This one time, at band camp, martin f krafft said: > also sprach Alexander Reelsen <alr@emplify.de> [2006.07.04.1247 +0200]: > > > Is there anything else you'd say I should check? > > You can check the available entropy via sysctl > > cat /proc/sys/kernel/random/entropy_avail > 1 > > Mh. Looks like Stephen was right after all. > > So what now? And why does this even concern me, as /dev/random is > never used...? If /dev/random was used, likely the TLS transaction would take place tomorrow, instead of 15 seconds from now. Remember that urandom uses real entropy preferentially, and only makes it's own when it would otherwise have to block. Having low entropy situations will affect both random and urandom reads, but urandom will continue, if slowly. Good luck. This is a recurring problem for me and for many others, judging by traffic on various lists (seanius is fighting with this at the moment as well). Try different kernel versions, see if you have forgotten to load the hw_random module (if Intel chipsets) etc. Generate lots of disk IO to help you limp along. I suppose you could take a look at all the modules for the machines, and see if they have been ported to the new entropy gathering routines, and port them if not. Take care, -- ----------------------------------------------------------------- | ,''`. Stephen Gran | | : :' : sgran@debian.org | | `. `' Debian user, admin, and developer | | `- http://www.debian.org | -----------------------------------------------------------------
Attachment:
signature.asc
Description: Digital signature