G: looks like you're using irssi and you might find http://todd.fries.net/pub/lose.pl (adjust regex to your common typos) as a script plugin ;-) toddf: interesting :) toddf: my problem is more that my fingers are one step ahead of my brain :) well, I hang in channels where people get annoyed at such typos. I don't, but .. it tries to do the right thing even if the brain to keyboard interface is missing some data. up_the_irons, so I have 4 buildworlds to complete.. one is running now.. would it be safe (you think) to run like 2 at a time? pjs: freebsd-update! hah you know in my 10 years as a fbsd user, I've never used freebsd-update. Always done it manually jpalmer, you recommend it? pjs: unless you are deviating from a standard binary in some way (custom patches, etc) then it's a heck of a lot more efficient. Obviously you'll want to do it in a non-production environment and get familiar witht he tool first. Yea pjs: the other option: build the world on the first box, then export an NFS mount to the others, for 'installworld' build once, install multiple. (assuming they are the same architecture, of course) jpalmer, ahh, thats not a bad idea actually jpalmer, I assume so.. they're all arp vps's :) just had 4 new ones setup last night Yea.. I think I'll do the NFS idea it's not a bad move. one of them will need a different kernel config (for Postgres Tuning) but I can just rebuild that kernel after the fact will freebsd-update go from like 8.1 to 8.2? yes and you can tell freebsd-update to ignore kernel updates if you custom compile a kernel. hrmm, seems like it'd be a good one to try maybe I do freebsd-update and just rebuild the kernel.. then again, since these are VPS's maybe it's smarter to squeeze out every ounce of performance.. man I've been in dev to long.. havnen't had to do any real sys admin'ing in a while pjs: i put your new VMs on different machines (turns out I had enough slots across the board), so you can run them all at the same time, no problem. I tend to distribute VMs of the same account on different machines so all your eggs aren't in the same basket :) mhoran: nice. but you can't say something like that without pasting a traceroute!! ;) G: good catch, it *is* an O, because someone else in the peeringdb has the same IP (so an old record) and it didn't let me put in the same IP jpalmer: yup, that week sure paid off, it's amazing. and yeah, support requests went down 25% and notice how I've not received one, not *one*, delegation request since then ;) so i guess this proves people don't give a shit about delegation if they can just update their PTRs in the Portal toddf: happy that you are impressed with the list. _insane_ ? really? why, peering is not that hard ;) toddf: and those are only the peers that actually have a BGP session on my route. there are many more peers on the Any2Easy IX and PE eXpress route servers that I peer with, but I didn't have an automated way of generating a list of those up_the_irons, is there a specific vlan for my VM's? I can't seem to talk to each other via vlan interfaces but I just gave them all a vlan of 1 (no doubt that's wrong lol) toddf: s/route/router/ pjs: yes, ARP puts each client on their own vlan pjs: there is a specific VLAN, yes, but that shouldn't matter on your side pjs: all your VMs should be able to talk directly pjs: you might be interested in knowing.. I just found a puppet module for bacula ;) it sets up the director service, the client service, installs all relevant software, and generates the SSL certificates. (ie, everything you'd need to do without puppet) up_the_irons, I did 'ifconfig vlan1 create && ifconfig vlan1 inet 192.168.1.1 netmask 255.255.255.0 vlan 1 vlandev em0" across all of them (changing internal IP's).. they can't ping each other pjs: don't use vlan interfaces, since your interfaces are not "trunk" ports (in cisco speak); pretty sure they have an MTU of 1500 so the VLAN tag won't fit up_the_irons, ahh ok.. so how do I give them internal IP's for use? I'd like to use a separate interface for this to make it easy on my pf rules :) pjs: to be honest, i've never tried that; i don't know if the VLAN tag survives pjs: i can give your VMs a second interface just pop in a support ticket with all the VM UUIDs up_the_irons, ahh ok.. thanks and i'll add a 2nd interface to them all up_the_irons: managed switches will do wierd things with vlan tags on untagged interfaces. they probably eat them. worst case they'll go to the vlan if it is accessable by the switch, but obviously not be able to be passed back to the system that sent out the packet. vlans inside vlans .. heh. up_the_irons: I'm not a bgp guru, just have serviced some pretty low level isp's and them having two peers (uplinks) was pulling teeth and spending a fortune. I guess I am not acclimated to the 'hosting systems near/in an IX location'. do you have a physcal ethernet cable for each or is it some network everybody writes in blood they will not spam and you hook up a 10G feed to it and talk to whoever is on the subnet there? up_the_irons, request submitted.. no rush :) when I was looking at nwax, it looked like it could be setup so everyone shared a vlan. You'd only be able to peer with groups you've agreed with and setup a route to+from, and for most of the ones in the list I assume they are only local traffic, NOT transit to anywhere on the other side of their network (which would be spendy) dferris: i noted the same toddf: there is a dedicated GigE for the "peering network", and yeah, everyone kinda promises to behave. But nevertheless, I have ACLs on the peering network to mitigate most problems anyway. Like, you're only supposed to accept packets destined for your own ASN, or else you could be providing free transit. So i have an ACL for that, etc... peering network = *each* peering network so 1 gigE for Any2 IX and 1 gigE for PE eXpress if I could get a cheap vlan trunked circuit to Equinix in LA, I could also join the Equinix peering exchange, which has more peers, but also has Equinix's snotty price tag (like $250 / mo or something) up_the_irons: so you're directly reachable from a bunch of networks, and so by peering with them you are reachable in general but you don't pass traffic between them? toddf: exactly. and you *could* pass traffic between them if you wanted, but why do it unless you get paid? ;) ic hmm. for some reason, my zxfer disk blew out all of the backup, and now it's transferring another 30GB putting me over my limit at the other ISP. :( so zxfer is *almost* usable. :) up_the_irons: Of course! http://pastie.org/private/vk0mf4emunf6xrfljuihww Limelight has pretty impressive peering, I'd say the majority of our US traffic goes out through them. We've got Tinet and NTT as well. Turns out transit is ridiculously cheap at 10Gbps.