fatal: Not enough entropy in RNG

Ben Lindstrom mouring at pconline.com
Fri Jul 21 04:16:35 EST 2000

On Thu, 20 Jul 2000, Lutz Jaenicke wrote:

> On Thu, Jul 20, 2000 at 02:00:23PM -0400, Matthew C. Weigel wrote:
> > On Thu, 20 Jul 2000, Lutz Jaenicke wrote:
> > 
> > > The integrated entropy collector should better not fail...
> > 
> > But it can, at least on boring machines.  My HP running NeXTSTEP can
> > *occasionally* get enough activity to let sshd start up, but tweaking
> > /etc/ssh_prng_cmds to very high values still doesn't always work.
> That's why I wrote PRNGD :-) You can get it from my site in the Postfix/TLS
> path. Since I announced it on openssl-users, I have seen several downloads
> in my ftp-log, but did not receive any comment about it.
> If you port it to NeXTSTEP, your input would be welcome :-)
Ermm.. May be something to look at in the future.  On my NeXT box I just
use egd.pl.  It works for the little traffic that occurs on that box at
this moment.  I would perfer the internal ssh prng generator to work on my
machine or a kernel based prng.. I'm not a fan of attempting to ensure
that some external prng software is running.

OH.BTW..I've not tried it.. Anyone running the NeXT port that has a system
handy (mine is not.. It's been brought down again for a while due to space
restrictions on my subnet) try doing "UseLogin YES" in the sshd_config ..
See if that temporary solves the utmp issue.  At least as a work around
until we figure out what the trouble is.

Ben Lindstrom

More information about the openssh-unix-dev mailing list