On Sun, Oct 28, 2018 at 11:41:14AM +0100, Bernhard Übelacker wrote: > Dear Maintainer, > just some more information, because I think I see this > difference in my qemu buster amd64 VM too. > > Before I could ssh into that machine just after some seconds. > > Now it takes some time until that line "random: crng init done" > appears in dmesg. > With logging in in the qemu window this line appears just after a > few seconds, when just trying via ssh it takes much longer.
I have a NAS running Debian[1] and there I only get the "crng init done"
message after more than 15 minutes (even though there is network activity and I
debugged the failure via rs232):
root@machine:~# dmesg | tail
[ 22.579884] audit: type=1400 audit(1529665920.708:3):
apparmor="STATUS" operation="profile_load" profile="unconfined"
name="man_filter" pid=299 comm="apparmor_parser"
[ 22.627877] audit: type=1400 audit(1529665920.708:4):
apparmor="STATUS" operation="profile_load" profile="unconfined"
name="man_groff" pid=299 comm="apparmor_parser"
[ 22.711102] 8021q: 802.1Q VLAN Support v1.8
[ 22.840150] mvneta d0074000.ethernet eth1: PHY
[d0072004.mdio-mii:01] driver [Marvell 88E1318S]
[ 22.856020] mvneta d0074000.ethernet eth1: configuring for
phy/rgmii-id link mode
[ 22.872501] IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready
[ 26.020202] mvneta d0074000.ethernet eth1: Link is Up - 1Gbps/Full -
flow control rx/tx
[ 26.027639] IPv6: ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
[ 1057.771583] random: crng init done
[ 1057.774739] random: 7 urandom warning(s) missed due to ratelimiting
I'm not aware the machine has a random number generator, so the solutions
presented here up to now don't help me.
I think the idea to mention this in the release notes for buster is a
good one (unless a solution is found until then of course).
Best regards
Uwe
[1] this i an armv7/mvebu machine with a Armada 370 CPU
signature.asc
Description: PGP signature
