[01:41] <benlake> roaksoax: noted, will add.
[01:44] <benlake> kiko: ThiagoCMC’s situation sounds familiar… but why would an update cause a redisovery of IPs, or is that more likely related to him probably adding subnets or a fabric or something?
[02:11] <xygnal> roaksoax i'll be around if you do come back, i'd like to verify what is needed in my curtin config to ensure it uses them.  Right now my storage blocks are not being used, though its not throwing errors either in the logs.
[08:41] <chrome0> Hi maas maintainers, do you have a timeline when you plan to pronounce 2.2 stable? I believe it's  currently labeled devel as it's in ppa:maas/next right?
[09:03] <chrome0> But maybe I'm confused about that devel label, as maas 2.2 is actually released with Zesty?
[12:51] <jamespage> congrats on the 2.2 release MAAS folks - looks awesome!
[13:07] <benlake> jamespage: where are you seeing 2.2 as released? I ask because I’m actually quite confused as to what package version is current and with what release
[13:08] <jamespage> benlake: https://docs.ubuntu.com/maas/2.2/en/release-notes
[13:08] <benlake> my 16.04.02 box installed 2.1.3 a few days ago, but according to packages.ubuntu.com it should only be 2.0.0-beta3   O.o
[13:08] <jamespage> I think its still +1 week off being in the stable PPA - roaksoax would be able to confirm that
[13:09] <jamespage> benlake:
[13:09] <jamespage>  maas | 2.1.3+bzr5573-0ubuntu1~16.04.1    | xenial-updates   | source, all
[13:09] <jamespage> 2.1.3 is in xenial-updates
[13:09] <benlake> even still I look at https://launchpad.net/maas/+packages and it says xenial should be at 2.1.5, but I don’t show an upgrade path to it via apt
[13:09] <jamespage> I think 2.2.x will end up there once its been through SRU
[13:10] <jamespage> benlake: try "rmadison maas"
[13:10] <jamespage> its a easier view of what is where
[13:10] <jamespage> 2.1.5 is in xenial-proposed ATM
[13:11]  * chrome0 would be eager to know when 2.2 lands in stable too 
[13:12] <benlake> oic https://launchpad.net/ubuntu/xenial/+source/maas
[13:12] <benlake> what is rmadison?
[13:29] <roaksoax> benlake: http://www.roaksoax.com/
[13:29] <roaksoax> benlake: http://www.roaksoax.com/2017/05/maas-2-2-0-released
[13:32] <benlake> ah, thanks
[13:40] <chrome0> aluria : ^^ maas 2.2 to be avail in maas/stable next week and in xenial in the following weeks
[13:41] <chrome0> roaksoax : typo alert, you probably meant "has _now_ been released..."
[13:42] <aluria> chrome0: thx for the heads-up
[13:54] <roaksoax> chrome0: thanks!
[13:55] <chrome0> nw at all
[13:56] <xygnal> roaksoax hello :)
[13:56] <roaksoax> xygnal: howdy! i'm otp for the next ~2-3 hours, but share with me whta you have and i can look as meetings go
[13:57] <xygnal> roaksoax: do I need more than a storage block to convince curtin to install on a different disk?  Just setting the disks, partitions, and mounts does not seem to be enough.
[14:02] <roaksoax> xygnal: checking that now.
[14:02] <xygnal>  roaksoax my co-worker says those blocks appeared to work, but only if the boot disk in the servers BIOS was set to boot that disk.
[14:05] <xygnal> roaksoax we also can't use the same hard-coded disk every time for every model.   Can we do some kind of pre-script for storage that grabs that and makes it a variable?
[15:39] <roaksoax> xygnal: what do you mean with the latter ?
[15:40] <xygnal> roaksoax storage section in curtin needs a hard-coded device name
[15:40] <xygnal> roaksoax but we have different models, where the device name woud not be identical
[15:40] <xygnal> roaksoax can we gather that information and pass it into curtin?
[15:40] <xygnal> so that it is no longer a hard coded disk path
[15:43] <roaksoax> xygnal: ok, so in the preseed you have access to the 'node' objects
[15:44] <xygnal> i've never played with tthose.  link me a example?
[15:45] <roaksoax> xygnal: i'll give you in a sec. So If I understand correctly
[15:46] <xygnal> I can set boot and root in the API, they just dont apply.  If I can tell Curtin to use that information, we essentially get around the 'not supported' problem.
[15:48] <ThiagoCMC> Guys, MaaS Next is weird!
[15:48] <ThiagoCMC> It is failing to test a server...
[15:48] <ThiagoCMC> Error, `smartctl --xall /dev/sda -d sat` returned 24! See the smartctl man page for return code meaning
[15:48] <ThiagoCMC> Say what?
[15:48] <ThiagoCMC> =/
[15:49] <ThiagoCMC> blade was just working fine until I tried to re-commision it
[15:51] <kiko> have you tried doing what the message suggests?
[15:51] <ThiagoCMC> yep... researching but, no clue
[15:52] <xygnal> start rescue mode on the node and try to execute that command via ssh
[15:52] <ThiagoCMC> Is there any way to ignore that?
[15:52] <ThiagoCMC> cool!
[15:53] <roaksoax> ThiagoCMC: that could indicate that there's an error with your disk :) but also if you havea a full pastebin we could look would be great
[15:53] <roaksoax> ThiagoCMC: that said, you can decide not to perform hardware testing on commissioning
[15:53] <ThiagoCMC> Hmm...
[15:56] <roaksoax> ThiagoCMC: iy could also mean smart is not supported on your disks
[16:38] <mcuenca> hi everyone! maas region controller and maas rack controller are created under ".maas" dns domain. Is it possible to change that?
[16:59] <roaksoax> mcuenca: https://bugs.launchpad.net/maas/+bug/1686678
[17:01] <mcuenca> roaksoax: thanks! so, it is a bug, no problem. For sure the guys will resolve it. I'll try a workaround using CNAMEs... should work, right?
[17:02] <roaksoax> mcuenca: yes it should work
[17:03] <mcuenca> roaksoax: thanks! :D
[17:03] <mcuenca> roaksoax: I'll try in 30 minutes and I'll let you know
[18:04] <mcuenca> roaksoax: solved :) thanks!
[19:13] <ThiagoCMC> Is there an easy way to move a VLAN / subnet from fabric-1, to another fabric?
[19:55] <pmatulis> ThiagoCMC, did you try via the CLI?
[20:06] <ThiagoCMC> pmatulis, not really, I'm not familiar with CLI... I'll research about it! Tks!
[22:36] <benlake> so, if you toast your network connection to a server after deployment, there isn’t a password for the ubunut user to use via IPMI KVM, is there?
[22:37] <benlake> I get SSH is public key only, but is the expectation if there is a desire to directly access the machine, you should set the password after deployment?
[22:52] <mpontillo> benlake: that is correct
[22:52] <benlake> thanks
[22:54] <mpontillo> benlake: np. with kvm console access you should be able to recover in a normal way, such as init=/bin/bash on the kernel command line or some similar incantatin (haven't had to do that in awhile so I don't remember exactly)
[22:54] <mpontillo> *incantation
[22:54] <mpontillo> of course, that assumes you'd be okay with rebooting
[22:54] <benlake> gotcha
[22:56] <mpontillo> benlake: right, I think something like this still applies, even though the post is pretty old. https://askubuntu.com/questions/24006/how-do-i-reset-a-lost-administrative-password
[22:56] <mpontillo> (heh, init=/bin/bash, that was a trick I used probably 20 years ago.)
[23:15] <benlake> with devices you can define a parent. Is there not a way to assign a machine to machine relationship similarly?
[23:16] <benlake> use case being kvm guests. would be nice to have the metadata as to which machines will die with the base machine
[23:28] <benlake> anyone else being bit by the bridge forwarding delay “out of range” issue? supposedly a fix is in kernel v4.4.8 - that’s a ways off for xenial, no?
[23:42] <catbus1> Hi, a question about the IP mode in https://docs.ubuntu.com/maas/devel/en/installconfig-commission-nodes#post-commission-configuration
[23:44] <catbus1> If it's a managed subnet and if auto assign is selected, the node will get the IP not in the reserved IP range. But if it's dhcp, the node will get the IP from the reserved dynamic IP range?
[23:45] <benlake> Re: forwarding delay. Value must be less than “hello” time; unless STP is off.
[23:45]  * catbus1 finds managed vs unmanaged and reserved IP range confusing. 
[23:46] <benlake> catbus1: I do as well.
[23:46] <benlake> so auto assign uses anything not reserved
[23:46] <benlake> dynamic range is for DHCP
[23:46] <catbus1> another question, can I assume the fabric will show up as 1:1 mapping to the physical interfaces on the machine?
[23:46] <benlake> reserved range is unused by MAAS
[23:47] <catbus1> but reserved IP range is used by MAAS and the subnet is in unmanaged mode.
[23:48] <benlake> when you say “is used by maas”, where did you see maas us it?
[23:48] <catbus1> *when* the subnet is in unmanaged mode.
[23:48] <catbus1> "When management is disabled for a subnet, the definition of a reserved IP range differs from the managed mode. Here, a reserved IP range tells MAAS to only allocate addresses from this range (via 'Auto assign'). In addition, DHCP will never lease addresses from an unmanaged subnet." from https://docs.ubuntu.com/maas/devel/en/installconfig-network-subnet-management
[23:48] <benlake> what version of maas are you using?
[23:49] <benlake> I’m 2.1.3 and I have no “mode” for a subnet.
[23:49] <catbus1> I am on 2.2
[23:49] <mup> Bug #1693644 opened: Enlistment does not validate min_hwe_kernel <MAAS:In Progress by ltrager> <https://launchpad.net/bugs/1693644>
[23:49] <benlake> oic, then that’s new to me.
[23:50] <benlake> but reading that, it apparently changes the “reserved range” that would otherwise not be used in “managed” mode, to the range that will explicitly be used in “unmanaged” mode.
[23:51] <catbus1> yes.
[23:51] <benlake> that ^^ is the answer to your original question :)
[23:51] <catbus1> but if it's unmanaged, why is maas giving out ip addresses?
[23:51] <benlake> did you reserve a range from this unmanaged subnet?
[23:52] <catbus1> I didn't do anything yet, I am trying to understand first what's the expected behavior.
[23:52] <benlake> if so, then MAAS will use it for any interface using “auto assign” (at least that was the 2.1 term)
[23:53] <benlake> if you don’t want an unmanaged subnet to be used in any way, then just defined it as unmanaged and leave it alone