[PLUG] CMR/PMR hard drive manufacturers and vendors
Ben Koenig
techkoenig at gmail.com
Mon Nov 2 23:52:06 UTC 2020
On 11/2/20 11:50 AM, TomasK wrote:
> Please note that Ben said archival/backup use.
>
> There are no data persistent guarantees on SSDs if not used for longer
> than about a year. That is probably an issue if you put data on SSD for
> archival/backup - have it in deposit box, fireproof safe, off-site or
> not used then realize 2-3 years later that you would like that data
> back.
>
> Also, when SSD is dead - it is dead - unlike with hard drives where you
> could attempt to read the data off platters in a recovery drive, if the
> data is important enough data.
>
> Context is important - SSD != HDD for all use cases or $$$ per GB
Exactly. The users decision needs to be based on what kind of data they
are working with and the various requirements posed by their use case
and workflow. SOHO users don't have the resources to go through a whole
R&D project so they often look these issues as black and white. Yes/No,
Better/Worse, etc.
SSD technology is actually still pretty new and in a state of constant
change. I can't give Rich a direct answer to his question because there
isn't one. But having talked to a lot of SOHO users about storage
strategies and backup I can say that CMR/PMR drives for the SMB/SOHO
market are still the best option. In some cases an SSD is used as a
working drive but the majority of the business data resides on on small
HDD RAID arrays.
Rich, you would use a S.M.A.R.T. reporting tool just like an HDD, but
the problem you will run into is that different SSD models have
different ways of reporting wear and tear. Over the years there have
been several implementations of wear leveling and each vendor fiddles
with it on a per-model basis. Firmware versions play into this reporting
as well so there is no "Do this to predict drive failure".
Out of curiosity, what is your backup strategy? Do you have any offsite
backups and are you following the 3-2-1 rule?
-Ben
> Best,
> -T
>
> On Mon, 2020-11-02 at 09:16 -0800, Rich Shepard wrote:
>> On Mon, 2 Nov 2020, Ben Koenig wrote:
>>
>>> Most end users will experience catastrophic data loss when using
>>> SSDs in a
>>> redundant RAID system. This is because most end users just assume
>>> that SSD
>>> technology is simply better and make no attempt to monitor the
>>> health of
>>> the drive. In server environment a typical "flash array" includes
>>> software
>>> that monitors the wear to the drive, and proactive ejects the SSD
>>> before
>>> it actually fails.
>> Ben,
>>
>> For a SOHO single-user with a 60G SSD holding the OS and a 2T HDD
>> holding
>> /home, /opt, and /data (all these, and /, incrementally backed up
>> daily),
>> what tool do you recommend to monitor the SSD's health?
>>
>> Regards,
>>
>> Rich
>>
>>
>> _______________________________________________
>> PLUG: https://pdxlinux.org
>> PLUG mailing list
>> PLUG at pdxlinux.org
>> http://lists.pdxlinux.org/mailman/listinfo/plug
> _______________________________________________
> PLUG: https://pdxlinux.org
> PLUG mailing list
> PLUG at pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
More information about the PLUG
mailing list