Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-02-02 Thread vicrandp

Hi Robert
 Is there an image available for the debian ReadonlyRoot for Beaglebone 
Black? 

https://wiki.debian.org/ReadonlyRoot 

Thanks
Vic
 

On Friday, January 23, 2015 at 7:39:28 AM UTC-8, RobertCNelson wrote:

 On Fri, Jan 23, 2015 at 9:31 AM, toni incog toni@gmail.com 
 javascript: wrote: 
  Sure, let's say just your latest greatest debian kernel 3.14 running 
  headless doing nothing only being great? 

 Still missing the point.. 

 So in that case, you have ext4, and / mounted rw with noatime.. 

 Let's say it's idle: 

 It'll survive a few power yanks and fsck.ext4 will fix it up most of the 
 time.. 

 I'd say about a week before you'll have to reflash the eMMC to due 
 file system corruption... 


  Any experiences with broken eMMCs? How important is using a backup 
 battery 
  in a daily unclean power down situation. What's your feeling: days, 
 weeks, 
  months, years, decades? 

 Really anything  then a week of unclean power down's, start looking 
 at read only root* and other file systems. 

 * We have a beagleboard-xm running on the roof of this building as a 
 webcam taking pictures all day, it's been running (off/on) for over 3 
 years on the same microSD card running Debian Squeeze, on a read only 
 rootfs.. It's powered by a solar array, so power is variable all day 
 long.. 

 https://wiki.debian.org/ReadonlyRoot 

  
  Hmmm, I guess I've some work to do in shutting down some bbb's properly 
  otherwise I can answer my own question. Thinking aloud a battery and 
  bringing the power button external. 

 Regards, 

 -- 
 Robert Nelson 
 http://www.rcn-ee.com/ 


-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-25 Thread toni incog
Interesting, I will (also) look into ro root solution, thanks!

Can I conclude that:

1. journaling ext4 is not as robust as me thought it was only capable of 
restoring limited damage 
2. Yanking power is corrupting the fs, *not* the eMMC.
3. and therefor we can repair the fs by reflash the eMMC without any 
hardware damage done.

thx,
Michiel

On Friday, January 23, 2015 at 4:39:28 PM UTC+1, RobertCNelson wrote:

 On Fri, Jan 23, 2015 at 9:31 AM, toni incog toni@gmail.com 
 javascript: wrote: 
  Sure, let's say just your latest greatest debian kernel 3.14 running 
  headless doing nothing only being great? 

 Still missing the point.. 

 So in that case, you have ext4, and / mounted rw with noatime.. 

 Let's say it's idle: 

 It'll survive a few power yanks and fsck.ext4 will fix it up most of the 
 time.. 

 I'd say about a week before you'll have to reflash the eMMC to due 
 file system corruption... 


  Any experiences with broken eMMCs? How important is using a backup 
 battery 
  in a daily unclean power down situation. What's your feeling: days, 
 weeks, 
  months, years, decades? 

 Really anything  then a week of unclean power down's, start looking 
 at read only root* and other file systems. 

 * We have a beagleboard-xm running on the roof of this building as a 
 webcam taking pictures all day, it's been running (off/on) for over 3 
 years on the same microSD card running Debian Squeeze, on a read only 
 rootfs.. It's powered by a solar array, so power is variable all day 
 long.. 

 https://wiki.debian.org/ReadonlyRoot 

  
  Hmmm, I guess I've some work to do in shutting down some bbb's properly 
  otherwise I can answer my own question. Thinking aloud a battery and 
  bringing the power button external. 

 Regards, 

 -- 
 Robert Nelson 
 http://www.rcn-ee.com/ 


-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-25 Thread vicrandp

Robert 

On Friday, January 23, 2015 at 7:39:28 AM UTC-8, RobertCNelson wrote:

 On Fri, Jan 23, 2015 at 9:31 AM, toni incog toni@gmail.com 
 javascript: wrote: 
  Sure, let's say just your latest greatest debian kernel 3.14 running 
  headless doing nothing only being great? 

 Still missing the point.. 

 So in that case, you have ext4, and / mounted rw with noatime.. 

 Let's say it's idle: 

 It'll survive a few power yanks and fsck.ext4 will fix it up most of the 
 time.. 

 I'd say about a week before you'll have to reflash the eMMC to due 
 file system corruption... 


  Any experiences with broken eMMCs? How important is using a backup 
 battery 
  in a daily unclean power down situation. What's your feeling: days, 
 weeks, 
  months, years, decades? 

 Really anything  then a week of unclean power down's, start looking 
 at read only root* and other file systems. 

 * We have a beagleboard-xm running on the roof of this building as a 
 webcam taking pictures all day, it's been running (off/on) for over 3 
 years on the same microSD card running Debian Squeeze, on a read only 
 rootfs.. It's powered by a solar array, so power is variable all day 
 long.. 

 https://wiki.debian.org/ReadonlyRoot 


Is there such a debian image available for beaglebone black ?

  

  
  Hmmm, I guess I've some work to do in shutting down some bbb's properly 
  otherwise I can answer my own question. Thinking aloud a battery and 
  bringing the power button external. 

 Regards, 

 -- 
 Robert Nelson 
 http://www.rcn-ee.com/ 


On Friday, January 23, 2015 at 7:39:28 AM UTC-8, RobertCNelson wrote:

 On Fri, Jan 23, 2015 at 9:31 AM, toni incog toni@gmail.com 
 javascript: wrote: 
  Sure, let's say just your latest greatest debian kernel 3.14 running 
  headless doing nothing only being great? 

 Still missing the point.. 

 So in that case, you have ext4, and / mounted rw with noatime.. 

 Let's say it's idle: 

 It'll survive a few power yanks and fsck.ext4 will fix it up most of the 
 time.. 

 I'd say about a week before you'll have to reflash the eMMC to due 
 file system corruption... 


  Any experiences with broken eMMCs? How important is using a backup 
 battery 
  in a daily unclean power down situation. What's your feeling: days, 
 weeks, 
  months, years, decades? 

 Really anything  then a week of unclean power down's, start looking 
 at read only root* and other file systems. 

 * We have a beagleboard-xm running on the roof of this building as a 
 webcam taking pictures all day, it's been running (off/on) for over 3 
 years on the same microSD card running Debian Squeeze, on a read only 
 rootfs.. It's powered by a solar array, so power is variable all day 
 long.. 

 https://wiki.debian.org/ReadonlyRoot 

  
  Hmmm, I guess I've some work to do in shutting down some bbb's properly 
  otherwise I can answer my own question. Thinking aloud a battery and 
  bringing the power button external. 

 Regards, 

 -- 
 Robert Nelson 
 http://www.rcn-ee.com/ 


On Friday, January 23, 2015 at 7:39:28 AM UTC-8, RobertCNelson wrote:

 On Fri, Jan 23, 2015 at 9:31 AM, toni incog toni@gmail.com 
 javascript: wrote: 
  Sure, let's say just your latest greatest debian kernel 3.14 running 
  headless doing nothing only being great? 

 Still missing the point.. 

 So in that case, you have ext4, and / mounted rw with noatime.. 

 Let's say it's idle: 

 It'll survive a few power yanks and fsck.ext4 will fix it up most of the 
 time.. 

 I'd say about a week before you'll have to reflash the eMMC to due 
 file system corruption... 


  Any experiences with broken eMMCs? How important is using a backup 
 battery 
  in a daily unclean power down situation. What's your feeling: days, 
 weeks, 
  months, years, decades? 

 Really anything  then a week of unclean power down's, start looking 
 at read only root* and other file systems. 

 * We have a beagleboard-xm running on the roof of this building as a 
 webcam taking pictures all day, it's been running (off/on) for over 3 
 years on the same microSD card running Debian Squeeze, on a read only 
 rootfs.. It's powered by a solar array, so power is variable all day 
 long.. 

 https://wiki.debian.org/ReadonlyRoot 

  
  Hmmm, I guess I've some work to do in shutting down some bbb's properly 
  otherwise I can answer my own question. Thinking aloud a battery and 
  bringing the power button external. 

 Regards, 

 -- 
 Robert Nelson 
 http://www.rcn-ee.com/ 


-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-23 Thread toni incog
Sure, let's say just your latest greatest debian kernel 3.14 running 
headless doing nothing only being great? 

Any experiences with broken eMMCs? How important is using a backup battery 
in a daily unclean power down situation. What's your feeling: days, weeks, 
months, years, decades?

Hmmm, I guess I've some work to do in shutting down some bbb's properly 
otherwise I can answer my own question. Thinking aloud a battery and 
bringing the power button external.

thx,
Michiel

On Friday, January 23, 2015 at 4:11:20 PM UTC+1, RobertCNelson wrote:

 On Fri, Jan 23, 2015 at 9:05 AM, toni incog toni@gmail.com 
 javascript: wrote: 
  Assume you mean: ... without /un/mounting the drives under Linux. 
  
  Is there any data (or guesses) on how many times I can power down 
 without 
  unmounting before the eMMC is going to the dodo? 

 Well... This guess needs more data.. Is the partition ro? What file 
 system? etc etc... 

 So without any of the important data, I'll give you the best answer: 
 zero... Yeap zero chance of a 100% data match of data on the drive 
 right before you yank power... 

 Regards, 

 -- 
 Robert Nelson 
 http://www.rcn-ee.com/ 


-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-23 Thread toni incog
Assume you mean: ... without /un/mounting the drives under Linux.

Is there any data (or guesses) on how many times I can power down without 
unmounting before the eMMC is going to the dodo?

thx,
Michiel

On Thursday, January 22, 2015 at 8:24:44 PM UTC+1, Gerald wrote:

 Most common is powering off the board without mounting the drives under 
 Linux.

 Gerald


 On Wed, Jan 21, 2015 at 3:00 PM, bgb...@gmail.com javascript: wrote:

 Please reference paragraph 5.10 on p37 of BBB SRM.
  
 Does memory device contamination occur due to open files in the Linux OS, 
 or due to improperly sequenced power to a memory devices inadvertently 
 writing to a memory device?
  
 Thanks,
  
 Bruce

 -- 
 For more options, visit http://beagleboard.org/discuss
 --- 
 You received this message because you are subscribed to the Google Groups 
 BeagleBoard group.
 To unsubscribe from this group and stop receiving emails from it, send an 
 email to beagleboard...@googlegroups.com javascript:.
 For more options, visit https://groups.google.com/d/optout.




 -- 
 Gerald
  
 ger...@beagleboard.org javascript:
 http://beagleboard.org/
 http://circuitco.com/support/
  

-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-23 Thread Robert Nelson
On Fri, Jan 23, 2015 at 9:05 AM, toni incog toni.in...@gmail.com wrote:
 Assume you mean: ... without /un/mounting the drives under Linux.

 Is there any data (or guesses) on how many times I can power down without
 unmounting before the eMMC is going to the dodo?

Well... This guess needs more data.. Is the partition ro? What file
system? etc etc...

So without any of the important data, I'll give you the best answer:
zero... Yeap zero chance of a 100% data match of data on the drive
right before you yank power...

Regards,

-- 
Robert Nelson
http://www.rcn-ee.com/

-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-23 Thread Robert Nelson
On Fri, Jan 23, 2015 at 9:31 AM, toni incog toni.in...@gmail.com wrote:
 Sure, let's say just your latest greatest debian kernel 3.14 running
 headless doing nothing only being great?

Still missing the point..

So in that case, you have ext4, and / mounted rw with noatime..

Let's say it's idle:

It'll survive a few power yanks and fsck.ext4 will fix it up most of the time..

I'd say about a week before you'll have to reflash the eMMC to due
file system corruption...


 Any experiences with broken eMMCs? How important is using a backup battery
 in a daily unclean power down situation. What's your feeling: days, weeks,
 months, years, decades?

Really anything  then a week of unclean power down's, start looking
at read only root* and other file systems.

* We have a beagleboard-xm running on the roof of this building as a
webcam taking pictures all day, it's been running (off/on) for over 3
years on the same microSD card running Debian Squeeze, on a read only
rootfs.. It's powered by a solar array, so power is variable all day
long..

https://wiki.debian.org/ReadonlyRoot


 Hmmm, I guess I've some work to do in shutting down some bbb's properly
 otherwise I can answer my own question. Thinking aloud a battery and
 bringing the power button external.

Regards,

-- 
Robert Nelson
http://www.rcn-ee.com/

-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


[beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-22 Thread bgb123
Please reference paragraph 5.10 on p37 of BBB SRM.
 
Does memory device contamination occur due to open files in the Linux OS, 
or due to improperly sequenced power to a memory devices inadvertently 
writing to a memory device?
 
Thanks,
 
Bruce

-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [beagleboard] Contamination, SD or eMMC, Root Cause

2015-01-22 Thread Gerald Coley
Most common is powering off the board without mounting the drives under
Linux.

Gerald


On Wed, Jan 21, 2015 at 3:00 PM, bgb...@gmail.com wrote:

 Please reference paragraph 5.10 on p37 of BBB SRM.

 Does memory device contamination occur due to open files in the Linux OS,
 or due to improperly sequenced power to a memory devices inadvertently
 writing to a memory device?

 Thanks,

 Bruce

 --
 For more options, visit http://beagleboard.org/discuss
 ---
 You received this message because you are subscribed to the Google Groups
 BeagleBoard group.
 To unsubscribe from this group and stop receiving emails from it, send an
 email to beagleboard+unsubscr...@googlegroups.com.
 For more options, visit https://groups.google.com/d/optout.




-- 
Gerald

ger...@beagleboard.org
http://beagleboard.org/
http://circuitco.com/support/

-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
BeagleBoard group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.