Maybe logs could be written to a USB memory stick? regards, Philip
Poul-Henning Kamp wrote: > In message <[email protected]>, Joel Jaeggli writes: > > The answer is "it depends on so many factors that you do not and > often cannot know, that you might as well give up predicting it at all." > > Unless you can get statistics out of the flash-adaptation layer in the > card, there is no way to project what kind of state the flash array > truely it is in. > > By the time you get the first bit of relevant information, in the > shape of a read or write error, it is already time to throw the > card out. > > If you use the card like it was intended, few writes of big files > (photos in a camera) you will generally be fine. If you do a lot > of small writes you will generally not be. > > For 'big', 'small', 'few' and 'a lot', insert any random numbers > you care to dream up, nobody knows the real numbers. (A few of > the high end manufacturers have data that is indicative, but they > will only share them under NDA.) > > NanoBSD and Fifolog (both in FreeBSD) were both written to specifically > avoid and when unavoidable, to optimize the write pattern for CF card > longevity. > > Poul-Henning > > _______________________________________________ Soekris-tech mailing list [email protected] http://lists.soekris.com/mailman/listinfo/soekris-tech
