Re: Advice on upgrading to SSD

[rhkramer]

On Thursday, March 05, 2020 01:56:27 PM Michael Stone wrote:
> On Thu, Mar 05, 2020 at 10:11:55AM -0500, rhkra...@gmail.com wrote:
> >On Thursday, March 05, 2020 03:27:09 AM to...@tuxteam.de wrote:
> >> On Wed, Mar 04, 2020 at 06:53:12PM -0500, Gene Heskett wrote:
> >> 
> >> [...]
> >> 
> >> > That is lieing to the user, and not nice at all.
> >> 
> >> Disable your caches! now!
> >> 
> >> What? Your computer is just 1/3 as fast as it used to be?
> >> 
> >> Computing is about lying -- and trying to not get caught.
> >
> >Not always, or not in this respect.  There are applications where, in
> >essence, a real time kernel is (leading to quick responses to real world
> >events or similar are) necessary, sometimes for the protection of life or
> >limb (as Gene has alluded to).
> 
> nice red herring, but those aren't going to be dependent on a commodity
> hard drive unless someone really screwed up.

What kind of hard drive would it use?  

Was this discussion limited to commodity hard drives?

Re: Subject: Re: Advice on upgrading to SSD

[David Wright]

On Thu 05 Mar 2020 at 18:01:57 (+), G.W. Haywood wrote:
> On Thu, 5 Mar 2020, 0...@caiway.net wrote:
> > > -dsr- wrote:
> > >
> > > There used to be (still?) a set of Western Digital drives that
> > > would go into a hard sleep and park their heads repeatedly. This
> > > tended to cause a shortened lifespan as well as terrible
> > > performance.
> > > with cron:
> > # prevent disks from sleeping, every minute:
> > * * *  **  /bin/touch /dev/sda &>/
> 
> No, I don't think so.  The discs in question I believe by default
> sleep after eight seconds and play particularly badly with Linux.

Is that what's discussed in   man hdparm   under option -J?

Cheers,
David.

Re: Advice on upgrading to SSD

[Michael Stone]

On Thu, Mar 05, 2020 at 10:11:55AM -0500, rhkra...@gmail.com wrote:

On Thursday, March 05, 2020 03:27:09 AM to...@tuxteam.de wrote:

On Wed, Mar 04, 2020 at 06:53:12PM -0500, Gene Heskett wrote:

[...]

> That is lieing to the user, and not nice at all.

Disable your caches! now!

What? Your computer is just 1/3 as fast as it used to be?

Computing is about lying -- and trying to not get caught.


Not always, or not in this respect.  There are applications where, in essence,
a real time kernel is (leading to quick responses to real world events or
similar are) necessary, sometimes for the protection of life or limb (as Gene
has alluded to).


nice red herring, but those aren't going to be dependent on a commodity 
hard drive unless someone really screwed up.

Subject: Re: Advice on upgrading to SSD

[G.W. Haywood]

Hi there,

On Thu, 5 Mar 2020, 0...@caiway.net wrote:


> -dsr- wrote:
>
> There used to be (still?) a set of Western Digital drives that
> would go into a hard sleep and park their heads repeatedly. This 
> tended to cause a shortened lifespan as well as terrible

> performance.
> 
with cron:

# prevent disks from sleeping, every minute:
* * *  **  /bin/touch /dev/sda &>/


No, I don't think so.  The discs in question I believe by default
sleep after eight seconds and play particularly badly with Linux.

--

73,
Ged.

Re: Advice on upgrading to SSD

[Klaus Singvogel]

David Wright wrote:
> > Wha not using "hdparm"?
> 
> Would you be more specific and include the options you are suggesting.
> There are over 60 available, some dangerous.

I don't use hdparm anymore and can't memorize.
But I'd gone to Google, and they tell me: "hdparm -I /dev/sdX -S 0"

Regards,
Klaus.
-- 
Klaus Singvogel
GnuPG-Key-ID: 1024R/5068792D  1994-06-27

Re: Advice on upgrading to SSD

[Tixy]

On Thu, 2020-03-05 at 11:18 -0600, David Wright wrote:
> On Thu 05 Mar 2020 at 08:26:58 (+0100), Klaus Singvogel wrote:
> > 0...@caiway.net wrote:
> > > with cron:
> > > # prevent disks from sleeping, every minute:
> > > * * *  **  /bin/touch /dev/sda &>/
> > 
> > Wha not using "hdparm"?
> 
> Would you be more specific and include the options you are
> suggesting.
> There are over 60 available, some dangerous.
> 

I'm assuming he means -S which sets the standby (spindown) time‐out for
the drive. I use it to force my NAS drives to spin down so the noise
doesn't annoy me. But a timeout of zero will disable spindown.

-- 
Tixy

Re: Advice on upgrading to SSD

[David Wright]

On Thu 05 Mar 2020 at 08:26:58 (+0100), Klaus Singvogel wrote:
> 0...@caiway.net wrote:
> > with cron:
> > # prevent disks from sleeping, every minute:
> > * * *  **  /bin/touch /dev/sda &>/
> 
> Wha not using "hdparm"?

Would you be more specific and include the options you are suggesting.
There are over 60 available, some dangerous.

Cheers,
David.

Re: Advice on upgrading to SSD

[rhkramer]

On Thursday, March 05, 2020 03:27:09 AM to...@tuxteam.de wrote:
> On Wed, Mar 04, 2020 at 06:53:12PM -0500, Gene Heskett wrote:
> 
> [...]
> 
> > That is lieing to the user, and not nice at all.
> 
> Disable your caches! now!
> 
> What? Your computer is just 1/3 as fast as it used to be?
> 
> Computing is about lying -- and trying to not get caught.

Not always, or not in this respect.  There are applications where, in essence, 
a real time kernel is (leading to quick responses to real world events or 
similar are) necessary, sometimes for the protection of life or limb (as Gene 
has alluded to).

Re: Advice on upgrading to SSD

[tomas]

On Wed, Mar 04, 2020 at 06:53:12PM -0500, Gene Heskett wrote:

[...]

> That is lieing to the user, and not nice at all.

Disable your caches! now!

What? Your computer is just 1/3 as fast as it used to be?

Computing is about lying -- and trying to not get caught.

Cheers
-- t


signature.asc
Description: Digital signature

Re: Advice on upgrading to SSD

[Klaus Singvogel]

0...@caiway.net wrote:
> with cron:
> # prevent disks from sleeping, every minute:
> * * *  **  /bin/touch /dev/sda &>/

Wha not using "hdparm"?

Regards,
Klaus.
-- 
Klaus Singvogel
GnuPG-Key-ID: 1024R/5068792D  1994-06-27

Re: Advice on upgrading to SSD

[0...@caiway.net]

> There used to be (still?) a set of Western Digital drives that
> would go into a hard sleep and park their heads repeatedly. This 
> tended to cause a shortened lifespan as well as terrible
> performance.
> 
> -dsr-
> 
with cron:
# prevent disks from sleeping, every minute:
* * *  **  /bin/touch /dev/sda &>/

Re: Advice on upgrading to SSD

[Gene Heskett]

On Wednesday 04 March 2020 22:26:34 Michael Stone wrote:

> On Wed, Mar 04, 2020 at 06:53:12PM -0500, Gene Heskett wrote:
> >On Wednesday 04 March 2020 18:49:06 Michael Stone wrote:
> >> On Wed, Mar 04, 2020 at 06:39:51PM -0500, Gene Heskett wrote:
> >> >On Wednesday 04 March 2020 12:48:57 rhkra...@gmail.com wrote:
> >> >>* I suspect most of us don't know whether our disks are
> >> >> spinning 24/7 -- some of my computers are up 24/7, but I suspect
> >> >> the disks "spin down" when unused for some period of time
> >> >
> >> >Would that not be logged as an access error, followed by a
> >> > reset/restart in the logs?  At the very minimum it should be
> >> > logged so that the user can know the drive is doing that.
> >>
> >> no, not unless something is horribly broken
> >
> >That is lieing to the user, and not nice at all.
>
> Why on earth would you expect an error? That's just nuts.

No its not. If the disk is asked for data from such and such a sector in 
order to issue a stop command to a machine, IMO it has its seek time + 1 
revolution to get that data under a head and read. If its spun down then 
it will have to also spin the drive up which may take 2 or 3 seconds.  
That IMNSHO is a huge error that in some scenarios, could get somebody 
maimed or killed.

So we'll have to agree to disagree.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 

Re: Advice on upgrading to SSD

[Michael Stone]

On Wed, Mar 04, 2020 at 06:53:12PM -0500, Gene Heskett wrote:

On Wednesday 04 March 2020 18:49:06 Michael Stone wrote:


On Wed, Mar 04, 2020 at 06:39:51PM -0500, Gene Heskett wrote:
>On Wednesday 04 March 2020 12:48:57 rhkra...@gmail.com wrote:
>>* I suspect most of us don't know whether our disks are spinning
>> 24/7 -- some of my computers are up 24/7, but I suspect the disks
>> "spin down" when unused for some period of time
>
>Would that not be logged as an access error, followed by a
> reset/restart in the logs?  At the very minimum it should be logged
> so that the user can know the drive is doing that.

no, not unless something is horribly broken


That is lieing to the user, and not nice at all.


Why on earth would you expect an error? That's just nuts.

Re: Advice on upgrading to SSD

[David Christensen]

On 2020-03-04 06:47, Tony van der Hoff wrote:
I guess I'll go with s single SSD onto which I'll copy the data from the 
RAID.


I would advise disconnecting the RAID, doing a fresh install onto the 
SSD, updating and configuring to suit, reconnecting the RAID, and then 
copying data.



Be sure to backup, archive, and image at key points before, during, and 
after the conversion.




Thank you all for helping me make up my mind.


YW.  :-)


David

Re: Advice on upgrading to SSD

[Gene Heskett]

On Wednesday 04 March 2020 18:49:06 Michael Stone wrote:

> On Wed, Mar 04, 2020 at 06:39:51PM -0500, Gene Heskett wrote:
> >On Wednesday 04 March 2020 12:48:57 rhkra...@gmail.com wrote:
> >>* I suspect most of us don't know whether our disks are spinning
> >> 24/7 -- some of my computers are up 24/7, but I suspect the disks
> >> "spin down" when unused for some period of time
> >
> >Would that not be logged as an access error, followed by a
> > reset/restart in the logs?  At the very minimum it should be logged
> > so that the user can know the drive is doing that.
>
> no, not unless something is horribly broken

That is lieing to the user, and not nice at all.


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 

Re: Advice on upgrading to SSD

[Michael Stone]

On Wed, Mar 04, 2020 at 06:39:51PM -0500, Gene Heskett wrote:

On Wednesday 04 March 2020 12:48:57 rhkra...@gmail.com wrote:

   * I suspect most of us don't know whether our disks are spinning
24/7 -- some of my computers are up 24/7, but I suspect the disks
"spin down" when unused for some period of time


Would that not be logged as an access error, followed by a reset/restart
in the logs?  At the very minimum it should be logged so that the user
can know the drive is doing that.


no, not unless something is horribly broken

Re: Advice on upgrading to SSD

[Gene Heskett]

On Wednesday 04 March 2020 12:48:57 rhkra...@gmail.com wrote:

> On Wednesday, March 04, 2020 09:47:47 AM Tony van der Hoff wrote:
> > Gene didn't address my problem, but made the very useful observation
> > that disks spinning 24/7 don't really die. Perhaps I shouldn't worry
> > about replacing them.
>
> I wouldn't count on that, for two reasons:
>
>* everything mechanical wears, and eventually wears out (well,
> maybe not if the bearings are a magnetic suspension type) (no moving
> parts in contact)
>
>* I suspect most of us don't know whether our disks are spinning
> 24/7 -- some of my computers are up 24/7, but I suspect the disks
> "spin down" when unused for some period of time

Would that not be logged as an access error, followed by a reset/restart 
in the logs?  At the very minimum it should be logged so that the user 
can know the drive is doing that.

The other item that would look identical in the logs is a hot red sata 
drive cable that is failing, which a hot red cable will do, often before 
5 years has elapsed. That particular dye has been known to be a problem 
child since CB radios were imported from the J.A.Pan Company in the 
early 1970's.

The red wire in the microphone cables was always the first to fail.  You 
could pour the copper out of the cut off end of the red wire as a rust 
colored powder.  No matter how much of it you cut off looking for good 
copper to resolder a good connection.  That same plastic dye is still 
destroying hot red sata drive cables. Test by moving the middle of the 
cable while the machine is running with a tail on the log.  If the log 
blows up, replace that cable, preferably with a different color.

In any event I would be watching such a disk with an eye toward 
warrantying it before the warranty expires if a smartctl command to 
disable that "feature" is not available or effective at stopping such 
drive killing non-sense. I understand that W.D. Greens do that so have 
never considered them for purchase.

Do I miss-understand their advertising?

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 

Re: Advice on upgrading to SSD

[Cindy Sue Causey]

On 3/4/20, Dan Ritter  wrote:
> rhkra...@gmail.com wrote:
>> some of my computers are up 24/7, but I suspect the disks "spin down" when
>>
>> unused for some period of time
>
> Some do, some don't. Drives sold as "laptop" or "green" drives
> are more likely to do so by themselves.
>
> There used to be (still?) a set of Western Digital drives that
> would go into a hard sleep and park their heads repeatedly. This
> tended to cause a shortened lifespan as well as terrible
> performance.


I remember seeing AND possibly sharing here on Debian-User about a
BIOS setting that let the User decide whether to let disks like that
make that decision on their own..

SOMETHING like that, anyway.. My Dell PC power supply started working
again.. and then finally fritzed for good so I can't go in to check
exactly what it said. I THINK that's the computer that had that
CHOICE.. Dell XPS 420 grin.

Cindy :)
-- 
Cindy-Sue Causey
Talking Rock, Pickens County, Georgia, USA

* runs with birdseed *

Re: Advice on upgrading to SSD

[Nate Bargmann]

* On 2020 04 Mar 10:02 -0600, Michael Stone wrote:
> On Wed, Mar 04, 2020 at 02:47:47PM +, Tony van der Hoff wrote:
> > Gene didn't address my problem, but made the very useful observation
> > that disks spinning 24/7 don't really die. Perhaps I shouldn't worry
> > about replacing them.
> 
> the speed advantages are such that I try to avoid spinning disks for OSs
> wherever possible these days since the costs for SSDs have gotten so low.

This.

I updated two desktops last week from spinning rust to SSDs, including
this machine, and the speed increase, lack of noise, much lower heat
inside the case, and now much lower price point makes the upgrade a no
brainer.

> > Andrei supports file-level copying, so I'll stick to that.
> 
> Yes, this is generally how I'd proceed. rsync existing filesystem onto the
> new one, bind mount /dev /proc and /sys to the new filesystem, then chroot
> into it, edit fstab on the new filesystem to point to the new root (I'd
> personally use the UUID) run grub-install to put the bootloader on the new
> disk, and then run update-grub. There are other approaches to setting up the
> bootloader as well. It's best to do all this in single user mode to avoid
> files changing while you're copying. Then you'll want to change the bios
> settings to boot off the ssd first rather than the hard drives.

I like rsync for this.  In the olde days I often updated over the LAN to
copy data from a laptop to a desktop and back to the laptop's new drive.
Years ago I bought a 500 GB USB drive that saved that step.  As rsync
gives compression with the '-z' option, doing so over the network was
tolerable.

I had to do a bit more dancing on this machine where I use secure boot
with UEFI.  I followed the steps from:

https://wiki.debian.org/GrubEFIReinstall 

including putting the suggested Refind image on a thumb drive and
booting to it after disabling secure boot but via EFI.  That was easy,
the next day I updated the BIOS which erased all of the user keys so I
had to do those steps over and ensure the Grub key was in NVRAM along
with the key I use to sign the Virtualbox kernel modules.

It was a learning experience!

- Nate

-- 

"The optimist proclaims that we live in the best of all
possible worlds.  The pessimist fears this is true."

Web: https://www.n0nb.us
Projects: https://github.com/N0NB
GPG fingerprint: 82D6 4F6B 0E67 CD41 F689 BBA6 FB2C 5130 D55A 8819



signature.asc
Description: PGP signature

Re: Advice on upgrading to SSD

[Dan Ritter]

rhkra...@gmail.com wrote: 
> On Wednesday, March 04, 2020 09:47:47 AM Tony van der Hoff wrote:
> > Gene didn't address my problem, but made the very useful observation
> > that disks spinning 24/7 don't really die. Perhaps I shouldn't worry
> > about replacing them.
> 
> I wouldn't count on that, for two reasons:
> 
>* everything mechanical wears, and eventually wears out (well, maybe not 
> if 
> the bearings are a magnetic suspension type) (no moving parts in contact)
> 
>* I suspect most of us don't know whether our disks are spinning 24/7 -- 
> some of my computers are up 24/7, but I suspect the disks "spin down" when 
> unused for some period of time

Some do, some don't. Drives sold as "laptop" or "green" drives
are more likely to do so by themselves.

There used to be (still?) a set of Western Digital drives that
would go into a hard sleep and park their heads repeatedly. This 
tended to cause a shortened lifespan as well as terrible
performance.

-dsr-

Re: Advice on upgrading to SSD

[rhkramer]

On Wednesday, March 04, 2020 09:47:47 AM Tony van der Hoff wrote:
> Gene didn't address my problem, but made the very useful observation
> that disks spinning 24/7 don't really die. Perhaps I shouldn't worry
> about replacing them.

I wouldn't count on that, for two reasons:

   * everything mechanical wears, and eventually wears out (well, maybe not if 
the bearings are a magnetic suspension type) (no moving parts in contact)

   * I suspect most of us don't know whether our disks are spinning 24/7 -- 
some of my computers are up 24/7, but I suspect the disks "spin down" when 
unused for some period of time

Re: Advice on upgrading to SSD

[G.W. Haywood]

Hi there,

On Wed, 4 Mar 2020, grumpy wrote:


if new drive is the same size or larger
install new drive
boot using usb drive, i use systemrescuecd
dd if=old_drive of=new_drive


If there might be any problem with readability on source disc sectors,
then 'ddrescue' is far superior to plain old 'dd'.

Be aware that there is more than one version of ddrescue, and I've
seen at least one version which was less than reliable.  From personal
experience the version packaged by Debian seems to be a good version.

--

73,
Ged.

Re: Advice on upgrading to SSD

[Michael Stone]

On Wed, Mar 04, 2020 at 02:47:47PM +, Tony van der Hoff wrote:
Gene didn't address my problem, but made the very useful observation 
that disks spinning 24/7 don't really die. Perhaps I shouldn't worry 
about replacing them.


the speed advantages are such that I try to avoid spinning disks for OSs
wherever possible these days since the costs for SSDs have gotten so 
low. 


Andrei supports file-level copying, so I'll stick to that.


Yes, this is generally how I'd proceed. rsync existing filesystem onto 
the new one, bind mount /dev /proc and /sys to the new filesystem, then 
chroot into it, edit fstab on the new filesystem to point to the new 
root (I'd personally use the UUID) run grub-install to put the 
bootloader on the new disk, and then run update-grub. There are other 
approaches to setting up the bootloader as well. It's best to do all 
this in single user mode to avoid files changing while you're copying. 
Then you'll want to change the bios settings to boot off the ssd first 
rather than the hard drives.

Re: Advice on upgrading to SSD

[grumpy]

On Wed, 4 Mar 2020, Tony van der Hoff wrote:


On Monday 02 March 2020 06:28:58 Tony van der Hoff wrote:


Hi,
I'm currently running Buster on a 5 year old GigaByte motherboard with
a 10-year old Raid-1 array on 2 500GB disks. Although it is running
fine, I'm becoming a bit concerned about the longevity of this
storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
from Crucial.

My plan would be to install the SSD in the cage, and dd the contents
of the array onto the SSD. I would then change the BIOS to boot from
the SSD, making the RAID array redundant.

Can someone please tell me whether this plan is feasible, and what
pitfalls I might encounter?



Thanks to all who replied.

Gene didn't address my problem, but made the very useful observation that 
disks spinning 24/7 don't really die. Perhaps I shouldn't worry about 
replacing them.


Dan gave a useful step-by-step procedure for copying at file level. This 
method is slightly problematical for me as it includes "make changes to new 
/etc/fstab" and "make changes to bootloader to set new root filesystem", 
neither of which I feel feel comfortable with, due to lack of knowledge. I'll 
certainly abandon my plan to dd the whole filesystem.


Basti and Deloptes propose to add the SSD as a third drive to the RAID. Very 
attractive, but would involve complications if the SSD is not identical in 
size to the old spinners.


Andrei supports file-level copying, so I'll stick to that.

Joe supports the idea of adding the drive to the RAID, but makes no reference 
to dealing wth the size difference.


Alex highlights the downside of size difference.

David and Sarunas suggests doing a system reinstall on the SSD, leaving my 
data on the RAID. Whilst a useful upgrade, I don't really understand how that 
addresses my problem. (if I have one!)


Michael and Deloptes poo-poo that suggestion, with which I concur.

Andy suggests that I should continue using RAID, instead of reverting to a 
single disk. Thank you very much for this helpful suggestion, I considered it 
originally, but abandoned it on cost grounds. When I originally built this 
system with RAID, 10 years ago, I was very reliant on data integrity. I now 
no longer have such a pressing need, and am quite happy to rely on a single 
disk, with nightly rsync backups to an off-site server. I certainly have no 
need for the always-on data that RAID provides.


Finally, Geoff supports Andy's suggestion, and in addition, interestingly, 
mentions NVME to replace SATA SSD. Attractive, but costly.


So, you have all given me plenty to think about, for which I'm grateful.
I guess I'll go with s single SSD onto which I'll copy the data from the 
RAID.

Thank you all for helping me make up my mind.



if new drive is the same size or larger
install new drive
boot using usb drive, i use systemrescuecd
dd if=old_drive of=new_drive

Re: Advice on upgrading to SSD

[Tony van der Hoff]

On Monday 02 March 2020 06:28:58 Tony van der Hoff wrote:


Hi,
I'm currently running Buster on a 5 year old GigaByte motherboard with
a 10-year old Raid-1 array on 2 500GB disks. Although it is running
fine, I'm becoming a bit concerned about the longevity of this
storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
from Crucial.

My plan would be to install the SSD in the cage, and dd the contents
of the array onto the SSD. I would then change the BIOS to boot from
the SSD, making the RAID array redundant.

Can someone please tell me whether this plan is feasible, and what
pitfalls I might encounter?



Thanks to all who replied.

Gene didn't address my problem, but made the very useful observation 
that disks spinning 24/7 don't really die. Perhaps I shouldn't worry 
about replacing them.


Dan gave a useful step-by-step procedure for copying at file level. This 
method is slightly problematical for me as it includes "make changes to 
new /etc/fstab" and "make changes to bootloader to set new root 
filesystem", neither of which I feel feel comfortable with, due to lack 
of knowledge. I'll certainly abandon my plan to dd the whole filesystem.


Basti and Deloptes propose to add the SSD as a third drive to the RAID. 
Very attractive, but would involve complications if the SSD is not 
identical in size to the old spinners.


Andrei supports file-level copying, so I'll stick to that.

Joe supports the idea of adding the drive to the RAID, but makes no 
reference to dealing wth the size difference.


Alex highlights the downside of size difference.

David and Sarunas suggests doing a system reinstall on the SSD, leaving 
my data on the RAID. Whilst a useful upgrade, I don't really understand 
how that addresses my problem. (if I have one!)


Michael and Deloptes poo-poo that suggestion, with which I concur.

Andy suggests that I should continue using RAID, instead of reverting to 
a single disk. Thank you very much for this helpful suggestion, I 
considered it originally, but abandoned it on cost grounds. When I 
originally built this system with RAID, 10 years ago, I was very reliant 
on data integrity. I now no longer have such a pressing need, and am 
quite happy to rely on a single disk, with nightly rsync backups to an 
off-site server. I certainly have no need for the always-on data that 
RAID provides.


Finally, Geoff supports Andy's suggestion, and in addition, 
interestingly,  mentions NVME to replace SATA SSD. Attractive, but costly.


So, you have all given me plenty to think about, for which I'm grateful.
I guess I'll go with s single SSD onto which I'll copy the data from the 
RAID.

Thank you all for helping me make up my mind.

--
Tony van der Hoff| mailto:t...@vanderhoff.org
Buckinghamshire, England |

Re: Advice on upgrading to SSD

[Geoff]

Andy Smith wrote:



Just as you have two HDDs in a RAID array for redundancy, you should
not rely on a single SSD.

If you cannot afford two SSDs then I second the suggestion to
replace one of the HDDs with the SSD and set the HDD write-mostly.

Do also check the exact sector counts as it is entirely possible
that a "500GB" SSD is actually slightly smaller than your "500GB" HDDs.

Cheers,
Andy



I did exactly this, though I replaced a 1TB disk with a 2TB sata SSD to avoid 
that problem, and at some point I'll replace the remaining disk with an nvme 
drive now that I've replaced the MB, then I'll grow the filesystems to fit.

Regards,
Geoff

Re: Advice on upgrading to SSD

[Andrei POPESCU]

On Lu, 02 mar 20, 10:39:11, Dan Ritter wrote:
> 
> file-level copying is better than block-level copying when going
> between drives of different technologies.

+1 on this.

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser


signature.asc
Description: PGP signature

Re: Advice on upgrading to SSD

[Andy Smith]

Hi Tony,

On Mon, Mar 02, 2020 at 11:28:58AM +, Tony van der Hoff wrote:
> My plan would be to install the SSD in the cage, and dd the contents of the
> array onto the SSD. I would then change the BIOS to boot from the SSD,
> making the RAID array redundant.

Just as you have two HDDs in a RAID array for redundancy, you should
not rely on a single SSD.

If you cannot afford two SSDs then I second the suggestion to
replace one of the HDDs with the SSD and set the HDD write-mostly.

Do also check the exact sector counts as it is entirely possible
that a "500GB" SSD is actually slightly smaller than your "500GB" HDDs.

Cheers,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting

Re: Advice on upgrading to SSD

[deloptes]

Michael Stone wrote:

> -1, unecessary busy work

+(-1) configurations grown during the years broken - hard to restore

Re: Advice on upgrading to SSD

[Alex Mestiashvili]

On 3/2/20 6:52 PM, Joe Pfeiffer wrote:
> Tony van der Hoff  writes:
> 
>> Hi,
>> I'm currently running Buster on a 5 year old GigaByte motherboard with
>> a 10-year old Raid-1 array on 2 500GB disks. Although it is running
>> fine, I'm becoming a bit concerned about the longevity of this
>> storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
>> from Crucial.
>>
>> My plan would be to install the SSD in the cage, and dd the contents
>> of the array onto the SSD. I would then change the BIOS to boot from
>> the SSD, making the RAID array redundant.
>>
>> Can someone please tell me whether this plan is feasible, and what
>> pitfalls I might encounter?
> 
> I've found the best way to migrate to new disks, especially if I've
> already got a RAID array, is to put the new disk in, add it to the
> array, wait for the array to synchronize, remove the old disks from
> array.  Now install grub on the new disk.  If the new disk is larger
> than the old ones were, now resize the filesystem to match.
> 
> This has worked well enough for me that I've got several one-disk RAID 1
> arrays in anticipation of eventual upgrades.
> 

Good idea, but most likely the 500GB SSD is not really 500GB as most
vendors leave some space for over provisioning area.

Re: Advice on upgrading to SSD

[Michael Stone]

On Mon, Mar 02, 2020 at 02:22:45PM -0500, Sarunas Burdulis wrote:

On 3/2/20 2:19 PM, David Christensen wrote:

On 2020-03-02 03:28, Tony van der Hoff wrote:

Hi,
I'm currently running Buster on a 5 year old GigaByte motherboard with
a 10-year old Raid-1 array on 2 500GB disks. Although it is running
fine, I'm becoming a bit concerned about the longevity of this
storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
from Crucial.

My plan would be to install the SSD in the cage, and dd the contents
of the array onto the SSD. I would then change the BIOS to boot from
the SSD, making the RAID array redundant.

Can someone please tell me whether this plan is feasible, and what
pitfalls I might encounter?

Thanks, Tony


I would suggest doing a fresh install onto a single, fast SSD, keeping
the system image and your home directory small, and leaving the bulk of
your data the 500 GB mirror.  Backup, archive, and image religiously
(especially before you start on this adventure).


+1. Nothing like a fresh install.


-1, unecessary busy work

Re: Advice on upgrading to SSD

[Sarunas Burdulis]

On 3/2/20 2:19 PM, David Christensen wrote:
> On 2020-03-02 03:28, Tony van der Hoff wrote:
>> Hi,
>> I'm currently running Buster on a 5 year old GigaByte motherboard with
>> a 10-year old Raid-1 array on 2 500GB disks. Although it is running
>> fine, I'm becoming a bit concerned about the longevity of this
>> storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
>> from Crucial.
>>
>> My plan would be to install the SSD in the cage, and dd the contents
>> of the array onto the SSD. I would then change the BIOS to boot from
>> the SSD, making the RAID array redundant.
>>
>> Can someone please tell me whether this plan is feasible, and what
>> pitfalls I might encounter?
>>
>> Thanks, Tony
> 
> I would suggest doing a fresh install onto a single, fast SSD, keeping
> the system image and your home directory small, and leaving the bulk of
> your data the 500 GB mirror.  Backup, archive, and image religiously
> (especially before you start on this adventure).

+1. Nothing like a fresh install.

-- 
Sarunas Burdulis
Systems Administrator, Dartmouth Mathematics
math.dartmouth.edu/~sarunas

· https://useplaintext.email ·



signature.asc
Description: OpenPGP digital signature

Re: Advice on upgrading to SSD

[David Christensen]

On 2020-03-02 03:28, Tony van der Hoff wrote:

Hi,
I'm currently running Buster on a 5 year old GigaByte motherboard with a 
10-year old Raid-1 array on 2 500GB disks. Although it is running fine, 
I'm becoming a bit concerned about the longevity of this storage, so I'm 
planning to upgrade it to a 500GB or maybe 1TB SSD from Crucial.


My plan would be to install the SSD in the cage, and dd the contents of 
the array onto the SSD. I would then change the BIOS to boot from the 
SSD, making the RAID array redundant.


Can someone please tell me whether this plan is feasible, and what 
pitfalls I might encounter?


Thanks, Tony


I would suggest doing a fresh install onto a single, fast SSD, keeping 
the system image and your home directory small, and leaving the bulk of 
your data the 500 GB mirror.  Backup, archive, and image religiously 
(especially before you start on this adventure).



David

Re: Advice on upgrading to SSD

[deloptes]

basti wrote:

> Why you do not simply prepare the SSD (partition/align) and after that
> add the SSD as 3rd drive to the raid?
> 
> When all is synced you can kick off the HDD's or that them to "Write
> mostly".

+1

However partitioning is still required, but the best way to copy the data is
to use the built in sync mechanism of linux raid.

I am just about to replace 2x1TB WD Green against 2x1TB WD Red using this
procedure, because the Green is very slow compared to the Red. 

Age is also a point, but as mentioned by Gene when running 24/7 its almost
forever. I think those Green disks are already 10y/o

Re: Advice on upgrading to SSD

[Joe Pfeiffer]

Tony van der Hoff  writes:

> Hi,
> I'm currently running Buster on a 5 year old GigaByte motherboard with
> a 10-year old Raid-1 array on 2 500GB disks. Although it is running
> fine, I'm becoming a bit concerned about the longevity of this
> storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
> from Crucial.
>
> My plan would be to install the SSD in the cage, and dd the contents
> of the array onto the SSD. I would then change the BIOS to boot from
> the SSD, making the RAID array redundant.
>
> Can someone please tell me whether this plan is feasible, and what
> pitfalls I might encounter?

I've found the best way to migrate to new disks, especially if I've
already got a RAID array, is to put the new disk in, add it to the
array, wait for the array to synchronize, remove the old disks from
array.  Now install grub on the new disk.  If the new disk is larger
than the old ones were, now resize the filesystem to match.

This has worked well enough for me that I've got several one-disk RAID 1
arrays in anticipation of eventual upgrades.

Re: Advice on upgrading to SSD

[basti]

On 02.03.20 16:39, Dan Ritter wrote:
> Tony van der Hoff wrote: 
>> Hi,
>> I'm currently running Buster on a 5 year old GigaByte motherboard with a
>> 10-year old Raid-1 array on 2 500GB disks. Although it is running fine, I'm
>> becoming a bit concerned about the longevity of this storage, so I'm
>> planning to upgrade it to a 500GB or maybe 1TB SSD from Crucial.
>>
>> My plan would be to install the SSD in the cage, and dd the contents of the
>> array onto the SSD. I would then change the BIOS to boot from the SSD,
>> making the RAID array redundant.
>>
>> Can someone please tell me whether this plan is feasible, and what pitfalls
>> I might encounter?
>>
> 
> I would recommend:
> 
> - install SSD
> - partition SSD, aligning partitions with new block sizes
> - create filesystems in SSD partitions
> - copy over filesystems with rsync
> - make changes to new /etc/fstab
> - make changes to bootloader to set new root filesystem
> - shut down machine
> - disconnect old disks
> - start machine, make BIOS changes if necessary
> - boot
> - test for functionality
> 
> file-level copying is better than block-level copying when going
> between drives of different technologies.
> 
> -dsr-
> 

Why you do not simply prepare the SSD (partition/align) and after that
add the SSD as 3rd drive to the raid?

When all is synced you can kick off the HDD's or that them to "Write
mostly".

Re: Advice on upgrading to SSD

[Dan Ritter]

Tony van der Hoff wrote: 
> Hi,
> I'm currently running Buster on a 5 year old GigaByte motherboard with a
> 10-year old Raid-1 array on 2 500GB disks. Although it is running fine, I'm
> becoming a bit concerned about the longevity of this storage, so I'm
> planning to upgrade it to a 500GB or maybe 1TB SSD from Crucial.
> 
> My plan would be to install the SSD in the cage, and dd the contents of the
> array onto the SSD. I would then change the BIOS to boot from the SSD,
> making the RAID array redundant.
> 
> Can someone please tell me whether this plan is feasible, and what pitfalls
> I might encounter?
> 

I would recommend:

- install SSD
- partition SSD, aligning partitions with new block sizes
- create filesystems in SSD partitions
- copy over filesystems with rsync
- make changes to new /etc/fstab
- make changes to bootloader to set new root filesystem
- shut down machine
- disconnect old disks
- start machine, make BIOS changes if necessary
- boot
- test for functionality

file-level copying is better than block-level copying when going
between drives of different technologies.

-dsr-

Re: Advice on upgrading to SSD

[Gene Heskett]

On Monday 02 March 2020 06:28:58 Tony van der Hoff wrote:

> Hi,
> I'm currently running Buster on a 5 year old GigaByte motherboard with
> a 10-year old Raid-1 array on 2 500GB disks. Although it is running
> fine, I'm becoming a bit concerned about the longevity of this
> storage, so I'm planning to upgrade it to a 500GB or maybe 1TB SSD
> from Crucial.
>
> My plan would be to install the SSD in the cage, and dd the contents
> of the array onto the SSD. I would then change the BIOS to boot from
> the SSD, making the RAID array redundant.
>
> Can someone please tell me whether this plan is feasible, and what
> pitfalls I might encounter?
>
> Thanks, Tony

Its been my experience here, that drives that run 24/7, tend to run 
essentially forever. Shutting down and starting back up are the most 
dangerous times in the life of spinning rust.  I have an old spinning 
rust 1T here that has over 100k hours on it, and has not added to its 
original 25 reallocated sectors in all that time. I was going to show 
its report, but apparently it got cycled into a drawer with the last 
rebuild of this box after I had a fire at one of the usb ports and had 
to replace that 12 year old mobo.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page