[04:02] using ppa:maas/stable on xenial, doing createadmin, the first time it failed due to proxy to fetch ssh keys, so i do it again, then it fails with an error message about fetching duplicate keys [04:03] which is fine if it's the last thing, but not fine if it's skipping some other setup as a result [04:04] (and it doesn't tell me) [05:27] * hallyn looks around for a no_proxy setting [07:11] hallyn: don't worry, fetching the keys comes last ;-) [08:06] Hey folks, is it possible to migrate from a deb install of a region/rack controller to a snap install? Didn't see anything in the docs about this. I guess the db will be the hard part. === frankban|afk is now known as frankban [12:45] zeestrat: maybe dump the database, and try to import in into the snapped environment. But i don't know if that is possibl [12:45] e [13:20] zeestrat: i wouldn't move the DB inside the snap [13:20] zeestrat: i would use external db and maas from the snaps [13:21] zeestrat: that way you, if in the future you decide to grow multiple regions (HA) [13:21] you are not constrained by having the DB inside a single snap [13:36] roaksoax: Thanks. That sounds like a good approach. [15:00] roaksoax: That is nice indeed. Normally if you install maas it installs the database it self and configures it self. But is this different for the snap? [15:01] BlackDex: the snap follows hte ssame principles as in the package [15:01] BlackDex: the sanp you tell it how you want to configure it [15:02] BlackDex: the same way you can do in MAAS installed by the packages, or as you can do in the snap, is forward it to a external DB [15:02] ah oke :) [15:03] something to tryout in the future :) [15:03] good to know === frankban is now known as frankban|afk === newell is now known as Guest25364 [17:49] hey guys, i'm getting https://paste.ubuntu.com/p/Pzj8ycBrXc/ [17:49] is there a way i can find out what the actual issue is? is it ECONNREFUSED, or a self signed cert or something? [17:52] Bug #1748875 opened: Unable to deploy Bionic on bare-metals with MaaS 2.3.0

[17:55] Bug #1748875 changed: Unable to deploy Bionic on bare-metals with MaaS 2.3.0

[18:06] tych0: that seems that there would be no rack controllers that can power on a machine [18:06] tych0: e.g. it cannot directly reach it [18:06] tych0: /win 4 [18:06] err [18:07] roaksoax: ok, but what does that mean? is it ECONNREFUSED? or dns doesn't resolve? [18:08] i can wget the url from the rack controller and i see output i expect [18:08] so i'm not sure what's idfferent from maas' environment [18:11] tych0: what power type is it ? [18:11] UCS manager [18:33] tych0: could be that there's a bug with the ucs manager power type. Are you using IP or DNS ? [18:34] IP [18:34] it's also possible that i have it pointed at the wrong rest API [18:35] but i don't know how to tell really [18:35] tych0: if you are using ssl, then yes, since we dont support ssl for power params [18:36] ah, so i need a http:// url? [18:37] this box gives a 301 to the https:// url :( [18:39] tych0: yeah http:// url indeed [18:41] ok, so if it enforces https with a redirect, i'm screwed? [18:42] ugh [18:42] is it all simple scripts, can it be easily hacked around/ [19:04] Bug # changed: 1585841, 1642916, 1702527, 1713771, 1729555, 1730626, 1730751, 1730991, 1733923, 1734347, 1735840, 1738478, 1741574, 1741915, 1742137, 1743005, 1744802 [22:45] roaksoax: so i just set up an ssl stripping proxy to try and use maas [22:45] and got, [22:45] Error:__init__() missing 4 required positional arguments: 'code', 'msg', 'hdrs', and 'fp' [22:46] i'm trying to find a full stack trace, but no luck so far [22:47] tych0: honestly i wouldn't know at this point, as last time we worked with one of those was a while ago and we dont have one in CI [22:47] tych0: maybe ucs has updated their API whjich is causing the breakage ? [22:48] ah, ok [22:48] i guess if it's not in CI then it's probably not going to work :) [22:49] roaksoax: ok, cool, thanks! [22:50] ok but surely [22:50] surely there is a way to get more detailed logs about where the %$%$% it's failing [22:51] tych0: are you able to tell whether it manages to connect to the CIMC and start sending msgs? [22:51] (i suppsoe maybe tcpdump is the way to go here) [22:52] (if maas doesn't want to help) [22:52] mitmproxy tells you [22:54] ah [22:54] ok so the connection is going through, but it fails while talking? [22:54] so mitmproxy should be able to dump the text of the conversation right? [22:54] no, it's not going through [22:54] tych0: are you pointing maas to their xml api endpoint ? [22:55] (but i can see other connections, for example if i use https, maas will complain about a self signed cert) [22:55] roaksoax: well, i'm pointing it to my ssl stripping proxy, which forwards the request with https on to the actual XML api [22:56] tych0: right, but for example, looking at the code I see that we add 'nuova' to the URL [22:56] oh, you guys add that? [22:56] tych0: yeah [22:56] i've added it in my url [22:56] self.api_url = urllib.parse.urljoin(self.url, 'nuova') [22:56] i'll remove it and check [22:57] lol [22:57] yeah, same erorr about __init__() missing 4 required positional arguments: 'code', 'msg', 'hdrs', and 'fp' [22:58] oh - so is that error being passed by the cimc/ucs back to maas? [22:58] if so then (a) i guess i see and (b) all is lost [22:58] no [22:58] i think the error is in maas before it ever tries to talk to the cimc [23:01] oh, whee [23:01] got them to talk, now i get [23:01] XML PARSING ERROR: The XML API method configResolveClass, has unrecognized child element(s) in the xml request. [23:01] tych0: that seems to me that the xml api could have changed since first enabled [23:02] yeah, that error definitely seems like it [23:04] value="2F5D552B-6BD5-4E2D-9B8B-E707032096FA"/> [23:04] is the thing that it doesn't like parsing. [23:07] oh, no, sorry [23:07] that's the request, derp. [23:07] [23:07] the error is from the cimc itself [23:07] it doesn't like the request maas is sending, i guess?