[00:42] <AskUbuntu> Squid Forbidden erro | http://askubuntu.com/q/488680
[07:23] <rvba> bigjools: thanks for looking into it.
[07:24] <jtv> Well done finding the cause of the slowness, bigjools.
[07:24]  * bigjools bows deeply
[07:25] <bigjools> not hard once profiled tbh
[07:25] <rvba> bigjools: the code change is not that complicated; but the diff makes it look so.
[07:25] <jtv> You got profiling working?
[07:25] <bigjools> jtv: yes, seem my email to maas-devel
[07:25] <bigjools> see*
[07:28] <bigjools> rvba: can we do a quick hangout and you can take me through your branch?
[07:28] <rvba> bigjools: sure
[07:29] <bigjools> rvba: just use daily hangout
[07:29] <rvba> k
[07:29] <rvba> allenap: testing your fix from yesterday now: http://d-jenkins.ubuntu-ci:8080/view/MAAS/job/utopic-adt-maas-manual/65/
[08:01] <rvba> allenap: gmb: Test in the CI lab is not finished yet… but we've passed the point where it was failing yesterday so I'm reasonably optimistic allenap's branch is sound.
[08:02] <gmb> rvba: Molto buona!
[08:06] <bigjools> we need to sort this out too https://bugs.launchpad.net/maas/+bug/1315051
[08:11] <allenap> rvba: \o/
[08:11] <allenap> rvba: Thanks for doing that, I really appreciate it.
[08:11] <rvba> allenap: Thank *you* for fixing it!
[08:16] <gmb> allenap: So ISTM that the best dynamic is this: I find broken things, other people fix them.
[08:18] <allenap> Graham "I see broken things” Binns
[08:18] <allenap> gmb: I’m using IPSet.__contains__(an_ip_address). Was the slow code using __contains__(an_ip_network)?
[08:19] <gmb> allenap: No it was is_disjoint(IPSet()…)
[08:19] <gmb> allenap: Could go either way really.
[08:19] <gmb> allenap: But IPRange seems to be far more efficient
[08:19] <gmb> Because I think it’s just doing maths behind the scenes
[08:20] <AskUbuntu> How do I grant access to someone to a MAAS node which is running a a private network? | http://askubuntu.com/q/488768
[08:20] <gmb> Whereas IPSet seems to walk the set.
[08:20] <rvba> allenap: As expected, the test was successful.  Let's get this branch MPed, reviewed and landed please.
[08:21] <rvba> (I'd like to merge trunk and do some more testing of my no-zone-pregeneration branch)
[08:22] <allenap> rvba: Okay. I need to write some tests :)
[08:23]  * gmb runs the tests with a deliberately large network… might need to go and cook a roast at this speed…
[08:25] <gmb> Crumbs, I might actually have killed my VM by testing with an IPv6 /16…
[08:25] <allenap> gmb: That’s going to take an awfully long time to complete. You’ve got time to start a farm and the breed the livestock before you start the roast.
[08:27] <gmb> allenap: Load was 30, dropping like a stone now that I’ve kill -9’d the test runner
[08:29] <gmb> Hah, yes.
[08:29] <gmb> allenap: New code = 7 seconds with a /16.
[08:29] <gmb> Old code … never completed, load 30, moribund VM
[08:30] <allenap> gmb: If we were American I’d do a high-5, but as we’re British we’ll have to make do with a virtual belly barge.
[08:30] <gmb> lol
[08:39] <gmb> allenap, rvba, bigjools: For the love: https://code.launchpad.net/~gmb/maas/speed-up-clean_ip_ranges/+merge/224766
[08:49]  * gmb -> off to find a power socket.
[09:09] <breakum-binns> Yay, faster test suite is faster.
[09:24] <jtv> \o/
[09:24] <jtv> Maybe we just need the existing code for "do these networks overlap" generalised to network ranges.
[10:26] <g0ldr4k3> hello everyone
[10:28] <g0ldr4k3> bigjools: coming back with about MaaS RC and CC while I'm reading the guide I found this "maas-cluster-controller - controls a group (“cluster”) of nodes including DHCP management." the link is https://maas.ubuntu.com/docs/install.html#pkg-install. For that I thought that CC manages the nodes.
[10:29] <g0ldr4k3> bigjools: anyway I've try to install maas-region-controller's package via apt on a clear Ubuntu Svr, but received the same error during the installer.
[10:31] <g0ldr4k3> bigjools: I'd like to separate the service on more Svr
[11:42]  * gmb -> lunch
[12:44] <allenap> gmb, rvba: Do either of you have time for a review? https://code.launchpad.net/~allenap/maas/ipv6-tftp-compat/+merge/224805
[12:45] <breakum-binns> allenap: Already on it.
[12:45] <breakum-binns> FFS
[12:45]  * gmb stabs colloquy
[12:45] <allenap> gmb: Thanks!
[12:53] <gmb> allenap: r=me with comments
[12:53] <allenap> gmb: Ta muchly.
[14:51] <rvba> allenap: I see you landed your ipv6-twisted branch but the build is still broken.  This is weird since my test from this morning was successful.
[14:52] <rvba> allenap: http://paste.ubuntu.com/7711494/
[14:53] <allenap> rvba: Ah, that’s interesting. I changed the code to only avoid link-local addresses. ::1:69 is a new thing to me. I wonder if it’s meant to be [::1]:69, i.e. 69 is the port?
[14:54] <allenap> I’ll investigate as soon as I’ve finished reviewing your branch.
[14:55] <rvba> Cool.
[14:56] <rharper> in maas do tag kernel_opts values concatenate when tagging a machine with more than one tag ?
[14:59] <rvba> rharper: No, you can have only one tag with a non-empty kernel_opts field.
[14:59] <rvba> It's probably a bit suboptimal, please file a bug if this matters to you.
[15:03] <allenap> rharper: The first tag, lexicographically, that defined kernel options is used, and these override the global kernel options. I agree this isn’t super-duper.
[15:04] <allenap> s/defined/defines/
[15:08] <rharper> rvba: allenap: ok -- yeah, it seems the last tag applied with kernel_ops wins -- which is OK for now, I just superset my tags, but yes, it would be very good to allow to compose kernel params based on tag concat
[15:09] <rvba> rharper: like allenap said, it's not the last tag that wins.  It's the first one, sorted by name.
[15:10] <rharper> rvba: ah, thanks for the clarification
[15:10] <rharper> it appears last in the list of tags when you look at the node on the ui
[15:10] <rharper> but it also is the first alphab-wise as well
[15:11] <allenap> rharper: It’s probably being sorted in the UI by database ID instead of name, or not sorted at all. That’s a bug :-/
[15:13] <rharper> hehe
[15:14] <rvba> allenap: (re ipv6) yes, ::1 is the loopback address.
[15:14] <allenap> rvba: Is CI running as root? Or is it using authbind to allow port 69? Maybe we just need another authbind rule.
[15:14] <rvba> allenap: I /think/ it's running as the jenkins user.
[15:14] <allenap> matsubara: ^
[15:15] <matsubara> allenap, tests run as root
[15:15] <rvba> Ah, right, inside the VM the tests run as root.
[15:16] <rvba> But the VM itself doesn't run as root on the host machine right?
[15:16] <allenap> matsubara: Is there an IPv6 localhost configured, i.e. ::1 on lo.
[15:16] <allenap> ?
[15:16] <matsubara> rvba, nope, the VM is started as the jenkins user
[15:16] <rvba> Okay.  Makes sense.
[15:17] <rvba> allenap: Why did my first test succeed?  Did you change your code significantly after that test?
[15:17] <matsubara> allenap, I'll need to check that but I don't think so (unless is configured by default)
[15:20] <allenap> rvba: I changed it to exclude link-local addresses, instead of only including all IPv4 and a limited subset of IPv6 addresses. The IPv6 localhost would have been excluded before. I can change it again, but I’m puzzled as to why it’s a problem.
[15:22] <rvba> allenap: yeah, it's probably best if we work out why it's failing.
[15:22] <matsubara> allenap, https://pastebin.canonical.com/112637/
[15:28] <rharper> rvba: allenap: bug https://bugs.launchpad.net/maas/+bug/1335175
[15:29] <rvba> rharper: ta
[15:33] <rharper> I've got a node that when it's started, it does nic renaming (eth0 -> em1) (trusty amd64 on some hp box) -- this is fine for most cases, but when we attempt to start a container on the machine,  it fails because br0 isn't available (it was configured to make br0 out of eth0, which is now em1)
[15:33] <rharper> is this a maas or a juju bug ?
[15:33] <rharper> or both ?
[15:36] <rharper> adding net.ifnames=1 biosdevname=0 to the kernel_opts  tag of the node fixes it, but that's where I ran into the bug I just filed as I need other options from another tag
[15:36] <rvba> I see.
[15:38] <blake_r> rharper: it could be a curtin issue as well
[15:38] <blake_r> rharper: curtin could touch /etc/udev.d/75-persistent-net-generator.rules
[15:39] <rharper> blake_r: ok, then I'll file it as a maas bug and we can add juju if needed
[15:49] <rharper> blake_r: bug  https://bugs.launchpad.net/maas/+bug/1335179
[16:06] <allenap> matsubara: Is authbind used in the CI virtual machine?
[16:09] <matsubara> allenap, If that's used by the services installed by the maas packaged, then yes. The CI VM doesn't do anything other than what the package provides
[16:11] <matsubara> allenap, s/packaged/packages/
[16:11] <allenap> matsubara: Ah ha, okay, I’ll look at that then.
[16:12] <matsubara> allenap,     exec /usr/bin/authbind --deep /usr/bin/twistd -n --uid=maas --gid=maas --pidfile=/run/maas-pserv.pid --logfile=/dev/null maas-pserv --config-file=/etc/maas/pserv.yaml
[16:12] <matsubara> allenap, that's from debian/maas-cluster-controller.maas-pserv.upstart
[16:16] <kirkland> can someone help me with a command line "maas" invocation to set the global kernel parameters?
[16:32] <roaksoax> kirkland: maas admin maas set-config name=kernel_opts value="console=ttyS0"
[16:32] <roaksoax> kirkland: maas <user> maas set-config name=kernel_opts value="console=ttyS0"
[16:32] <kirkland> roaksoax: fantastic, thanks
[18:48] <AskUbuntu> When I trey to add new node to MAAS, dose the node need physical internet connection? | http://askubuntu.com/q/488968
[21:02] <AskUbuntu> How Can I set the MAAS Network adaptor to do nat on ubuntu 14.04? | http://askubuntu.com/q/489029