[07:21] <jam> mwhudson: poke about your recent patches
[07:50] <rvba> gmb: I'm done with the commissioning tests; where are you at with the enlistment test?  We should probably briefly talk about that because I suspect those tests are almost identical.
[07:51] <rvba> Re: "allenap | 17:58:18> rvba: What's the direct_network setting for in maas-test?"
[07:52] <rvba> allenap: that's the definition of the node's network, the one MAAS' DHCP manages.  If you leave it blank '192.168.2.0/24' will be used, but you might want to use a specific network for some reason, hence the presence of the (optional) parameter.
[07:57] <rvba> gmb: commissioning *test*
[08:01] <bigjools> rvba: can gmb give you his tests migration task? I want him to sort out a caching proxy.
[08:01] <rvba> bigjools: no problem
[08:01] <gmb> Cool.
[08:01] <bigjools> rvba: you are great
[08:02] <gmb> rvba: I haven't gotten much further than the code you passed me yesterday — it was close to my EoD and the MTTF on this has been killing me.
[08:02] <bigjools> now, I need to know why allenap put "str=None" in the templates for our code
[08:02] <gmb> bigjools: To force us to use unicode and byte strings?
[08:02] <bigjools> since I think I need it to convert a byte to a char
[08:02] <gmb> Ew.
[08:03] <bigjools> more to the point, I have a bytes buffer from a DHCP server
[08:03] <bigjools> and 4 of those bytes are an IP address I want to get out of it
[08:03] <bigjools> str(byte) works perfectly.
[08:03] <rvba> gmb: all right, no worries.  Like I said, the enlistment test is very similar to the commissioning test.  I'm counting on you for the review though ;)
[08:03] <gmb> rvba: Deal.
[08:04] <gmb> rvba: bigjools and I have agreed that I'm about to engage in anger-driven development to make maas-test faster — or at least more informative.
[08:04]  * gmb *cracks knuckles*
[08:04] <rvba> Sounds like a very good idea.
[08:04]  * bigjools winces
[08:05] <bigjools> rvba: yeah it sounds like you are hacking around slowness at the moment
[08:05] <rvba> Indeed.
[08:05] <bigjools> hence me wanting gmb to do this proxy
[08:05] <bigjools> since I believe that's what everyone said was best ;)
[08:06] <rvba> I agree :).
[09:22] <allenap> bigjools: bytes is a synonym for str in Python 2.
[09:25] <bigjools> allenap: I can't do a straight replacement though, my code fails
[09:25] <bigjools> good morning BTW
[09:25] <allenap> bigjools: Morning! Want to share you code and I'll take a look?
[09:26] <bigjools> allenap: in Python3 I can do:
[09:26] <bigjools> buffer = b'\x10\x00\x00\xaa'
[09:27] <bigjools> and then something like
[09:27] <bigjools> '.'.join(str(byte) for byte in buffer)
[09:27] <bigjools> et voila I get an IP
[09:35] <mwhudson> jam: oops, bed time
[09:35] <jam> mwhudson: np. I think I've queued all your stuff to land
[10:05] <bigjools> allenap: gah only python3 has an ipaddress module
[10:05] <rvba> bigjools: just use netaddr then.
[10:05] <rvba> (If that suits your needs)
[10:06] <bigjools> rvba: where?
[10:06] <rvba> What do you mean "where?"?
[10:06] <bigjools> where is netaddr?
[10:07] <bigjools> hmm maybe socket.inet_ntoa
[10:07] <rvba> We use it in MAAS already.
[10:08] <rvba> http://pythonhosted.org/netaddr/
[10:08] <bigjools> >>> socket.inet_ntoa(buffer[245:249])
[10:08] <bigjools> '16.0.0.170'
[10:08] <bigjools> \o/
[10:09] <bigjools> allenap: FYI ^
[10:09] <allenap> bigjools: Yay!
[10:11] <bigjools> allenap: is it worth me taking the time to make a class look like a byte buffer or should I just stick with a property on the class to get the buffer out ...?
[10:20] <bigjools> allenap: also this worked: '.'.join('%d' % ord(char) for char in buffer[245:249])
[10:20]  * bigjools EODs
[10:22] <allenap> bigjools: Stick a property on.
[10:57] <gmb> rvba: mass-test took 34m41.123s from start to test_dummy() on Canonistack lcy01. Haven't got timings for individual components yet but packages + PXE files definitely took the longest, anecdotally.
[10:57] <gmb> That was with trunk, btw.
[11:30] <rvba> gmb: it took 40m5.924s to running locally (and I don't have the same bandwidth canonistack has *including* the enlistment and the commissioning tests.
[11:30] <rvba> gmb: care to have a look btw? https://code.launchpad.net/~rvb/maas-test/real-tests/+merge/195748
[11:47] <gmb> rvba: Interesting. And it looks like uvt-kvm create tends to get stuck in a read() for long periods of time; can't figure out why yet.
[11:47] <gmb> rvba: Anyway, I'll look at your branch now.
[12:04]  * gmb -> lunch and errands
[13:35] <HumpyDumpy> I just installed ubuntu MAAS, logged in for the first time to the web gui, and it says I have no boot images
[13:36] <HumpyDumpy> how do I get them, do i do it from the gui or from command line?
[13:36] <HumpyDumpy> cos I don't want to break my srv
[13:41] <HumpyDumpy> http://pastebin.com/ncnDLfjV
[13:41] <HumpyDumpy> anyone?
[13:41] <HumpyDumpy> look at that and help humpy
[13:43] <HumpyDumpy> it's ok, false alarm... i googled it
[13:43] <HumpyDumpy> ;D
[13:54] <matsubara> HumpyDumpy, yes, you can trigger that with the command line or from the gui. (through the cli, run as root maas-import-pxe-files or click the Import boot images in the web ui)
[13:54] <HumpyDumpy> yay, someone woke up
[13:55] <HumpyDumpy> what PXE settings do I need in my windows dhcp to make MAAS work?
[13:57] <HumpyDumpy> OMG how many boot files are there
[13:58] <HumpyDumpy> >:(
[16:28] <HumpyDumpy> i got my first srv up and running, trying to add a node to it, the node is accepted and ready to do it's thing, but i can't PXE boot it, it's timing out.
[16:29] <HumpyDumpy> they talk about avahiboot but i can't find anything on it
[16:29] <HumpyDumpy> i added in my dhcp the boot ip for the maas srv and filename
[16:57] <HumpyDumpy> i had to restart the pserv
[16:57] <HumpyDumpy> ;D
[16:58] <HumpyDumpy> i got a node installed now
[16:58] <HumpyDumpy> woot
[21:57] <bigjools> HumpyDumpy: looking at backscroll, I suggest you read http://maas.ubuntu.com/docs