Boston Linux & UNIX was originally founded in 1994 as part of The Boston Computer Society. We meet on the third Wednesday of each month at the Massachusetts Institute of Technology, in Building E51.

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Discuss] SSD drives vs. Mechanical drives



On Tue, May 6, 2014 at 2:47 PM, Richard Pieri <richard.pieri at gmail.com> wrote:
> Bill Bogstad wrote:
>> 1.  SSDs are constantly moving data around in order to do wear leveling.
>
> Not constantly. SSD on-board controllers automatically perform garbage
> collection when they're idle. Not when the systems are idle; when the
> controllers are idle. Or they can be sent a trim command to initiate
> garbage collection.

Fair enough.   More accurate would have been to say that it is
difficult to know when
an SSD isn't moving data around (i.e. when it is done with internal
garbage collection, etc.).

>> 2. Not all SSDs have batteries/super capacitors to finish those
>> activities if power is lost.
>
> I'm unaware of any consumer-grade SSDs with batteries or supercaps.
> That's typical of enterprise-grade devices but those tend to cost
> substantially more than consumer kit.

The original poster didn't say what kind of SSD he had.   Given that
most people around here
are talking about consumer devices, this is certainly possible.

> SSD performance is largely dependent on massive DRAM caches on the
> controllers. Seagate's super-fast Cheetah disks have 16MB DRAM caches;
> the Samsung SSD you referenced previously has a 512MB DRAM cache. That's
> where the performance comes from, and that's one place where data loss
> is going to clobber you hard if the power fails.

Making sudden power loss even more likely to screw things up.   Noting
that this can
happen with a hard drive as well if you allow write caching, but I
would think less likely
to result in catastrophic failure.   Again due to wear leveling, an
SSD could lose track
of old data which the OS may not have read or written for a long time.
 A hard drive is only
going to lose the most recently written data against which many
filesystems have been
hardened anyway.

> There are other possible problems; I leave it to the reader to ask
> Google about SSD power failures.

Power failure vs. forced power off.   Is there a difference from the
prospective of an SSD?

Moving this conversation in a slightly different direction.   Does
anybody know how to tell a generic
SSD to go into a consistent state?   If the software sends a standby
or spin down command to an
SSD will it intepret that as a command to stop garbage collection?
If not is the only solution
to wait some arbitrary amount of time and hope the SSD is finished
with its internal magic?

Bill Bogstad



BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org