[Soekris] SMTP on Soekris?

John.Simon@sun.com John.Simon at sun.com
Wed Feb 5 05:06:15 UTC 2003


And according to this paper it will really depend on how you have your filesystem laid out, anyone with a little more experiance in BSD/Linux filesystems want to speak to this point?

http://fmg-www.cs.ucla.edu/classes/239_1.fall97/papers/flashfs.ps

>Yeah, this is the kind of stat I'd love to replace.  A sample
>of one without criteria "seeming" slow after "a while" isn't
>statistically useful, although it "does' give me a clue what
>to look for.
>
>Thanks for the info point, however.
>
>chuck
>
>Quoting John Vernier Simon (john.simon at sun.com):
>> I haven't taken a 100 CF's and written a lot of data to them. I have 
>> however left logging on a net4501 with a 32 meg firewall, ran great for 
>> about 6 months but then finally the writes got so slow it started 
>> erroring out. I still have the CF card and it works, sort of, if you 
>> don't mind the wait. It wasn't a "high end" CF, it was one of  those 
>> really cheap generic types, so I don't have any info on the newer 
>> 1,000,000+ write CF cards, but I can speak from experience, logging will 
>> eventually kill a CF card, sort of. I have also had the same experience 
>> with an older digital camera I have, my wife is an artist so she takes 
>> thousands of pictures of her artwork and the more used CF cards I have 
>> are noticeably slower then those that have seen less use (same brand, 
>> bought at the same time, some see camera duty a couple others see only 
>> PDA duty).
>> 
>> 
>> Chuck Yerkes wrote:
>> 
>> >We keep hearing numbers, but don't see evidence for *anything*.
>> >I know Sam Leffler was looking for information, but I haven't see
>> >anything.  Has anyone taken 100-200 CF disks and hammered them with
>> >data until they failed?  It's not sensible for hobbyists, but one
>_____________________________________________________________________
>Soekris Engineering, technical discussion mailing list
>[un]subscribe: http://lists.soekris.com/mailman/listinfo/soekris-tech
>



More information about the Soekris-tech mailing list