mup | Bug #1630817 changed: [ERROR] Unable to find a copy of bootppc64.bin in the SimpleStream or a previously downloaded copy. <MAAS:New> <https://launchpad.net/bugs/1630817> | 02:58 |
---|---|---|
Sujeet_ | Hi Kiko | 03:30 |
Sujeet_ | How to handle errors in the commissioning script? | 03:30 |
=== frankban|afk is now known as frankban | ||
mup | Bug #1631358 opened: [2.1] Incorrect logging message - showing SERVICE_STATE.ON <MAAS:New> <https://launchpad.net/bugs/1631358> | 13:04 |
mup | Bug #1631358 changed: [2.1] Incorrect logging message - showing SERVICE_STATE.ON <MAAS:New> <https://launchpad.net/bugs/1631358> | 13:13 |
mup | Bug #1631358 opened: [2.1] Incorrect logging message - showing SERVICE_STATE.ON <MAAS:New> <https://launchpad.net/bugs/1631358> | 13:25 |
mup | Bug #1630745 changed: [2.1b1] Nodes faild to deploy due to 503 error in archive <cdo-qa> <cdo-qa-blocker> <MAAS:Invalid> <https://launchpad.net/bugs/1630745> | 14:04 |
mup | Bug #1630745 opened: [2.1b1] Nodes faild to deploy due to 503 error in archive <cdo-qa> <cdo-qa-blocker> <MAAS:Invalid> <https://launchpad.net/bugs/1630745> | 14:22 |
mup | Bug #1630745 changed: [2.1b1] Nodes faild to deploy due to 503 error in archive <cdo-qa> <cdo-qa-blocker> <MAAS:Invalid> <https://launchpad.net/bugs/1630745> | 14:25 |
* baldpope sigh | 14:39 | |
baldpope | figured out my networking issue from inside lan - forgot to turn on nat at perimeter firewall.. | 14:39 |
fbarilla | hi, can't restart maas-dhcpd | 14:51 |
fbarilla | root@hnode1-8:~# service maas-dhcpd restart Failed to restart maas-dhcpd.service: Unit maas-rackd.service is masked. | 14:51 |
roaksoax | fbarilla: restart maas-rackd | 14:57 |
roaksoax | fbarilla: that should restart maas-dhcpd | 14:57 |
mup | Bug #1631403 opened: [2.1b2] Service 'ntp' is SERVICE_STATE.ON but not in the expected state of 'running', its current state is 'exited'. <oil> <MAAS:New> <https://launchpad.net/bugs/1631403> | 14:57 |
fbarilla | root@hnode1-8:~# service maas-rackd restart Failed to restart maas-rackd.service: Unit maas-rackd.service is masked. | 14:57 |
mup | Bug #1631403 changed: [2.1b2] Service 'ntp' is SERVICE_STATE.ON but not in the expected state of 'running', its current state is 'exited'. <oil> <MAAS:New> <https://launchpad.net/bugs/1631403> | 15:03 |
mup | Bug #1631403 opened: [2.1b2] Service 'ntp' is SERVICE_STATE.ON but not in the expected state of 'running', its current state is 'exited'. <oil> <MAAS:New> <https://launchpad.net/bugs/1631403> | 15:06 |
mup | Bug #1631420 opened: [2.1 UI] Images page "Queued for download" is confusing when selections are not saved <MAAS:New> <https://launchpad.net/bugs/1631420> | 16:15 |
mup | Bug #1631421 opened: [2.1b2] failure to deploy all systems that lasted until reboot of maas container <oil> <MAAS:New> <https://launchpad.net/bugs/1631421> | 16:15 |
braven_ | hello | 16:33 |
braven_ | Is anyone on line? | 16:34 |
brendand | yes | 16:36 |
brendand | braven_, ^ | 16:36 |
braven_ | I am having an issue. I keep getting a internal error | 16:38 |
braven_ | when try to save a interface to dhcp | 16:38 |
braven_ | The cluster controller will not accept the changes | 16:42 |
=== frankban is now known as frankban|afk | ||
braven_ | Code<UNKNOWN>: Unknown Error | 16:56 |
brendand | braven_, was it you was asking about the secret yesterday? | 17:01 |
braven_ | no | 17:06 |
braven_ | We have the cluster Controller connected | 17:07 |
braven_ | it was working. We tried adjust the dhcp range. It would not save | 17:09 |
braven_ | when we try to save the changes to the interface we get "Internal server error." | 17:10 |
sean5384 | hi there | 17:10 |
braven_ | hello | 17:10 |
sean5384 | i hope someone can shed some light for me, i am trying to install MAAS and it says a few sentences down about the clusters tab, where the heck is it at? | 17:11 |
braven_ | what version of MAAS | 17:12 |
sean5384 | i think it is 2.0, let me check | 17:12 |
sean5384 | if i'm running ubuntu 16.04 LTS, it should be 2.0 right? | 17:13 |
braven_ | yeap. | 17:13 |
sean5384 | ok then, thats what i have running then. | 17:13 |
sean5384 | is the clusters tab, no more then or something? | 17:14 |
braven_ | I am using the old version because my hardware. But do remember read they change it | 17:15 |
sean5384 | i also can't seem to get my private network functioning correcty within ubuntu itself either for some reason, would that have something to do with it maybe? | 17:15 |
braven_ | They something new called Fabric | 17:15 |
sean5384 | ok, i have seen that. | 17:16 |
brendand | sean5384, do you have a link to the docs you are reading? | 17:16 |
braven_ | Have used the older version of MAAS | 17:16 |
sean5384 | i was trying to install via the documentation, but you can't follow that now i can see. | 17:16 |
brendand | braven_, which version are you using? | 17:16 |
braven_ | 1.8 | 17:17 |
brendand | ahhh | 17:17 |
brendand | on which version of ubuntu | 17:18 |
braven_ | http://maas.io/docs/en/installconfig-rack for Sean | 17:18 |
braven_ | 14.04 | 17:18 |
braven_ | the hardware I have is not supported by 16 | 17:18 |
sean5384 | i just saw the documentation on top of this site..... | 17:18 |
braven_ | lol | 17:18 |
sean5384 | this makes more sense now. | 17:19 |
brendand | braven_, 1.8 is not really supported. 1.9 is the version in Trusty | 17:19 |
sean5384 | i was getting distraught because there was no clusters tab at the top of the page anywhere. | 17:19 |
sean5384 | lol | 17:19 |
braven_ | MAAS is awesome when it works. We were able to deploy 60 nodes in about 3 hour | 17:20 |
sean5384 | wow | 17:20 |
braven_ | how do I know what version I am running? | 17:20 |
sean5384 | i'm trying to get just 5 going right now. | 17:20 |
brendand | braven_, it should say at the bottom of the web ui | 17:20 |
brendand | or you can do dpkg -l | grep maas on the maas server | 17:21 |
brendand | braven_, that largely depends on your hardware/network too, we regularly deploy 40+ nodes in about 15-20 mins | 17:21 |
brendand | haven't timed it though so it could be a bit longer | 17:22 |
braven_ | wow | 17:22 |
braven_ | I wonder when we configured wrong | 17:22 |
brendand | it will use a lot of bandwidth, is one thing | 17:23 |
braven_ | 1.7.6 | 17:23 |
braven_ | so can we upgrade without loosing all the nodes? | 17:23 |
brendand | braven_, you *should* be able to | 17:24 |
braven_ | do know why it might be give that error when I am trying to save a cluster interface | 17:27 |
brendand | could be anything really. haven't used 1.7 much | 17:29 |
brendand | braven_, but some logs would be helpful | 17:33 |
brendand | braven_, if you're willing to upgrade that would be a good first step, then if it still happens, file a bug | 17:35 |
smgoller | Hi all, if I want to control VMs on a vCenter server, what has to be on the same subnet as the rack controller, the vCenter appliance, or the individual esxi servers? | 17:39 |
batkins61_ | Nodes won't enlist, even though booted under maas. Metadata service appears unreachable. Version 1.9 on 14.04. | 19:39 |
batkins61_ | Where should I look for that? | 19:39 |
batkins61_ | Is there any help on this channel? Is there somewhere else to ask questions? | 19:54 |
=== spammy is now known as Guest28928 | ||
=== CyberJacob is now known as Guest90146 | ||
=== med_ is now known as Guest94967 | ||
=== Guest94967 is now known as medberry | ||
baldpope | with recent update to conjure, i get error after selecting which type of openstack to deploy | 20:20 |
stokachu | baldpope: which is? | 20:25 |
mup | Bug # changed: 1470202, 1473254, 1553647, 1563779, 1563799, 1563807, 1592666, 1604901 | 20:25 |
baldpope | conjure-up -> openstack for maas -> hawk (only option) -> error | 20:27 |
baldpope | unable to list models: error cannot list models: no api addresses | 20:27 |
stokachu | baldpope: sounds like a failed bootstrap | 20:29 |
stokachu | what does juju status return? | 20:29 |
baldpope | error no api address | 20:29 |
stokachu | yea your bootstrap failed | 20:29 |
baldpope | i'm seeing this on the ubuntu maas host | 20:30 |
baldpope | all nodes in ready - none currently deployed - i thought this prepped the first node? | 20:30 |
baldpope | or can I make this maas head the juju controller? | 20:31 |
stokachu | yea so you have a state where juju thinks a controller exist but maas has nothing deployed | 20:32 |
stokachu | i would juju kill-controller hawk | 20:32 |
baldpope | Unable to open API: no API addresses Unable to connect to the API server, destroying through provider | 20:32 |
stokachu | thats fine it'll just kill it manually and delete it from it's database | 20:33 |
stokachu | then you can try conjure-up again | 20:33 |
stokachu | baldpope: nah we have a seperate queue that handles all that logic so you can proceed until you see the screen showing the bootstrap output | 20:40 |
baldpope | looks like it's making progress | 20:44 |
baldpope | i see progress in the conjure window - hadn't see that before | 20:44 |
stokachu | baldpope: :D | 20:45 |
baldpope | stokachu, hm, seems like it's stalled again | 20:50 |
stokachu | baldpope: the machine is still in deploying state? | 20:50 |
baldpope | nope, deployed | 20:50 |
stokachu | what the wait screen showing? | 20:51 |
baldpope | oh wait | 20:51 |
baldpope | just finihed | 20:51 |
baldpope | wow - long time? | 20:51 |
baldpope | it's now deploying the rest | 20:51 |
stokachu | yea maas is depending on the hardware | 20:51 |
stokachu | so if the systems take a minute to boot up then it'll prolong that wait | 20:52 |
stokachu | etc | 20:52 |
baldpope | gotcha | 20:52 |
* baldpope w000t-- progress | 20:52 | |
stokachu | but honestly aws takes about the same amount of time from bootstrap to deploy | 20:52 |
baldpope | i don't mind failure as long as I learn something in the progress | 20:52 |
stokachu | how many machines do you have available in maas? | 20:53 |
stokachu | after the first one deployed? | 20:53 |
baldpope | 5 total | 20:53 |
baldpope | blade(1-5) | 20:53 |
stokachu | ok cool | 20:53 |
stokachu | all with 2nics? | 20:53 |
baldpope | the first, is where juju (I believe) completed the it's deployment | 20:53 |
baldpope | 6 technically, but 2 are plumbed | 20:53 |
stokachu | yea the first deployed machine is your controller node | 20:53 |
stokachu | ok | 20:54 |
stokachu | baldpope: and your network interfaces are coming up eth0/eth1 as well? | 20:54 |
baldpope | side note, i'm all in on openstack for the foreseeable future at work - so anything you guys need to test - i'm available | 20:54 |
stokachu | i know we fixed that yesterday | 20:54 |
baldpope | on blade1, it's shows as br-enp9s0f0 | 20:55 |
baldpope | as the active interface | 20:55 |
baldpope | and then en01, enp9s0f0, enp9s0f1, ens3f1, 2, 3, | 20:55 |
stokachu | ok neutron may fail with a 'config-changed' hook that we can fix with juju config neutron-gateway br-ext=en01 etc | 20:55 |
stokachu | and then juju resolved neutron-gateway/0 | 20:56 |
stokachu | you could probably set that now if you want | 20:56 |
stokachu | to avoid the hook error | 20:56 |
baldpope | so from the maas head, run what? | 20:57 |
stokachu | juju config neutron-gateway br-ext=<second interface name> | 20:57 |
baldpope | ok, quick question before I do that - the remaining 4 are now in deployed state | 21:02 |
baldpope | but everything is in waiting or maintenance state | 21:04 |
stokachu | yea they are all going through processing hooks etc | 21:05 |
stokachu | all the things needed to get the software up | 21:05 |
baldpope | should I still attempt to change the neutron-gateway? | 21:06 |
stokachu | yea it should queue it up and apply it | 21:06 |
baldpope | k | 21:06 |
stokachu | if the charm breaks because of performing that action it is a bug in the charm | 21:07 |
stokachu | unless the interface was entered incorrectly | 21:07 |
baldpope | and the second int name is the one on the node, correct? | 21:07 |
stokachu | though it should really let you know via status-set blocked | 21:07 |
stokachu | baldpope: yea the one on that node that is running neutron-gateway | 21:07 |
baldpope | ERROR unknown option "br-ext" | 21:09 |
baldpope | juju config neutron-gateway br-ext=enp9s0f1 | 21:10 |
baldpope | and conjure-up just errored out - neutron-gateway/0: hook failed: config-changed | 21:11 |
stokachu | sec | 21:11 |
stokachu | baldpope: ah you need to setup bridge-mappings | 21:12 |
stokachu | https://jujucharms.com/neutron-gateway/xenial/3#charm-config-bridge-mappings | 21:12 |
stokachu | so like if you have a bridge setup on that node you can do | 21:14 |
stokachu | juju config neutron-gateway bridge-mappings=mynet:br0 | 21:14 |
baldpope | stokachu, while setting up a bridge makes sense (and I'd want to get there eventually) I did not setup bridge mode (either in maas or on the switch) | 21:14 |
baldpope | i assumed the br-enp9s0f0 was stock from the existing charm? | 21:15 |
stokachu | does that bridge exist on the system? | 21:16 |
stokachu | if you juju ssh neutron-gateway/0 | 21:16 |
stokachu | more info as well: https://ask.openstack.org/en/question/94206/how-to-understand-the-bridge_mappings/ | 21:16 |
baldpope | http://pastebin.ubuntu.com/23290786/ | 21:17 |
stokachu | yea so you can assign the bridge mapping to br-enp9s0f0 | 21:17 |
stokachu | baldpope: additional info https://github.com/openstack/charm-neutron-gateway#port-configuration | 21:18 |
stokachu | you can access all those config options from conjure-up if you click on configure->advanced configuration | 21:19 |
baldpope | silly question stokachu - i feel like I've completely skipped a ton of reading - attempting to just follow the ubuntu docs on deploying open stack - is there a primer I skipped/missed? | 21:20 |
stokachu | baldpope: networking is complex and you'll want to be familiar with that | 21:20 |
=== Guest28928 is now known as spammy | ||
batkins61 | Do target hosts have to have internet access during elistment? | 21:36 |
baldpope | stokachu, not gonna lie - a lot to take in | 21:54 |
stokachu | baldpope: yea openstack is a complex beast | 21:55 |
baldpope | stokachu, for the sake of conversation i've got the 6 ports total, 2 on board, and 4 on the add-on card - the long-term idea being that I'd use the some combination to dedicate shared storage and some remain to dedicate to service networks | 21:56 |
baldpope | all that is done through juju ahead of deployment, or post deploy? | 21:57 |
baldpope | i can see where that made little sense, i'll try again... dedicate 4 ports to the shared storage, and the 2 onboard to network services | 21:59 |
baldpope | with the appropriate lag/bridge ports in place for redundancy and throughput | 22:00 |
baldpope | ok, so I did the following: juju config neutron-gateway bridge-mappings=br-enp9s0f0 | 22:10 |
baldpope | and am re-running conjure against the existing deployment | 22:10 |
baldpope | we'll see | 22:10 |
baldpope | bridge-mappings appears to be the current non-deprecated command | 22:11 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!