bradm | bigjools: you about? | 00:20 |
---|---|---|
bigjools | bradm: yarp | 00:20 |
bradm | bigjools: whats the right way to reinstall a node in MaaS? should I stop it, then delete, then reboot it to re-enlist? or should just stop, commision do a reinstall? | 00:21 |
bigjools | bradm: what do you mean by "reinstall" ? | 00:21 |
bigjools | exactly, I mean | 00:22 |
bradm | bigjools: I want to format the disks and reuse the server | 00:22 |
bigjools | what is the desired outcome | 00:22 |
bigjools | ok | 00:22 |
bigjools | so just unallocate the node so it goes Ready | 00:22 |
bigjools | then the next user to acquire it will make it reinstall | 00:22 |
bigjools | no need to delete | 00:22 |
bigjools | or commission | 00:22 |
bradm | bigjools: right. I want to re-bootstrap my juju environment, I have some lingering nodes and services that just won't die | 00:24 |
bigjools | bradm: when you destroy env, juju should release the nodes and maas should turn them off. | 00:28 |
bigjools | as in power down | 00:28 |
bigjools | next time the node is allocated it will be powered up and reinstalled | 00:28 |
bradm | I'm not sure we have ipmi setup right here | 00:33 |
bigjools | sounds like not | 00:40 |
bigjools | the key is that maas must be able to boot stuff on demand | 00:40 |
bradm | bigjools: yeah, its not a huge issue for us right now, once we have the deployment down pat we're not going to be moving things around too much | 00:51 |
bigjools | bradm: well, still best to get IPMI right :) | 00:51 |
bigjools | you never know when you will depend on it | 00:51 |
bradm | bigjools: yeah, its just not a huge priority right now | 00:51 |
bradm | bigjools: these are all HP Proliants, it shouldn't be too hard to figure it out | 00:52 |
bigjools | bradm: indeed | 00:52 |
roaksoax | bigjools: howdy!! there's no way to accept all cluster controllers from the cli? | 00:56 |
bigjools | roaksoax: uhm | 00:57 |
roaksoax | bigjools: seems not for wshat I';ve been looking for | 00:57 |
bigjools | roaksoax: there's a mass-accept but you have to specify all the UUIDs | 00:58 |
roaksoax | bigjools: yeah that's what I see | 00:58 |
roaksoax | bigjools: alright, no worries, I'll work on that assumption | 00:58 |
bigjools | roaksoax: you'd need a script to iterate over what list() returns | 00:58 |
bigjools | it's a bit of a dangerous method to blindly accept anything pending | 00:59 |
bigjools | you might not accept what you think, it's racy | 00:59 |
roaksoax | bigjools: yeah, but this is for the CHARM | 00:59 |
bigjools | charming! | 01:00 |
roaksoax | bigjools: so basically, all the clusters that are being registered to maas will need to be accepted | 01:00 |
roaksoax | bigjools: oh and to get the API key? the only way is to get it from the DB right? | 01:00 |
bigjools | roaksoax: there's a maas command | 01:00 |
roaksoax | bigjools: but you don't need to be logged in? | 01:01 |
bigjools | no | 01:01 |
roaksoax | ah cool | 01:01 |
bigjools | it's a maas command :) | 01:01 |
bigjools | roaksoax: "maas apikey" | 01:01 |
roaksoax | bigjools: awesome! cool! thanks | 01:01 |
bigjools | welcome | 01:01 |
bigjools | roaksoax: we are going to need a project called Crochet packaging. https://pypi.python.org/pypi/crochet/1.0.0 Are you the man for this or should I poke someone else? | 01:22 |
roaksoax | bigjools: yup , but allenap pinged me earlier about him wanting to package and was looking for poinnters | 01:25 |
bigjools | roaksoax: ah ok hope you pointed him appropriately then :) | 01:26 |
roaksoax | bigjools: yup! He did good and gave him some pointers, so should be packaged by tomorrow | 01:26 |
bigjools | \o/ | 01:26 |
=== mwhudson is now known as zz_mwhudson | ||
=== zz_mwhudson is now known as mwhudson | ||
=== CyberJacob|Away is now known as CyberJacob | ||
=== mwhudson is now known as zz_mwhudson | ||
bigjools | jtv, rvba: how cool is YouCompleteMe! http://i.imgur.com/aj88n4d.png | 10:24 |
jtv | Yup. I'm used to that already. | 10:26 |
melmoth | hola ! Once a node has been enlisted in maas. Is there a way to have maas alerted of a change in hardware ? (like adding more ram or cpu ?) | 10:28 |
rvba | bigjools: indeed, pretty cool. Out-of-the-box in sublimetext http://people.canonical.com/~rvb/yeah.png | 10:28 |
melmoth | or am i suppose to remove the node from the cluster and re enlist again ? | 10:28 |
bigjools | melmoth: re-commission | 10:28 |
bigjools | no other way right now | 10:28 |
rvba | melmoth: you need to take it out of rotation and re-commission | 10:28 |
melmoth | ok. thanks | 10:28 |
bigjools | rvba: does you SublimeText also show you docstrings from funcs you're about to call? | 10:29 |
bigjools | your* | 10:29 |
rvba | bigjools: no, it doesn't. YouCompleteMe! wins. | 10:30 |
bigjools | \o/ | 10:30 |
tomixxx | hi, it seems i have some DNS resolve problems: when i enter ssh cloud1.master it says: "ssh: Could not resolve hostname cloud1.master: Name or service not known" | 13:38 |
tomixxx | i have added two nodes to he maas-server, one is called "cloud1.master" and the other "cloud2.master" | 13:39 |
tomixxx | however, when i enter the nodes ip, it seems to work: "ssh ubuntu@10.0.0.100 | 13:41 |
tomixxx | hi, i have no wake-on-lan. however, when i manually power on a node, the maas-server will select the node if it needs it? | 15:14 |
=== CyberJacob is now known as CyberJacob|Away | ||
tomixxx | hi, can u help me plz | 15:59 |
tomixxx | when i call juju status i have some problem i guess | 16:00 |
tomixxx | can someone have a look on it please: http://pastebin.ubuntu.com/6844443 | 16:00 |
tomixxx | first of all: i have only 2 physical nodes in my cloude, why are there listed other nodes= | 16:01 |
tomixxx | ? | 16:01 |
tomixxx | and then, multiple juju services are pending i guess because they try to run each one on a different machine?? | 16:01 |
axisys | I am looking for a maas/juju/hadoop setup in google.. any suggestions? | 17:00 |
axisys | goal is to using juju intsall hadoop on multiple systems | 17:00 |
axisys | so first one will be master and then rest will be nodes | 17:01 |
axisys | ok installing maas on a ubuntu server | 17:47 |
axisys | and now when I try to import image it fails | 17:47 |
axisys | $ maas-cli maas node-groups import-boot-images | 17:47 |
axisys | usage: /usr/lib/python2.7/dist-packages/maascli/__main__.py [-h] COMMAND ... | 17:47 |
axisys | /usr/lib/python2.7/dist-packages/maascli/__main__.py: error: argument COMMAND: invalid choice: u'maas' (choose from u'list', u'login', u'logout', u'refresh') | 17:47 |
axisys | substitute in a different profile name for ‘maas’ if you have called yours something else .. | 17:48 |
axisys | how do I find out the profile name? | 17:48 |
axisys | ok I do have a profile now.. how do I import boot images? | 18:00 |
axisys | $ maas-cli maas node-groups import-boot-images <== not valid | 18:00 |
axisys | import-boot-images is an invalid argument to node-groups | 18:01 |
axisys | node-groups: error: argument COMMAND: invalid choice: u'import-boot-images' (choose from 'register', 'list', 'refresh-workers', 'accept', 'reject') | 18:01 |
=== CyberJacob|Away is now known as CyberJacob | ||
axisys | sudo maas-import-pxe-files seems to be working.. | 19:24 |
axisys | http://maas.ubuntu.com/docs/install.html#import-the-boot-images does not seem to align with what I am experiencing | 19:24 |
=== zz_mwhudson is now known as mwhudson | ||
axisys | looks like that is a known bug from maas 1.2 still exists on 1.3.1 | 19:45 |
axisys | $ sudo maas --version | 19:45 |
axisys | 1.3.1 | 19:45 |
=== CyberJacob is now known as CyberJacob|Away |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!