http://blog.cryptographyengineering.com/2012/10/attack-of-week-cross-vm-timing-attacks.html
yay, packet loss
mtr
up_the_irons, jdoe: fwiw, on smartos you can quota cpu/io
(and growing a disk attached to a vm is as easy as vmadm update cebbbd8e-b747-4da5-b0bc-ba7a113f73aa quota=500)
jbergstroem: but does the filesystem on the guest grow?  I would think not, or at least not on all guests (FreeBSD filesystems don't always growfs(8) correctly)
it's also illumos-based, which may/may not be here in N years.
up_the_irons: it does grow, but isn't allocated until use
jdoe: good to know
their dev community is (I think) still pretty small.
jbergstroem: i'd have to see that for myself, b/c i've never seen a system that will grow both Linux (ext3/4) and FreeBSD / OpenBSD (UFS) filesystems on-the-fly.  Linux, yeah, but not the UFS stuff.
up_the_irons: not kvm stuff, just the smartos vm's
up_the_irons: for that you need gparted and such'n'such. you still have quota options though.
also: http://blog.smartcore.net.au/smartos-command-line-fu/
jbergstroem: ok, yeah, if it's a native smartos vm, i believe it
up_the_irons: basically similar to fbsd + zfs + jail I would presume
jbergstroem: yeah, that is the impression i get as well
jlgaddis: how is the packet loss going?
sorry, got distracted.  gone now.
well, s/gone/normal/
:)
up_the_irons: are vm's supposed to run ntpd/the lkes? i thought i got clocked passed on from kvm
jbergstroem: no magic passes clock from the kvm host. I tried to talk to the qemu devs about supporting the wmware timedelta via vmt(4) on OpenBSD but they basically shut me down.  If you want this to happen, best develop some virtio clock info to pass to the guest os, and implement a timedelta sensor ontop of the virtio(4) framework already in e.g. openbsd. no it does not exist yet. yes it works well for vmt(4) on real vmware systems.
toddf: thanks for the info. went openntpd.
actually drifted 8 minutes in ~2 months.
O_o