Reply
Thread Tools
Posts: 114 | Thanked: 7 times | Joined on Jan 2006
#1
I know flash cells have only a certain lifespan to them, in terms of rewrites, and I'm hoping to have my little buggers around for awhile.

I'm sure those numbers are quite high, and difficult to reach.. perhaps for the average user. But, what about those of us that are flashing and reflashing images, installing and reinstalling and uninstalling... how quickly will be reach that threshhold? And, what happens at that point?

I'm sure it's all hypothetical, that there's no way I could ever fathom reaching that point..but.. what if? I found the HowTo on moving the filesystems over to the MMC card, and I'll be attempting that within the next few days.. easier to swap out the ever-cheaper tiny bits of plastic than replacing the entire unit...

anyone?
 
Posts: 286 | Thanked: 259 times | Joined on Jan 2006 @ Cambridge, England
#2
Hi

Depending on the flash inside, it should last 100,000 writes to a block and with wear levelling should last a while. Lots about it on the web, a couple of links here,

kernel list

Oesf forums


Though it looks like Nokia in IT2006 may have tried to free memory by no longer using a memory tmpfs fileststem for logging and temporay files and now uses flash for these files. This frees up some memory for performance, but could wear out flash quicker. /var and /tmp should be in ram, not flash. Though I'm no expert. See this here - maemo developers

Rich
 
Odin's Avatar
Posts: 207 | Thanked: 3 times | Joined on Jun 2006 @ Texas
#3
Originally Posted by spycedtx
I know flash cells have only a certain lifespan to them, in terms of rewrites, and I'm hoping to have my little buggers around for awhile.

I'm sure those numbers are quite high, and difficult to reach.. perhaps for the average user. But, what about those of us that are flashing and reflashing images, installing and reinstalling and uninstalling... how quickly will be reach that threshhold? And, what happens at that point?
I recently flashed IT2006 and it took a few minutes. I didn't log the time, but let's be conservative and say it took five minutes. Let's also say that you get (and I think this is quite reasonable) 100,000 flashes out of the memory. That comes out to around 350 days, about a year, of continuous flashing. Of course, you have to flash repeatedly every five minutes and that isn't very reasonable.

I would say that even if you use the 770 to excess (I can only use it so much, but I'm old), it's unlikely that you will wear it out before you get the needful urge for the next model. Like all dealings with memory, back up your unique data often and backup to multiple locations.
 
Posts: 286 | Thanked: 259 times | Joined on Jan 2006 @ Cambridge, England
#4
Originally Posted by richie
Though it looks like Nokia in IT2006 may have tried to free memory by no longer using a memory tmpfs fileststem for logging and temporay files and now uses flash for these files. This frees up some memory for performance, but could wear out flash quicker. /var and /tmp should be in ram, not flash. Though I'm no expert. See this here - maemo developers
Actually using a terminal osso xterm with IT2006, please can someone enter
Code:
mount [enter]
and post output here?

thanks
Rich
 
johsua's Avatar
Posts: 449 | Thanked: 18 times | Joined on Apr 2006 @ Eureka, CA
#5
/dev/mtdblock4 on /mnt/initfs type jffs2 (rw)
none on /mnt/initfs/proc type proc (rw)
none on /mnt/initfs/sys type sysfs (rw)
none on /mnt/initfs/tmp type tmpfs (rw)
/dev/mtdblock4 on / type jffs2 (rw,rpsize=1024,rpuid=0,rpuid=30000)
none on /tmp type tmpfs (rw)
/proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
usbfs on /proc/bus/usb type usbfs (rw)
none on /dev type tmpfs (rw)
devpts on /dev/pts type devpts (rw)
/dev/mmcblk0 on /media/mmc1 type vfat (utf8,uid=29999,shortname=mixed)
 
Posts: 286 | Thanked: 259 times | Joined on Jan 2006 @ Cambridge, England
#6
Originally Posted by johsua
/dev/mtdblock4 on /mnt/initfs type jffs2 (rw)
none on /mnt/initfs/proc type proc (rw)
none on /mnt/initfs/sys type sysfs (rw)
none on /mnt/initfs/tmp type tmpfs (rw)
/dev/mtdblock4 on / type jffs2 (rw,rpsize=1024,rpuid=0,rpuid=30000)
none on /tmp type tmpfs (rw)
/proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
usbfs on /proc/bus/usb type usbfs (rw)
none on /dev type tmpfs (rw)
devpts on /dev/pts type devpts (rw)
/dev/mmcblk0 on /media/mmc1 type vfat (utf8,uid=29999,shortname=mixed)
Hi Johsua

Thanks for that. Please could you also post the output of
Code:
df -h [enter]
thank you
Rich
 
Posts: 286 | Thanked: 259 times | Joined on Jan 2006 @ Cambridge, England
#7
Looks like IT2006 has /var/run wearing flash needlessly, reported as a bug

https://www.maemo.org/bugzilla/show_bug.cgi?id=598

Filesystem Size Used Available Use% Mounted on
/dev/mtdblock4 2.0M 2.0M 0 100% /mnt/initfs
none 512.0k 48.0k 464.0k 9% /mnt/initfs/tmp
/dev/mtdblock4 123.5M 82.9M 40.6M 67% /
none 512.0k 48.0k 464.0k 9% /tmp
none 1.0M 48.0k 976.0k 5% /dev
/dev/mmcblk0 979.5M 165.9M 813.6M 17% /media/mmc1
 
Posts: 191 | Thanked: 10 times | Joined on Feb 2006
#8
i tried symlinking /var/run to /tmp and had to reflash (as it wouldn't boot anymore)
 
Posts: 286 | Thanked: 259 times | Joined on Jan 2006 @ Cambridge, England
#9
Originally Posted by disq
i tried symlinking /var/run to /tmp and had to reflash (as it wouldn't boot anymore)
I wonder if moving /var/run on a running system did it, it might work if the image is setup that way and flashed or if /var/run was mounted as a tmpfs filesystem.
 
Posts: 191 | Thanked: 10 times | Joined on Feb 2006
#10
nope it kept working until i tried to reboot
i suspect that it used /var/run (hence, /tmp) before it had a chance to mount the tmpfs, so after it got mounted the old files and named pipes weren't there anymore.
 
Reply

Thread Tools

 
Forum Jump


All times are GMT. The time now is 20:03.