Re: I've run out of entropy!

Author	All Replies
deblin Dark Side of the Moon Premium,MVM join:2001-09-01 Middletown, DE	reply to justin Re: I've run out of entropy! Hmm, seems like having alternative entropy sources would be something the kernel should include by default. Maybe it's been sufficient in the past to use hard drive access to build entropy? Or perhaps the particular kernel version is broken and not properly building entropy from hard drive access? That article specifically says "IDE timings" and mentions using hdparm, but it's unclear whether it actually means only ATA/IDE hard drives are sensed for entropy or not. If the server has SCSI, I would hope it would still be useful to build entropy. Are you planning on patching the kernel to include network interrupt support entropy? -- "Talk is cheap because the supply is greater than the demand" - Shelby Friedman
	to forum · permalink · · 2006-06-21 12:40:20 ·
justin Australian join:1999-05-28 Brooklyn, NY	the server isn't used for SSL so I'm ok with urandom. Rebuilding a kernel for this problem is something I'd rather wait for a better excuse to do.
justin Australian join:1999-05-28 Brooklyn, NY	to forum · permalink · · 2006-06-21 12:52:21 ·
sempergoofy Premium join:2001-07-06 Smyrna, GA ·AT&T Southeast	reply to deblin said by deblin : Hmm, seems like having alternative entropy sources would be something the kernel should include by default. Maybe it's been sufficient in the past to use hard drive access to build entropy? Or perhaps the particular kernel version is broken and not properly building entropy from hard drive access? That article specifically says "IDE timings" and mentions using hdparm, but it's unclear whether it actually means only ATA/IDE hard drives are sensed for entropy or not. If the server has SCSI, I would hope it would still be useful to build entropy. Are you planning on patching the kernel to include network interrupt support entropy? For long key generation on a "headless/keyboardless/mouseless" system needing entropy bits from /dev/random, I have ususally cranked up a few concurrent backgrounded dd commands from /dev/sda and other scsi devices targetted to /dev/null. I would presume (perhaps mistakenly) that doing similar with IDE drives would give the same results. dd if=/dev/sda of=/dev/null bs=1024k & dd if=/dev/sdb of=/dev/null bs=1024k & # do work needing to read from /dev/random here There needs to be multiple concurrent of these running, otherwise one could make a stronger case that the randomness was not random because the sectors were always ascending from one drive. -- nohup rm -fr /&
	to forum · permalink · · 2006-06-21 15:00:51 ·


Friday, 27-Nov 19:20:46	Terms of Use \| Privacy Policy \| Hosting by www.nac.net - DSL,Hosting & Co-lo \| feedback \| contact over 10 years online! © 1999-2009 dslreports.com.republican-creole page compression OFF