[mythtv-users] Mirroring root partition on SSD to a hard drive -- bad idea?
David Madsen
david.madsen at gmail.com
Fri Jan 15 09:45:49 UTC 2016
I've been running my MythTV system on an Intel X25-M SSD for several years
with no issues. I have a rather large database and the SSD has sped up the
operation of the system tremendously. I'll never go back. This drive was
in use in a separate machine for a while before running in the MythTV
system, but here's some statistics for the curious.
9 Power_On_Hours 0x0002 100 100 000 Old_age
Always - 46119
12 Power_Cycle_Count 0x0002 100 100 000 Old_age
Always - 1165
233 Media_Wearout_Indicator 0x0002 086 086 000 Old_age
Always - 0
225 Host_Writes_32MiB 0x0000 199 199 000 Old_age
Offline - 196422
That comes out to ~6 TB of writes over the lifetime of the drive. This is
a snapshot of the same statistics from 2011 (about 5 years of use all in
the MythTV system)
9 Power_On_Hours 0x0002 100 100 000 Old_age
Always - 11860
12 Power_Cycle_Count 0x0002 100 100 000 Old_age
Always - 1106
233 Media_Wearout_Indicator 0x0002 096 096 000 Old_age
Always - 0
225 Host_Writes_32MiB 0x0000 200 200 000 Old_age
Offline - 73719
That comes out to about 4 TB of writes in ~5 years of use. If you look at
the Media_Wearout_Indicator, that starts at 100, and drops as the drive is
written to. According to Intel's estimates, over the 5 years of use I've
only used 10% of the drives write endurance (and those estimates are
usually hugely conservative). By the time this system wears out the drive,
it'll be long obsolete. As others have said, I wouldn't worry too much
about write endurance of the SSD. The internal wear leveling will take
care of most of the really problematic stuff. SSDs are much more resilient
than running on an SD card or something like that.
For those interested, here is a great experiement showing the true
endurance of a sampling of several SSDs:
http://techreport.com/review/27909/the-ssd-endurance-experiment-theyre-all-dead
While the drives varied in endurance based on flash technologies (MLC vs
TLC) and whether the controller implemented compression, they all easily
passed 100TB worth of write without issue. The stalwarts actually made it
to over 2,000TB.
--Dave
On Fri, Jan 15, 2016 at 7:18 AM, <mythtv at phipps-hutton.freeserve.co.uk>
wrote:
>
> Quoting Calvin Dodge <caldodge at gmail.com>:
>
> For backups I do something like this, because it lets me keep multiple
>> old version of files.
>>
>> 1) designate an archive location
>> 2) Every day, make a copy of files in that location with "cp -al" (the
>> -l means "make hard links of files, rather than copying them) to a new
>> directory (my script creates on named YYYYMMDD). For me the location
>> is /backups, with the rsync directory (see step 3) being
>> /backups/latest, with daily directories like /backups/20160101,
>> /backups/20160102, etc.
>> 3) rsync -a --delete source to destination. Since any files will be
>> deleted and replaced, rather than changed, the old versions of such
>> files will be safe in those daily directories.
>>
>> Calvin
>>
>
> I think you just reinvented rsnaphot
>
>
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://lists.mythtv.org/mailman/listinfo/mythtv-users
> http://wiki.mythtv.org/Mailing_List_etiquette
> MythTV Forums: https://forum.mythtv.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mythtv.org/pipermail/mythtv-users/attachments/20160115/ba6b7a7c/attachment.html>
More information about the mythtv-users
mailing list