[00:01] mercutio: bhyve can do pci-passthru if you have VT-d [00:01] daca: cool [00:01] most systems have vt-d these days [00:01] well systems you may want to virtualise on. [00:02] vt-d is acutally more secure if not doing passthrough too [00:02] how do you know so much about this? [00:02] consultant? [00:22] mercutio: Parallels has a GUI option for selecting isolation, however I haven't found any technical description of what that means [00:22] For future reference, the UEFI FBSD builds were not recognized [04:04] i didn't realised freebsd had uefi support [04:04] is that new in 10? [06:56] *** dj_goku has quit IRC (Ping timeout: 240 seconds) [07:35] mnathani_: What's I've done is set up VMs in the VirtualBox GUI (sometimes fiddling with advanced settings via VBoxManage), and then start them with with the VBoxHeadless command [07:37] I've also used VirtualBox's remote display server + xfreerdp when I need to peek at the console of a headless VM [15:53] I would appreciate suggestions for high-performance HTTP reverse proxies. I'm considering eliminating an entire web server in a stack that only uses one for HTTP reverse proxying to Node.js instances [16:46] trafficserver [16:47] I'd be looking for something without the caching feature [16:47] nginx [18:27] haproxy [18:30] grody: I looked at that. Have you run it in production? [18:31] production.. no, not i [18:32] relayd? httpd? :P (I can't speak from experience, but know they can both do that sort of frontend/load balancer activity) [18:34] I hadn't found relayd (http://relayd.org/about.html) thanks brycec [18:35] (Count on the OpenBSD nut to know about OpenBSD things) [18:35] relayd is a bit finicky under freebsd [18:35] it's stable under openbsd though [18:36] (and nobody is really surprised about that ^ :P) [18:36] the finicky part or stable under openbsd? [18:36] I'm surprised it runs on non-OpenBSD at all. [18:36] mercutio: stable [18:36] yeah [18:37] brycec: it's just libevent [18:37] and pf [18:37] which freebsd has too [19:45] kellytk: http://www.apsis.ch/pound/ [19:51] pound is perl and so low performance? [19:51] he said high performance rather than flexible. [19:52] maybe performance isn't too bad [19:53] it isn't perl, it just uses pcre [19:53] daca, thanks, pound is on my list [19:53] oh it's python/c? [19:53] it is c [19:53] ahh the tarball had a .py file in it [19:53] but it may be just part of the compile process. [19:54] I quote: If the PCRE, tcmalloc (from the Google perftools package) and/or Hoard are available Pound will link against them. This will provide a significant performance boost and is highly recommended. [19:54] cool [19:54] varnish seems popular [19:54] but never gave me a good impression [19:54] well he asked no caching :) [19:55] yeh i dunno what he has against caching [19:55] As I already have nginx running well, I've decided to remain with it until I'm ready to run benchmarks, which is several months out [19:55] I don't have anything against caching, I simply don't have need for it at present [20:17] *** dj_goku has joined #arpnetworks [20:29] *** NiTe has quit IRC (Ping timeout: 246 seconds) [20:30] *** toddf has quit IRC (Ping timeout: 246 seconds) [20:30] *** NiTe has joined #arpnetworks [20:31] *** rendrag has quit IRC (Ping timeout: 246 seconds) [20:31] *** gizmoguy has quit IRC (Ping timeout: 246 seconds) [20:31] *** up_the_irons has quit IRC (Ping timeout: 246 seconds) [20:31] *** gizmoguy has joined #arpnetworks [20:31] *** rendrag has joined #arpnetworks [20:31] *** neish_ has quit IRC (Ping timeout: 246 seconds) [20:32] *** toddf has joined #arpnetworks [20:32] *** ChanServ sets mode: +o toddf [20:35] *** up_the_irons has joined #arpnetworks [20:35] *** ChanServ sets mode: +o up_the_irons [20:37] *** neish has joined #arpnetworks