=== harlowja is now known as harlowja_away === harlowja_away is now known as harlowja === harlowja is now known as harlowja_away [10:23] smoser: don't forget your promise :P [13:06] harmw, i'll look at that right now. [13:06] are you going to paris ods ? [13:27] no, afraid not [13:29] booo [13:43] :) [14:07] smoser: let me know if you need feedback on reviewing that merger, I haven't got much time though [14:12] k [14:59] 30minutes left :P [17:50] smoser: I think LP is broken [17:50] it still didn't send me any notification about an accepted merge request :> [17:50] :) [17:51] it should have sent you some info though [17:51] i've been going through bugs in cloud-init all day. [17:51] i do sitll paln to get on that mp thoug h today. [17:51] haha [17:51] well, I don't think I've ever submitted bugs to c-i [17:51] hence I didn't get any mail today :p [17:52] but nice, fixing bugs is also needed :) [17:53] https://code.launchpad.net/~cloud-init-dev/cloud-init/trunk <- th [17:53] this the main source, right? [17:53] yep [17:54] I might try to submit my problem with not being able to use NoCloud and a seed as a bug [17:54] bring some closure to the issue, well at least for me :) [17:54] go ahead :) === zz_gondoi is now known as gondoi [18:15] harmw, you still there? [18:49] smoser harmw if u guys want, https://code.launchpad.net/~harlowja/cloud-init/schema-validate/+merge/231950 , probably useful for documentation purposes and validation and all that... [18:49] can be progressively expanded to include other modules [18:52] i'm ok with that. i like it even :) [18:53] harlowja, if we can make jsonschema usage optional that would be goodl [18:53] ie, wherever it loaded it it would just try/except importerror [18:53] and if not there, then say "looks good to me" [18:53] hmmmm, and then no do validation? [18:53] k [18:53] thats easy with that way of decorating stuff [18:55] if done correctly we can suck these schemas into the docs somehow [18:56] yeah. i really like that harlowja [18:56] yeah, if jsonschema isn't available, then dont' do validation. [18:57] i'm only thinking that i still care some about precise. [18:57] and probably other linuxes dont have jsonschema [18:57] RaginBajin, you there? [18:57] yep [18:58] good [18:58] smoser k, updated with optionalness [18:58] so lookoing at your mp [18:58] RaginBajin, i dont like lazy unmount [18:58] it just seems prone to error. [18:58] Which part do you mean [18:59] cloud-init admittedly should not do this... but during search-for-datasource it often mounts and unmounts devices [18:59] smoser i think i might have a new cloud-init contributor guy that might want to/be ok with flushing out all the rest of the modules there [18:59] and if a lazy unmount happened, and then another came in and mounted... [18:59] then i think unexpected behavior [18:59] RaginBajin, why did you add "lazy mode" for umount [19:00] because in FreeBSD there is no such thing as the -l option [19:00] but we didn't use -l before [19:00] so you didnt' have to do anything [19:00] it's in the code that I pulled. [19:01] http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/head:/cloudinit/util.py#L1302 [19:01] umount_cmd = ["umount", '-l', umount] [19:01] stop showing me that i'm wrong! [19:01] ha sorry [19:02] If there is a better way, I'm all ears. That's the only way I could think that would make that option work [19:03] I figured my crappy code would be critiqued and shown a better way. [19:04] bah. that code blames to harlowja [19:04] harlowja, why did we 'umount -l' [19:04] hmmm [19:04] errr [19:04] not sure, lol [19:06] Well probably because in linux it's makes systems less angry a bit.. You say just get rid of it so I can't see it, and when I'm no longer in the directory or something is using it do the work in the background [19:06] instead FreeBSD wants to choke you out till you umount it cleaning [19:06] cleanly* [19:07] RaginBajin, yeah, i understand why you want lazy [19:07] but i explicitly think its a bad idea here. [19:07] i want it unmounted now! [19:08] maybe we should have that contextmanager have a lazy=True or lazy=False option? [19:10] smoser didn't u add http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/550/cloudinit/util.py?start_revid=550#L697 ;) [19:10] before JH arrived, lol [19:11] ;) [19:12] back-in-them-olden-days [19:12] i'm not making this stuff up, harlowja [19:12] http://paste.ubuntu.com/8116527/ [19:13] i mak ea lot of stuff up. but i dont htink i'm making htis up [19:13] ;) [19:13] you even snuck it in [19:14] lol [19:14] anyway. i'm gonna take it out. [19:14] http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/revision/517 ? [19:14] lol [19:15] RaginBajin, so there is no 'mount -t auto' ? [19:16] correct [19:16] hm.. yeah, that looks bad, harlowja [19:16] why would i have done that. [19:16] :-P [19:17] You have to be explicit in everything. But they don't even name their cdrom like everyone else. iso9660 vs cd9660 [19:18] smoser u were young back in 2011 [19:18] times were differnent [19:18] had more hair, etc right? [19:18] that is true. [19:18] i hate you all. [19:19] next time i say that one of you is wrong, just accept it. [19:19] rather than proving that it is me that is wrong [19:19] lol [19:19] ;) [19:37] RaginBajin, what does platform.system() show ? [19:37] can you let me into a system RaginBajin ? i guess i can just luanch one on ec2... [19:39] 'FreeBSD' [19:44] thanks. [19:44] seems like a better long term thinking idea to be able to specify a list of types [19:45] RaginBajin, so if its not a cdrom [19:45] if its not a cd device [19:45] then we'd not try to mount with -t [19:45] It's a CDROM device [19:46] but it's called something else. [19:46] not always. [19:46] atleast in freebsd [19:46] the config drive does not have to be a cdrom [19:46] and it can be vfat formated [19:46] right. [19:46] so it could just be /dev/sdb with either iso9660 format or vfat format [19:50] RaginBajin, ok. lets say i booted this "freebsd" thing.. [19:50] how would i a.) do anything [19:50] b.) install python [19:51] python should come already installed. [19:51] It shoudl work like your linux box [19:51] just a little bit more explicitness is needed most of the time. [19:51] $ python -c 'import platform; print(platform.system())' [19:51] python: not found [19:51] http://www.daemonology.net/freebsd-on-ec2/ [19:52] check /usr/local/bin/python [19:52] actually. [19:52] /usr/local/bin/python2.7 [19:52] so you may have to create a symlink to /usr/local/bin/python [19:53] good user experience, that :) [20:10] RaginBajin, are you wliling to fix things up a bit ? [20:10] absolutely Just let me know what you need. I'm a little tight for time today, but over the weekend I can [20:10] k [20:10] i'll put comments in there. [20:21] RaginBajin, do we have a prayer at mounting windows [20:21] er... vfat on freebsd [20:28] RaginBajin, thank you [20:45] smoser I'm not sure about windows or vfat. I don't know if it has support. I assume it does [20:46] i dont know. [20:46] do my comments make sense? [20:56] I haven't looked at them just yet [20:57] k [21:06] harmw, finally, smoser looked at your MP. === gondoi is now known as zz_gondoi