Hi Gerald,

you are warning about SSD, with many technical arguments and questions, 
down to the cell level (one cell is not one bit). That's great.

Do you have any practical experience with SSD technology you would like 
to share with others, to argue all the facts your are highlighting ?

Regards
Karim

gerald.eggenberger at sunwave.ch wrote:
>> - Lifetime: Actual technology provide an MTBF of at least 1M Hours.
>> Laptop HD have MTBF of around 2M Hours. You'll replace your laptop
>> before replacing your SSD, and you have more chance(risk) to have a
>> failing HD than a failing SSD :)
>
> Sorry Karim, but the specifications are 10'000 Write-Cycles per Cell 
> (MLC SSD). If you have a log file for example (just to mention one) 
> which writes all the time in the same few Cells, these cells reach the 
> 10'000 writes in a moderate time and perhaps before you're System runs 
> out of Life-Cycle. And yes i know there are the possibility of 
> log-rotation, but this also depends on proper 
> system-design/configuration.
>
>> - Hotspot: the selfhealing feature of ZFS (checksum) is perfect to
>> counter "badblocks" for dying cells.
> You're right, but ZFS also multiplexes the Content of each metafile 
> and therefore increase the amount of hot-spots. What if all these 
> multiplexed metafiles on the same disk becomes unaccessable at the 
> same time?
> This Hot-Spots covers one or more Cells. Multiplexing also increase 
> the speed in 'killing' Cells.
> I don't know what exactly will happen when this 10'000 write-operation 
> per Cell are reached. When only some Blocks become unusable you're 
> perhaps happy when you've used a filesystem which can handle this. But 
> perhaps it is also could your dead. What if your system heals itself 
> and move content of dead cells to cells still alive? Once the amount 
> of healty Cells are equal to Zero! Busted, if you don't realise it 
> early enough.
>
> Have you also mentioned what will happen to your aged hfs Plus 
> Filesystem or NTFS (Windows XP, VISTA) when the Content in defect 
> Cells are unaccessable?
>
>> My advice: give a try and then consider the "risks".
> What kind of Sysadmin are you with such an attitude?!
> Well, just jump out of the Window - the feeling will be great?!
>
> Perhaps you won't be the first with the experience of a touch down... 
> but im sure there will be some whiny gadget-Freaks with no 
> technological Background saying:
>
> No one warned me. No one said to me, that the lifetime depends on 
> System-Usage/System-Configuration. Oh! There were different SSD's with 
> different Firmware/Controller? My cheaper MultiLevelCell SSD has a 10x 
> shorter Lifetime than the SingleLevelCell SSD? I don't want to get the 
> answers on this question when i need them.
>
> So please Karim don't distribute a Headless Mindset on this Mailinglist.
> I may be wrong in some points, then help me filling the gaps. But 
> deliver Facts! And not the kind of high gloss paper...
>
> Thanks for understanding.
>
> regards
> G?rald
>
> _______________________________________________
> ug-chosug mailing list
> ug-chosug at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/ug-chosug

Reply via email to