[01:13] <mup> Bug #1660211 changed: MAAS fails to properly configure two NICs with different subnets <MAAS:Invalid> <https://launchpad.net/bugs/1660211>
[04:19] <mup> Bug #1633663 changed: [2.1] Unlink interfaces creates a new link <MAAS:Expired> <https://launchpad.net/bugs/1633663>
[09:16] <BlackDex> oke, i can't seem to find the correct place for the changelog
[09:16] <BlackDex> where can i find the changelog for 2.1.3, and also for the dev versions?
[10:58] <thiagolib> Hi someone problem with maas-region-controller trying to reinstall
[10:58] <thiagolib> https://pastebin.ubuntu.com/23875672/
[11:05] <pmatulis> thiagolib, this seems to come up often. dunno. roaksoax? ^^^
[11:09] <thiagolib> pmatulis: So I'm just trying to install the mass-region-controller without losing its dbconfig files.
[11:51] <jlec__> hi
[11:52] <jlec__> when deploying CentOS7 images I see "tar... mkdir: no space left on device"
[11:52] <jlec__> doing the same deploy using CentOS6 or ubuntu image works fine
[11:52] <jlec__> Any idea?
[12:42] <jlec__> roaksoax: please ping me when ever you have time.
[13:00] <BlackDex> Hello there, i have a custom ubuntu image loaded into maas. And it worked perfectly until a few weeks ago. Now it doesn't configure the network config to be DHCP or static or any thing.
[13:00] <BlackDex> How can i force maas or curtain to put the network to dhcp in /etc/network/interface?
[15:09] <jlec__> What is the current way to create custom images for Maas-2?
[15:19] <BlackDex> jlec__: you can do that several ways. What i did i just used kvm as a client, and extracted/tarred out all the data from that image, and uploaded that image to maas :)
[15:19] <BlackDex> don't forget to install cloud-init if  you want key's and network etc.. to be correct
[15:20] <jlec__> BlackDex: so the whole magic is stage 3 tarball plus cloud-init?
[15:20] <jlec__> nice
[15:20]  * jlec__ likes that
[15:22] <vogelc> jlec__: Also make sure to include curtin.
[15:23] <jlec__> vogelc: does curtin come from pip?
[15:23] <jlec__> mmh no
[15:30] <vogelc> jlec__: I extracted it from / of build using the centos provided by MAAS.
[15:31] <jlec__> That's also possible
[15:32] <vogelc> Has any successfully implemented LVM for use on Centos?
[15:33] <vogelc> During build curtin keeps trying to install grub on /dev/dm-0 and it complains it is an LVM device.
[15:33] <jlec__> vogelc: curtin is distutils installable. so easy thing
[15:35] <jlec__> vogelc: https://bazaar.launchpad.net/~curtin-dev/curtin/trunk/view/head:/doc/topics/storage.rst seems to be some support for LVM in curtin
[15:38] <vogelc> jlec__:  thats exactly what I follow.  Let me do a build and show the output.
[15:38] <jlec__> vogelc: I am also fighting with CentOS and MAAS
[15:43] <vogelc> jlec__:  here is a link to my storage config.  http://chopapp.com/#n7d4c4w1
[15:43] <jlec__> so you are really chunking it up?
[15:44] <jlec__> Might be my problem too.
[15:44] <vogelc> If I use standard partitions no LVM it works.
[15:47] <vogelc> Here is the error: grub2-install: warning: File system `ext2' doesn't support embedding. grub2-install: error: embedding is not possible, but this is required for RAID and LVM install.
[15:48] <vogelc> This is what is being run: ['chroot', '/tmp/tmpt0crqezh/target', 'grub2-install', '--recheck', '/dev/dm-0']
[16:03] <mup> Bug #1660370 opened: [UX Improvement] Surface the filter selection better on the UI <ui> <MAAS:New> <https://launchpad.net/bugs/1660370>
[17:17] <roaksoax> jlec: pong
[17:35] <g3> went to commission my nodes
[17:35] <g3> and now I see this
[17:35] <g3> [error] ceph-mon1: Unable to set any default storage layout because it has no writable disks.
[17:36] <g3> It was working, but we had to reset everything
[17:36] <g3> now it isn't finding any of the disks?
[17:58] <g3> also what username / password does MAAS create on the machine? I can't phyiscally login at the host!
[18:33] <g3> went to recommission a node
[18:33] <g3> [error] ceph-mon1: Unable to set any default storage layout because it has no writable
[18:33] <g3>             disks.
[18:33] <g3> [error] ceph-mon1: Unable to set any default storage layout because it has no writable disks
[18:48] <mup> Bug #1660418 opened: Uploading a new custom image does not remove the old image <MAAS:Triaged> <https://launchpad.net/bugs/1660418>
[18:59] <vogelc> is anyone aware of an update to knife for MAAS so it will work with api/2.0?
[19:00] <roaksoax> g3: commission your nodes with an HWE kernel
[19:00] <roaksoax> g3: the stock kernel probabaly doesnt; have the drivers
[19:00] <roaksoax> g3: and we use SSH to access a deployed system
[19:22] <jlec> roaksoax: is there some up2date guide around custom images?
[19:23] <jlec> roaksoax: I need ubuntu LTS including python, and I would like to create CentOS 7.3 images
[19:24] <jlec> roaksoax: secondly, I am trying to deploy CentOS. 6.6 works, but 7.0 with the same settings has problems with filesystem and network
[19:55] <roaksoax> jlec: what are your issues with centos7 ? the images that maas provides are the latest centos available as per their released cloud-images
[19:58] <jlec> roaksoax: interesting. They are still labelled 7.0 in Maas. As I cannot deploy them I neever could check what they really are.
[19:58] <jlec> roaksoax: How can I debug a failed deploy if the network setup fails?
[19:59] <jlec> roaksoax: Is there anything special I need ot do to the FS? I just set the whole disk to be / with xfs. Any suggestions?
[20:04] <roaksoax> jlec: yeah there's a bug for that
[20:05] <roaksoax> jlec: maas provides an installation log (you can see it in the WebUI or get it with: "maas admin node-results read system_id=4tbcnh result_type=1 | grep data | cut -d"\"" -f4 | base64 --decode"
[20:06] <roaksoax> maas <admin> node-results read system_id=<system_id_of_machine> result_type=1 | grep data | cut -d"\"" -f4 | base64 --decode
[20:06] <roaksoax> jlec: do that on a failed deployment machine
[20:06] <roaksoax> jlec: or you ncan get rsyslog form /var/log/maas/rsyslog/<machine-=name>
[20:07] <jlec> roaksoax: thanks. I will look into that tomorrow. Perhaps I can figure out what it is.
[20:08] <roaksoax> jlec: yeah, let me know./ We know there are some deployment failures with centos when deploying with UEFI
[20:08] <roaksoax> which we are working on fixing
[20:09] <jlec> roaksoax: this is with a old IBM blade setup
[20:11] <jlec> roaksoax: only BIOS and manual power cycle
[20:12] <vogelc> roaksoax: have you run into an error where curtin trys to install grub2 on /dev/dm-0 when using LVM on CentOS?  This does not happen when using flat storage.
[20:15] <roaksoax> jlec: strange!
[20:15] <roaksoax> vogelc: maas doesn't support lvm+centos
[20:18] <g3> roaksoax, how do you specify which kernel to use when commissioning? I've only seen that option via deploy
[20:19] <roaksoax> g3: in the settings page you can select which kernel you want to use.
[20:20] <g3> Oh interesting
[20:20] <g3> I'm curious as to why it was working before?
[20:21] <vogelc> roaksoax: Is lvm and centos support on the roadmap?
[20:42] <mup> Bug #1660439 opened: MAAS should automatically "vacuum" the database to remove deleted images. <MAAS:Triaged> <https://launchpad.net/bugs/1660439>
[20:42] <mup> Bug #1660440 opened: MAAS should notify users when boot image storage space is low <MAAS:Triaged> <https://launchpad.net/bugs/1660440>
[20:51] <roaksoax> vogelc: it is but not in the next 6 months, we are prioritizing other work atm
[22:27] <netmonk> hello does it works with something different than ubuntu ?
[22:28] <netmonk> too nice to be good
[22:28] <netmonk> what if i want to deploy centos or arch ?
[22:29] <netmonk> im stuck to deploy only ubuntu ?
[22:56] <g3> roaksoax, unfortunately that didn't work..
[23:17] <g3> hmp
[23:34] <roaksoax> netmonk: you can deploy centos
[23:34] <roaksoax> netmonk: maas provides centos images
[23:35] <g3> Are you chatting with me?
[23:35] <roaksoax> g3: what didn't work? maas discovering storage on a HWE kernel ?
[23:35] <g3> Yes you are correct
[23:35] <g3> From the commissioning output of lshw it sees all of the drives
[23:36] <roaksoax> g3: what types of drives are these ?
[23:36] <g3> but according to the logs in /var/log/maas it states that it is Unable to set any default storage layout because it has no writable disks.
[23:36] <g3> it was working
[23:36] <g3> but I swizzled some subnet things then started over
[23:36] <g3> they are sata ssd's
[23:36] <g3> and a pcie ssd
[23:36] <g3> and on the OSD machines there are 12 sata spinners, 2 sata ssd's and 2 pcie ssd's
[23:37] <g3> All 5 machines are reporting the same error of the disks are not writable
[23:37] <roaksoax> g3: uhmm strange... did you by anychance manually delete those disks and commissioning wont recognize them again ?
[23:37] <g3> I didn't
[23:37] <g3> I deleted the machines
[23:37] <g3> and started over
[23:37] <roaksoax> g3: can you pleas efell a bug with details on what errors it is showing you and attaching logs so we can look into that ?
[23:37] <g3> I can
[23:38] <g3> bug comming but also a forumn post
[23:38] <g3> https://ubuntuforums.org/showthread.php?t=2351047
[23:38] <g3> well basically there is a bug already
[23:38] <g3> https://bugs.launchpad.net/maas/+bug/1573084
[23:48] <roaksoax> g3: are you using MAAS 1.9 ?
[23:49] <g3> no I jsut realized that
[23:49] <g3> 2.1.3
[23:50] <roaksoax> g3: please do file a bug and attach to it the cloud-init logs which you can get from /var/log/maas/rsyslog/<machine-name>
[23:54] <mup> Bug #1660495 opened: Can't commission node because it has no writable disks <MAAS:New> <https://launchpad.net/bugs/1660495>
[23:58] <g3> done done and done