[mythtv-users] Recommendation for SSD boot drive
Gregorio Gervasio, Jr.
gregorio.gervasio at gmail.com
Thu Jan 24 06:16:23 UTC 2013
>>>>> Raymond Wagner writes:
> On 1/21/2013 16:01, John Drescher wrote:
>> On Mon, Jan 21, 2013 at 3:31 PM, Fred Watt<fredwattmythtv at gmail.com> wrote:
>>> Anyone recommend a super fast ssd for mythfrontend/slave boot? Currently use
>>> a 13Gb partition, but I would welcome fast start up times.
>>
>> I recommend for good wear leveling to get a 64GB or 128GB.
> That seems a bit excessive. A MythTV install is only going to take a
> few GB, even on modern bloated distros. The only writing it's going to
> be doing is log data and the theme cache, and the occasional system
> update. I doubt you would see 128GB of writes over the life of the
> system, certainly not the thousand times that to necessitate a drive
> that large for the purposes of wear leveling.
I agree. For reference, my *master* backend has had an 80GB
Intel 320 SSD for almost exactly a year. It's partitioned like this:
Filesystem 1MB-blocks Used Available Use% Mounted on
/dev/sda3 21138MB 12744MB 7322MB 64% /
/dev/sda4 52874MB 22666MB 30208MB 43% /home
This is a Fedora installation which I haven't really bothered to trim
down. I do a system update every 1-2 weeks and the drive has gone
through a full upgrade to F17, then F18.
The mythconverg DB is using just over 1GB on /. /home is mostly
DB backups, daily remote backups, and a mythtv git master checkout
(ie. stuff you wouldn't normally put on a master backend boot drive).
Every 2-3 weeks, I build mythtv on another (faster) machine, then rsync
to the local drive, and install to /usr/local. Overall, I'd say the
drive sees quite a bit more writes than a normal MythTV backend
boot/root drive, let alone one on a frontend/slave.
Anyway, after a year of use, SMART reports Media_Wearout_Indicator
at 98%, with Host_Writes_32MiB at 295279 (just over 9TiB of writes).
--
Gregorio Gervasio, Jr.
gregorio.gervasio at gmail.com
More information about the mythtv-users
mailing list