[00:42] harlowja, well, i'd replace any usage of the templates that we have in trunk to use the new format. [00:43] smoser right, i adjusted all the ones in /templates [00:43] and those templates would in some way declaratively state that they were of this new type. [00:43] yup [00:44] check out the merge smoser when u get some time [00:44] *review/merge [00:46] harlowja, thanks for all your help. [00:46] np boss [00:46] :) [00:46] like 100 distros in cloud-init now [00:46] haha [00:46] pretty interesting [00:47] horah! [00:47] def === ctracey is now known as ctracey|away [04:09] howdy? I'm stuck with vcloud director and am trying to add cloud-init to my vms because vcloud's vm customization is picky about the OS versions that it supports. [04:10] I've created a user-data.txt file and genisoimage'd it, uploaded, attached, and booted. But, /var/log/cloud-init.log doesn't show it noticing the ISO [04:11] i'm using the EPEL 6.3 version of cloudinit on centos 6.5 [04:21] any tips? It doesn't show "Alt cloud" as a datasource in the log, just nocloud, configdrive, ec2 and OVF. I guess I need to find the 7.5 rpm or build one, right? === harlowja is now known as harlowja_away [04:51] i just noticed that OVF requires userdata to be in XML. I just used the make-iso script to convert my user-data into bases64 and shove it into the XML file. I'll try it out tomorrow, since I'm unable to upload the ISO right now. === harlowja_away is now known as harlowja === harlowja is now known as harlowja_away [09:43] seanwbruno: you've had a chance to test the freebsd stuff? [09:45] smoser: anything needed from my part to have the pending merge request in cirros be fullfilled? [09:45] *accepted === oobx_ is now known as oobx [14:00] harmw, i'll try to do that today. [14:00] i'm really sorry. i want to run it in lxc once. (even though i made you boot an unholy ubuntu system to try that :) [14:31] oh it worked on my fedora system :) [14:31] and sure, np [17:15] any RH folks around? i was wondering what version of cloud-init will be in rhel7 [17:15] *and more importantly with what cloud.cfg [17:16] * harlowja might save me some time rebuilding a rhel7 rpm if i can avoid building it in the first place :-P [17:37] harlowja: does it even ship with ci? I dont see it at either ftp://ftp.redhat.com/redhat/linux/enterprise/6Server/en/os/SRPMS/ or the path for the 7beta [17:37] hmmm, was talking to the folks here, and i think they said it does, let me double check, i just got access to a rhel7beta [17:37] I could be wrong ofc [17:38] k, think u are right, doesn't seem in the rhel-beta repo, wonder where the guy saw it [17:38] perhaps epel, since it's include there [17:40] http://mirror.nl.leaseweb.net/epel/6Server/SRPMS/cloud-init-0.7.4-2.el6.src.rpm [17:40] thats for rhel6/centos6 [17:41] http://mirror.nl.leaseweb.net/epel/beta/7/SRPMS/cloud-init-0.7.2-8.el7.src.rpm [17:41] k, found it [17:41] http://dl.fedoraproject.org/pub/epel/beta/7/SRPMS/repoview/cloud-init.html [17:41] and thats 7 [17:41] ah yes harlowja [17:41] ya, looking inside it, seeing what they did :-P [17:41] hhe [17:42] ya, it seems like they just took over the fedora one :-P [17:42] distro: fedora... [17:42] in cloud.cfg [17:43] ah yes, I believe I've patched that in my private repo to read rhel or something :p [17:43] ya, some systemd changes also [17:43] are those patches fedora specific or should thy be included 'upstream'? [17:43] not 100% sure :-P [17:43] they don't seem so specific [17:44] the nodevconsole patch looks interesting [17:44] ya [17:46] hey, at least rhel7 is 'Python 2.7.5' [17:46] lol [17:46] :-/ [17:47] harlowja: I was tinkering with my desktop and may have missed help directed at me. Thanks for the cloud-init epel links for CENTOS 7. Now, I'm seeing 7.4 instead of 6.3 in http://dl.fedoraproject.org/pub/epel/6/x86_64/ [17:48] :) [17:48] there u go [17:48] smoser: got that lxc container up already? :p [17:58] harmw, so in my test here.. i see the disk output of the host. [17:58] you do? hm [17:58] proc/partitions has indeed all the hosts stuff [17:59] but lsblk didn't [18:09] i'll lok at this a bit more. [18:09] i think i might use lscpu [18:09] as it can tele me cores x sockets x threads [18:10] and thats included in cirros already? [18:10] yeah. [18:10] damn [18:10] and the lslk might differ between yours and mine. [18:10] (the lxc container might differ) [18:10] yea could be the case [18:10] but at this point whatever we see is probably not valid. [18:27] harmw, lscpu --parse [18:27] that looks like it would be nice. [18:27] but it seems to not work :) [18:27] oh. i see. [18:28] it does. but doesn't give as much info as you'd like. [18:28] damn, nifty output [18:29] though it's all zeroes whn running --parse [18:32] well, its not. [18:32] thats only for the first cpu [18:32] and you only have 1 [18:32] :) [18:32] indeed I have [18:32] i didn' realize cirros /bin/sh supported ${FOO//from/to} [18:32] that is not posix [18:34] awk '$1 == "MemTotal:" {print $2}' /proc/meminfo [18:34] you can do the divide in awk too if you wanted. and even then ditch the echo. [18:55] harmw, http://paste.ubuntu.com/7045706/ [18:55] that prints out: [18:55] arch: x86_64 [18:55] total: 16 cpu @ 1600.000 MHz [18:55] cores/sockets/threads: 4x2x2 [19:12] arch: x86_64 [19:12] total: 1 cpu @ 2009.101 MHz [19:12] cores/sockets/threads: 1x1x1 [19:12] nice [20:10] === system information === [20:10] Arch: x86_64 [20:10] CPU(s): 1 @ 2009.101 MHz [20:10] Cores/Sockets/Threads: 1/1/1 [20:10] Virt-type: AMD-V [20:10] Hypervisor: [20:10] I think I like that output a little better [20:11] (and to figure out the hypervisor I'd probably only have dmidecode as option) [20:14] harm. tha tlooks fine to me. [20:14] one cleanup to what i had, you should initialize variables to 'NA' or something. [20:17] harmw, http://paste.ubuntu.com/7046165/ [21:09] smoser: pushed [21:09] and I realy can't wait for dmidecode... :) [21:10] harmw, dmidecode is "interesting" [21:10] i recently fixed two bugs in cloud-init to *not* use dmidecode [21:10] haha [21:10] as on arm, it crashes vms [21:10] and even, reportedly, crashes hardware. [21:10] ouch [21:10] ie, you run that command and "poof" your hardware goes MIA. [21:11] well, perhaps /dev/mem is just to delicate :p [21:11] anyway, time for a new cirros release? [21:13] hm [21:13] dmesg|grep DMI: [21:13] looks like a simple way of telling which platform we're on [21:14] [ 0.000000] DMI: Red Hat Inc. OpenStack Nova, BIOS Bochs 01/01/2011 [21:14] kindof [21:14] dmesg|grep DMI: [21:14] DMI: Supermicro X7SPA-HF/X7SPA-HF, BIOS 1.0b 01/19/2010 [21:16] (or var/log/messages for that matter) [21:17] smoser: I think I like that line as part of the extended info stuff [21:17] thoughts? [21:18] cut until BIOS shows up that is === ctracey|away is now known as ctracey [21:30] $ dmesg | sed -n 's/.*[ ]\+DMI:[ ]\+//p' [21:30] LENOVO 7417CTO/7417CTO, BIOS 7UET91WW (3.21 ) 12/06/2010 [21:31] harmw, lets do something simple now. and get you some of the dat ayou want. [21:31] and then work on getting something nicer. [21:43] ok [21:43] grep DMI /var/log/messages | sed 's/.\+] DMI: $.\+$, BIOS.\+/\1/' [21:47] no need for grep. [21:47] sed is doing that above. [21:47] and in yours too. [21:48] $ sed -n 's/.*[ ]\+DMI:[ ]\+//p' /var/log/kern.log [21:48] LENOVO 7417CTO/7417CTO, BIOS 7UET91WW (3.21 ) 12/06/2010 [21:49] use '-n' [21:51] well, using grep at first seems a little faster [21:54] $ grep DMI: /var/log/messages | sed -n 's/.*[ ]\+DMI:[ ]\+$.\+$, BIOS.*/\1/p' [21:54] Red Hat Inc. OpenStack Nova [21:57] really? [21:57] are you sure? [21:59] yeah. wow. it is. [22:00] :) [22:00] I'm pushing it [22:05] http://paste.ubuntu.com/7046723/ thats how it looks like [22:08] idealy, the sudo fix for reading ssh keys shouldn't have gone into this branch [22:13] ok, and using awk to get the ram size [22:15] right. you dont need sudo there. [22:15] right? [22:15] well, not at boot time no [22:15] ah. i see. [22:15] but you do if you run the cirros-status manually [22:15] for whatever reason [22:15] right. other option is just to output only what is avilable as non-root. [22:16] and say "if you were root, you'd know more" [22:16] lets merge this, release cirros and then change that :p [22:17] :) [22:18] i have to run. [22:18] ok :)