD I could buy was 320GB.
I tried a 500GB disk in it: no dice, but 128GB SSDs were around and
that's what it got. Puugged in, worked immediately.

I'm pleased with that. Currently the machine is running 47/7 doing
protein 3D shape fitting. The trimmer currently reports tidying up about
2GB of blocks per week, so evidently the protein mangler uses more disk
work space than I'd imagines it would. But, the oddity is that in all my HDD-based systems, smartd reports around 40-45 hours of spun-up time per
week for this T440 and maybe 50-60 hours per week for the for the always-
on house server, but the SSD-equipped R61i which runs the protein folding
stuff 24/7 never reports a weekly 'spun up' time of more than 1-2 hours.

My best guess is that this is due to the SSD reporting that its parked
whenever its not actually reading or writing, but that's just a guess:
what do I know about SSDs?


--
--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 10.11.2021 at 21:52, Martin Gregorie scribbled:

On Wed, 10 Nov 2021 22:10:15 +0100, Aragorn wrote:

What you definitely want to stay away from are the WD Green drives,
if they're still making those. They spin down and back up all of
the time, and they don't last for very long. The WD Blue and Black
drives are okay, though.

I've, fortunately or so it seems, never looked at a Green drive, but
I have wondered about getting a WD Red for my house server, which
runs 24/7. Have you tried that colour?

Not personally, but I've heard good things about them. They seem to be
quite reliable, from what I hear.

Currently the machine is running 47/7 doing protein 3D shape fitting.

Ah, the time dilation thing. :p I'm not very good at that; the
gravity well gets me every time, so I can't manage more than 24 hours in
a day. But there's hope, though: Earth's rotation is slowing down.

After all, the dinosaurs had to get by with only 20 hours in a day, and
they had to put in as much work as we do in order not to get fired.
No, wait...

:p

My best guess is that this is due to the SSD reporting that its
parked whenever its not actually reading or writing, but that's just
a guess: what do I know about SSDs?

It's probably due to that, yes. The absence of moving parts makes
switching between the different power settings of an SSD negligible in
terms of performance, so when they are idle, they appear "spun down" to
a regular SAS/SATA interface.

--
With respect,
= Aragorn =

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Am Wed, 10 Nov 2021 22:10:15 +0100
schrieb Aragorn <thorongil@telenet.be>:

Same thing here. The only old drive that ever failed on me was still
of the PATA variety — it might have been a Quantum but I'm not sure anymore. All other drive failures I've had were with new drives, and
they all went belly up within one or two months after acquiring the
machine.

I had many drive fails.
10 year old WD 1200 SATA, many Seagate Momentus laptop drives (not even
10 years old) and many PATA drive fails (Maxtor DiamondMax 9 Plus in
2021, 2 Hitachi Travelstar in 2014/2015 (both IDE).

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 10/11/2021 11:28, Marco Moock wrote:

Am Wed, 10 Nov 2021 10:59:28 +0000
schrieb AJH <news@loampitsfarm.co.uk>:

In the past I simply took a copy of home and installed a complete new
version is this still the safest way or is there an upgrade path that
retains my current home and installed programs?

I don't recommend that, only copy the files you know what they are for. Sometimes old configurations files for the desktop environment cause problems.

If I re install I am thinking of having home on a separate partition
on a new SSD, is it worth keeping the 5 year old hard drive in the
machine for storing bigger files or is there likely to be a
reliability problem due to age?

No, 5 years should be ok, but check SMART parameters. If they are bad, replace the disk.

Thanks for that. sda passes a basic smart check and no faults.

I'll save user files and re install

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 11/11/2021 06:24, Marco Moock wrote:

Am Wed, 10 Nov 2021 22:10:15 +0100
schrieb Aragorn <thorongil@telenet.be>:

Same thing here. The only old drive that ever failed on me was still
of the PATA variety — it might have been a Quantum but I'm not sure
anymore. All other drive failures I've had were with new drives, and
they all went belly up within one or two months after acquiring the
machine.

I had many drive fails.
10 year old WD 1200 SATA, many Seagate Momentus laptop drives (not even
10 years old) and many PATA drive fails (Maxtor DiamondMax 9 Plus in
2021, 2 Hitachi Travelstar in 2014/2015 (both IDE).

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Am Fri, 12 Nov 2021 13:51:04 +0100
schrieb Steve <steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Am Fri, 12 Nov 2021 13:29:09 +0000
schrieb Chris Green <cl@isbd.net>:

... and when your disk gets corrupted it backs up the corrupted data?

If the data is corrupted it backs up corrupted data. If the sectors
can't be read it can't even read the file. But I keep more than 1
version of a file to avoid that and also do full backups sometimes.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Marco Moock <mo01@posteo.de> wrote:

Am Fri, 12 Nov 2021 13:51:04 +0100
schrieb Steve <steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

... and when your disk gets corrupted it backs up the corrupted data?

--
Chris Green
·

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/11/2021 14:29, Chris Green wrote:

Marco Moock <mo01@posteo.de> wrote:

Am Fri, 12 Nov 2021 13:51:04 +0100
schrieb Steve <steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

... and when your disk gets corrupted it backs up the corrupted data?

If you have automatic backups with multiple copies (e.g. once per day,
keep them for 30 days) then you can restore to the backup you took
before the corruption occurred. Veeam will do that for you.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Steve <steve@nospam.invalid> wrote:

On 12/11/2021 14:29, Chris Green wrote:

Marco Moock <mo01@posteo.de> wrote:

Am Fri, 12 Nov 2021 13:51:04 +0100
schrieb Steve <steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

... and when your disk gets corrupted it backs up the corrupted data?

If you have automatic backups with multiple copies (e.g. once per day,
keep them for 30 days) then you can restore to the backup you took
before the corruption occurred. Veeam will do that for you.

But something has to notice the corrupted file/disk for this to work.

If, say, a chunk of your photo archive gets corrupted in some way are
you sure that you'd notice within 30 days? I don't think I would, I
certainly don't scan through my 20000 plus images that frequently! :-)

I do however keep incremental backups (of the important things like my
photo archives) for *much* longer than 30 days. I have them back for
over a year on the current backup system and for several years on an
older NAS which is now retired (but did work when I powered it up a
few months ago).

--
Chris Green
·

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Fri, 12 Nov 2021 18:35:03 +0100, Steve wrote:

On 12/11/2021 17:09, Chris Green wrote:

Steve <steve@nospam.invalid> wrote:

On 12/11/2021 14:29, Chris Green wrote:

Marco Moock <mo01@posteo.de> wrote:

Am Fri, 12 Nov 2021 13:51:04 +0100 schrieb Steve
<steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

... and when your disk gets corrupted it backs up the corrupted data?

If you have automatic backups with multiple copies (e.g. once per day,
keep them for 30 days) then you can restore to the backup you took
before the corruption occurred. Veeam will do that for you.

But something has to notice the corrupted file/disk for this to work.

If, say, a chunk of your photo archive gets corrupted in some way are
you sure that you'd notice within 30 days? I don't think I would, I
certainly don't scan through my 20000 plus images that frequently! :-)

I do however keep incremental backups (of the important things like my
photo archives) for *much* longer than 30 days. I have them back for
over a year on the current backup system and for several years on an
older NAS which is now retired (but did work when I powered it up a few
months ago).

Good plan. My 30 days was just an example, nothing else.

Just don't put them all on the same backup drive. However, a cycle of two
or three backup disks is enough if they're stored offline, preferably in
a firesafe or another building, and only one od out of safe storage at
the time.

Yes, I know I have 4 weeks worth of rsnapshot backups on one disk that
lives next to the computer it backs up, but thats really just fat finger protection: my cycle of weekly backups are kept in a firesafe and only
one is ever outside the firesafe at a time.


--
--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Fri, 12 Nov 2021 16:09:28 +0000, Chris Green wrote:

But something has to notice the corrupted file/disk for this to work.

If, say, a chunk of your photo archive gets corrupted in some way are
you sure that you'd notice within 30 days? I don't think I would, I certainly don't scan through my 20000 plus images that frequently! :-)

That's easy. I have a cron job that:
- mounts the USB backup disk
- stops Postgresql
- makes a backup using rsnapshot
- restarts Postgresql
- runs "fsck -n " against the backup disk
- unmounts the backup disk
- collects a log of everything written to stdout and stderr
while the script was running
- emails the log to myself as its last action.

So, I'll know about any disk problems as soon as I read my mail next
morning and the output from the last but one rsnapshot run should be
clean.

So, just write a similar script that runs whatever your preferred backup program may be. I formerly used compressed tar backups, but when that run exceeded 3 hours each night, I switched to using rsnapshot, which only
takes 10 minutes.



--
--
Martin | martin at
Gregorie | gregorie dot org

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/11/2021 17:09, Chris Green wrote:

Steve <steve@nospam.invalid> wrote:

On 12/11/2021 14:29, Chris Green wrote:

Marco Moock <mo01@posteo.de> wrote:

Am Fri, 12 Nov 2021 13:51:04 +0100
schrieb Steve <steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

... and when your disk gets corrupted it backs up the corrupted data?

If you have automatic backups with multiple copies (e.g. once per day,
keep them for 30 days) then you can restore to the backup you took
before the corruption occurred. Veeam will do that for you.

But something has to notice the corrupted file/disk for this to work.

If, say, a chunk of your photo archive gets corrupted in some way are
you sure that you'd notice within 30 days? I don't think I would, I certainly don't scan through my 20000 plus images that frequently! :-)

I do however keep incremental backups (of the important things like my
photo archives) for *much* longer than 30 days. I have them back for
over a year on the current backup system and for several years on an
older NAS which is now retired (but did work when I powered it up a
few months ago).

Good plan. My 30 days was just an example, nothing else.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Martin Gregorie <martin@mydomain.invalid> wrote:

On Fri, 12 Nov 2021 18:35:03 +0100, Steve wrote:

On 12/11/2021 17:09, Chris Green wrote:

Steve <steve@nospam.invalid> wrote:

On 12/11/2021 14:29, Chris Green wrote:

Marco Moock <mo01@posteo.de> wrote:

Am Fri, 12 Nov 2021 13:51:04 +0100 schrieb Steve
<steve@nospam.invalid>:

Disk failures will happen at a time you least want them to.
Backup your data regularly. I recommend this:

https://www.veeam.com/linux-backup-free.html

I let Deja Dup create a backup daily on my other hard disk.

... and when your disk gets corrupted it backs up the corrupted data? >>>>

If you have automatic backups with multiple copies (e.g. once per day, >>> keep them for 30 days) then you can restore to the backup you took
before the corruption occurred. Veeam will do that for you.

But something has to notice the corrupted file/disk for this to work.

If, say, a chunk of your photo archive gets corrupted in some way are
you sure that you'd notice within 30 days? I don't think I would, I
certainly don't scan through my 20000 plus images that frequently! :-)

I do however keep incremental backups (of the important things like my
photo archives) for *much* longer than 30 days. I have them back for
over a year on the current backup system and for several years on an
older NAS which is now retired (but did work when I powered it up a few
months ago).

Good plan. My 30 days was just an example, nothing else.

Just don't put them all on the same backup drive. However, a cycle of two
or three backup disks is enough if they're stored offline, preferably in
a firesafe or another building, and only one od out of safe storage at
the time.

Yes, my main backup system is out in the garage which is 30 metres or
so from the house.

Yes, I know I have 4 weeks worth of rsnapshot backups on one disk that
lives next to the computer it backs up, but thats really just fat finger protection: my cycle of weekly backups are kept in a firesafe and only
one is ever outside the firesafe at a time.

Yes, I do that as well, very handy (as you say) for "fat finger
protection". :-)

--
Chris Green
·

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)