ts-7000
[Top] [All Lists]

Re: [ts-7000] TS-7553 MicroSD longevity

To:
Subject: Re: [ts-7000] TS-7553 MicroSD longevity
From: Jim Jackson <>
Date: Tue, 17 May 2011 15:44:58 +0100 (BST)


On Mon, 16 May 2011, walter marvin wrote:

> Some devices may have advanced wear levelling techniques. But OS' tend to
> write meta data in the same place, so that's what wears out first.

No. Because wear levelling is basically a technique that maps the block the
OS _thinks_ it is writing to (the logical block number) to a real physical
flash block. Thus allowing the physical block used to be moved around the
device to level the write wear.

> Solution: Journaling.

This is a non-sequitor.


Can I suggest you read up on what wear levelling actually means?

Try the Wikipedia article at

  http://en.wikipedia.org/wiki/Wear_leveling

Then try this from an actual manufacturer...

  http://download.micron.com/pdf/.../nand/tn2942_nand_wear_leveling.pdf

I'm sure google will throw loads of extra articles/papers.

> On Sun, 15 May 2011, walter marvin wrote:
>
>
>
> > It the writes per cell that is the limit not total writes journaling will
>
> > repair a cell that has had too many writes
>
>
>
> Yes but misleading. All flash based devices like USB sticks and SD cards,
>
> CFdisks etc contain wear levelling. Where the journal is stored will move
>
> about the flash blocks to spread the wear out.
>
>
>
> I repeat, increasing the total number of writes to the flash devices
>
> increases the wear on the flash memory. Whether this is important in one's
>
> application is upto the engineer to determine. Just ignoring it with bland
>
> platitudes is not sensible engineering.
>
>
>
> >
>
> > --- On Sun, 5/15/11, Jim Jackson <> wrote:
>
> >
>
> > From: Jim Jackson <>
>
> > Subject: Re: [ts-7000] TS-7553 MicroSD longevity
>
> > To: 
>
> > Date: Sunday, May 15, 2011, 5:27 PM
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >  
>
> >
>
> >
>
> >
>
> >  
>
> >
>
> >
>
> >    
>
> >      
>
> >      
>
> >      
>
> >
>
> >
>
> >
>
> > On Sun, 15 May 2011, walter marvin wrote:
>
> >
>
> >
>
> >
>
> > > My info is infinite reads but limited writes this means that you must use
>
> >
>
> > > a journaling file system or build in track remapping
>
> >
>
> >
>
> >
>
> > Not sure that makes sense. In fact a journalling file system INCREASES
>
> >
>
> > writes - first to write the update details to the journal, then to
>
> >
>
> > do the updates, then the make the update done in the journal.
>
> >
>
> >
>
> >
>
> > The use of a journalling file system helps if the system is likely
>
> >
>
> > to suffer unexpected power loss etc. When it comes back up, any uncompleted
>
> >
>
> > transactions can be verified and completed, and there is usually no need
>
> >
>
> > for a very lengthy file system integrety check.
>
> >
>
> >
>
> >
>
> > As an engineer it's up to you decide the trade offs and decide which is
>
> >
>
> > more important. My own systems have battery backup, allowing graceful
>
> >
>
> > shutdown, so I use ext2 (mounted with "noatime") to reduce writes. If I
>
> >
>
> > didn't use battery backup, I'd use ext3, and stand the write hit to give me
>
> >
>
> > better file system integrety.
>
> >
>
> >
>
> >
>
> > >
>
> >
>
> > > --- On Sun, 5/15/11, parkranger_dan <> wrote:
>
> >
>
> > >
>
> >
>
> > > From: parkranger_dan <>
>
> >
>
> > > Subject: [ts-7000] TS-7553 MicroSD longevity
>
> >
>
> > > To: 
>
> >
>
> > > Date: Sunday, May 15, 2011, 1:37 PM
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >  
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >  
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >    
>
> >
>
> > >      
>
> >
>
> > >      
>
> >
>
> > >       Hey guys.  I posted awhile back when having some functionality 
> > > issues with a TS-7200 and realized the 7553 was really the platform I 
> > > should have been working with from the get-go. 
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > Sweet little box this 7553.  I've gotten all my software ported over and 
> > > it's working perfect.  After only a couple weeks with it, I think I'm 
> > > ready for deployment. 
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > We'll be preparing quite a few of these little guys.  Having the ability 
> > > to insert a pre-imaged MicroSD, set one jumper, and run one command (rm 
> > > linuxrc; ln -sf /linuxrc-sdroot /linuxrc; save) to change the boot is 
> > > very attractive in terms of quick deployment, and easy field upgrade 
> > > (send customer a new MicroSD card, done!).
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > One question I had was in regard to the longevity of these little MicroSD 
> > > cards, and their resilience to repeated power loss.
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > Customers will not have the ability to shutdown nicely, power will always 
> > > be removed to turn off.  I based my image from the latest.dd image 
> > > available from the Technologic website, which if I remember correctly is 
> > > formatted ext3. 
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > Has anyone had any experience with a similar setup?  How are the boxes 
> > > holding up, and have the MicroSD cards been lasting? 
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > I've been also pondering making use of the xnand drive with a custom 
> > > busybox that includes the compiled apps i need and just script the 
> > > flashing process to ease prep/deployment.  I know bootup time would be 
> > > significantly better than my current 1 minute timeframe, and resilience 
> > > would be better.  Downside would be that I lose the ability to do remote 
> > > software updates.  Anyway, I have yet to break ground on that idea, or 
> > > even wrap my brain around how that's done.
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > Thanks!
>
> >
>
> > >
>
> >
>
> > > Dan
>
> >
>
> > >
>
> >
>
> > > 
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >    
>
> >
>
> > >     
>
> >
>
> > >
>
> >
>
> > >    
>
> >
>
> > >    
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > > 
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >  
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> >    
>
> >     
>
> >
>
> >    
>
> >    
>
> >
>
> >
>
> > 
>
> >
>
> >
>
> >
>
> >  
>
> >
>
> >
>
> >
>
> >
>
>    
>     
>
>    
>    
>
>
> 
>
>
>
>  
>
>
>
>

------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/ts-7000/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/ts-7000/join
    (Yahoo! ID required)

<*> To change settings via email:
    
    

<*> To unsubscribe from this group, send an email to:
    

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/

<Prev in Thread] Current Thread [Next in Thread>
Admin

Disclaimer: Neither Andrew Taylor nor the University of NSW School of Computer and Engineering take any responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the birding-aus mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU