=== freeflying_away is now known as freeflying | ||
=== _bjf is now known as bjf | ||
=== julianwa_ is now known as julian | ||
=== julian is now known as juliawa | ||
=== juliawa is now known as julianwa | ||
=== CyberJacob|Away is now known as CyberJacob | ||
nesusvet | hello everyone. I have a little problem with the proxy. My local server downloads packages trough a remote proxy server, I set up the maas to work with this proxy and it looks like everything works well. But it seem JuJu tries to get information about servers through the proxy also. Could please help me to disable the requests to the proxy server please. | 07:18 |
---|---|---|
nesusvet | hello everyone. I have a little problem with the proxy. My local server downloads packages trough a remote proxy server, I set up the maas to work with this proxy and it looks like everything works well. But it seem JuJu tries to get information about servers through the proxy also. Could please help me to disable the requests to the proxy server please. | 08:42 |
allenap | nesusvet: Is that the Juju client, or Juju running on the bootstrap node, or another machine in your environment? | 08:47 |
nesusvet | Hi allenap, This is "Juju running on the bootstrap" | 08:47 |
allenap | nesusvet: Do you have a log of the "JuJu tries to get information about servers through the proxy also" bit? | 08:48 |
nesusvet | I got the 401 error with mention about my proxy host | 08:50 |
allenap | nesusvet: I really need a log to understand exactly where that error is being encountered. | 08:51 |
nesusvet | Ok, http://pastebin.com/mgVawVG0 - without proxy | 08:53 |
nesusvet | http://pastebin.com/NrS3u3wA - with proxy | 08:54 |
rvba | nesusvet: why are you using '--upload-tools" (which you should do) only when you define the proxy? I.e. could you try: juju -v bootstrap --upload-tools -e maas | 08:56 |
nesusvet | Doing. | 08:58 |
allenap | nesusvet: Right, to summarise: without proxy Juju can talk to MAAS but not to the outside world; with proxy Juju can talk to the outside world, but not MAAS. | 08:59 |
nesusvet | http://pastebin.com/kXykMF53 - here is the result | 08:59 |
allenap | nesusvet: Try that again without http_proxy. | 09:00 |
allenap | As in, unset/blank http_proxy. | 09:00 |
nesusvet | http://pastebin.com/gy24D7RD | 09:01 |
allenap | nesusvet: juju destroy-environment -e maas, then try again. | 09:01 |
nesusvet | http://pastebin.com/nsGfiVsz | 09:02 |
racedo | ping rvba | 09:02 |
rvba | racedo: hi | 09:03 |
racedo | hey rvba | 09:03 |
allenap | nesusvet: I'm not sure if we should worry about the warning. Otherwise that looks like success. | 09:03 |
nesusvet | allenap, http://pastebin.com/fMKNrhtt on this step I alway get the "time out" | 09:04 |
racedo | rvba: we have split the maas-region-controller and the maas-cluster-controller into two servers | 09:04 |
nesusvet | I mean server do nothing | 09:04 |
racedo | rvba: now we want to install multiple cluster controllers, like 7 or 8 | 09:04 |
racedo | rvba: so the region controller can communicate to all of them through the same network, right? | 09:05 |
allenap | nesusvet: Give it 5-10 minutes and try again; juju bootstrap returns before the bootstrap node is actually up and running (yeah, weird, I know). | 09:05 |
rvba | racedo: yes | 09:05 |
racedo | rvba: because every cluster controller will be managing different networks | 09:05 |
racedo | rvba: we will have 7 different networks (1 per cluster controller) | 09:05 |
rvba | racedo: right, they all need to have there own network, which they control, plus a way to contact the region | 09:05 |
racedo | rvba: exactly perfect! | 09:06 |
racedo | rvba: thanks :) | 09:06 |
racedo | rvba: so the region controller doesn't provide anything other than API, Web UI and database, right? | 09:06 |
rvba | racedo: that plus DNS | 09:07 |
racedo | rvba: oh ok | 09:07 |
racedo | rvba: but that doesn't affect the communication as DNS doesn't need to be in the same network | 09:07 |
rvba | racedo: right | 09:07 |
nesusvet | allenap, waiting... | 09:07 |
racedo | rvba: thanks again | 09:07 |
rvba | np | 09:08 |
nesusvet | allenap, http://pastebin.com/yc1H9KzY | 09:14 |
nesusvet | the result | 09:14 |
nesusvet | ohh, I forgot to add "-e maas" | 09:15 |
nesusvet | to the same step with a new option | 09:15 |
allenap | nesusvet: Can you ssh into the host - ubuntu@GPT-bart-FE - then look at /var/log/cloud-init-output.log. | 09:30 |
allenap | ? | 09:30 |
nesusvet | yes | 09:31 |
nesusvet | can't reach the host (permissions denied) | 09:32 |
allenap | nesusvet: Can you check in /var/log/maas/rsyslog on the region controller to see if there's anything. Also look at /var/log/maas/pserv.log and the Apache logs for anything suspicious. | 09:47 |
nesusvet | I have restarted the node | 09:47 |
allenap | nesusvet: Have you set the http_proxy in MAAS too? | 09:48 |
nesusvet | yes | 09:48 |
nesusvet | 2013-10-01 03:34:16 DEBUG juju.state open.go:88 connection failed, will retry: dial tcp 127.0.0.1:37017: connection refused | 09:49 |
nesusvet | here is the message on the node | 09:49 |
allenap | nesusvet: Okay, I'm wondering if that's the problem. Your proxy can't see internal hosts, but MAAS doesn't know which hosts it should use the proxy for, and which it shouldn't. | 09:49 |
nesusvet | yes, it's a true | 09:49 |
nesusvet | so for this reason I need to restrict the actions between node and proxy | 09:50 |
nesusvet | I mean local network and proxy | 09:50 |
nesusvet | for this reason I want to disable proxy | 09:53 |
nesusvet | on the clients and on the bootstrap machine | 09:53 |
=== freeflying is now known as freeflying_away | ||
allenap | nesusvet: I agree that's the best thing to try next. MAAS does install a proxy on the region controller. You could configure by hand that to use your existing proxy as an upstream proxy, and remove all proxy settings from MAAS. | 10:05 |
=== freeflying_away is now known as freeflying | ||
stokachu | what version of maas was http://maas.ubuntu.com/docs/cluster-configuration.html written for? ive got maas 1.2 on ubuntu precise and the gui doesn't have any of these options | 13:22 |
smoser | rvba, around ? | 13:31 |
smoser | where do we stand on maas and saucy right now | 13:32 |
rvba | smoser: Hi. Well, same as yesterday, the old script errors when trying to download the Saucy images. The new script is still not fully integrated but jtv is working on it. | 13:34 |
smoser | "old script" | 13:35 |
smoser | there are 2 scripts that import things | 13:35 |
stokachu | actually what version of maas was http://maas.ubuntu.com/docs/ written for | 13:35 |
smoser | import-ephemeral and import-pxe-files | 13:35 |
rvba | In this case I meant the import-ephemeral script | 13:35 |
rvba | Besides, the new script is only a replacement for maas-import-ephemerals, I wonder if tych0 is planning to port maas-import-pxe-images. | 13:35 |
rvba | tych0: ? | 13:35 |
smoser | well, thats an interesting question.. | 13:35 |
tych0 | not atm, busy with cloud installer stuff | 13:36 |
smoser | because we need server side data if we want that. | 13:36 |
smoser | i *do* want that | 13:36 |
smoser | but .... | 13:36 |
smoser | that probably woudl have to happen "right now". | 13:36 |
smoser | rvba, i really, really want to have something uploaded to ubuntu end of day here that at least doesn't fail to start. | 13:36 |
smoser | so is import-ephemeral the only blocker there ? | 13:36 |
smoser | wait. | 13:37 |
rvba | smoser: the only solution is to remove saucy from the list of series to consider. | 13:37 |
smoser | what is that a solution to ? | 13:37 |
smoser | both import-ephemeral and import-pxe-files ? | 13:37 |
smoser | we can have an ephemeral image later today. | 13:37 |
smoser | can/will | 13:38 |
tych0 | FWIW, i just landed https://code.launchpad.net/~tycho-s/maas/port-ephemerals-configs/+merge/187931 | 13:38 |
smoser | but if import-pxe-files is busted, we need to fix that. | 13:38 |
rvba | A solution to have a package which does not blow up. | 13:38 |
tych0 | which i thought was the only holdup to switching to .py | 13:38 |
rvba | tych0: well, that plus hooking up the new script in MAAS; plus actual testing :) | 13:38 |
tych0 | rvba: right, i meant the only hold up to hooking up the new script | 13:39 |
tych0 | and testing, obviously :-) | 13:39 |
smoser | oh we dont actually test anythin ghere. | 13:39 |
smoser | we just type make test | 13:39 |
* smoser was rude. sorry. | 13:40 | |
smoser | rvba, is import-pxe-files broken still ? | 13:40 |
smoser | in the current version? | 13:40 |
smoser | what is that bug ? | 13:40 |
rvba | Bug 1232174 | 13:41 |
ubot5 | bug 1232174 in MAAS "maas-import-pxe-files fails because Saucy's images information is not available." [Critical,Fix committed] https://launchpad.net/bugs/1232174 | 13:41 |
rvba | No idea why the lander marked it as fixed. | 13:41 |
stokachu | the man page for maas-cli doesn't mention anything about " maas-cli maas node-groups import-boot-images" as stated in the online docs so im not sure what version of maas these docs were written for | 13:42 |
rvba | stokachu: these docs are for the version about to be released in saucy. | 13:42 |
smoser | stokachu, i'm really sorry. you're runing 12.04 i suspect. | 13:43 |
stokachu | yea im on precise with maas 1.2 | 13:43 |
smoser | stokachu, we're working on getting you a sane solution for 12.04. | 13:43 |
stokachu | ah ok | 13:43 |
smoser | the plan is "cloud-tools" pocket, which is part of the cloud-archive. | 13:43 |
smoser | that pocket will have juju and maas from 13.10 (and dependencies) running on 12.04 | 13:43 |
stokachu | ah ok, are those bits available now? or not until saucy goes gold? | 13:44 |
smoser | stokachu, well, they're available... but in a non-yet-working state :) | 13:44 |
smoser | i was hoping by end of today that we'd have a at-least-working state. | 13:44 |
stokachu | smoser: gotcha, mind if i pm you real quick? | 13:45 |
smoser | thats fine. | 13:46 |
roaksoax | o/ | 14:01 |
roaksoax | rvba: so how are we looking? everything ready to be released? | 14:01 |
rvba | roaksoax: Hi. | 14:03 |
roaksoax | rvba: o/ | 14:04 |
roaksoax | :) | 14:04 |
rvba | roaksoax: I'm relying on smoser's judgment here. We're not in a position to use tych0's replacement script yet. And the old import scripts error because of 1232174. The consequence is not dramatic because the script is written in such a way that all the other images are downloaded first. But I would be more comfortable removing "Saucy" from the list of images to download. | 14:05 |
smoser | bug 1232174 | 14:06 |
ubot5 | bug 1232174 in MAAS "maas-import-pxe-files fails because Saucy's images information is not available." [Critical,Triaged] https://launchpad.net/bugs/1232174 | 14:06 |
smoser | rvba, ok... i'll look at fixing that script for saucy. | 14:07 |
smoser | i think actually the issue is unfortunately deeper than you think. | 14:07 |
rvba | oh? | 14:07 |
roaksoax | rvba: but we can simply disable the import of saucy for the time being... can tych0's replacement make use of saucy' | 14:08 |
roaksoax | rvba: but we can simply disable the import of saucy for the time being... can tych0's replacement make use of saucy's devel images? | 14:08 |
roaksoax | smoser: devel images are already up right? | 14:08 |
rvba | roaksoax: I'd be in favor of reverting http://bazaar.launchpad.net/~maas-maintainers/maas/trunk/revision/1641. | 14:09 |
smoser | rvba, revert the config file change | 14:11 |
smoser | but not the import_pxe_files | 14:11 |
smoser | and just ship thtat for today | 14:11 |
smoser | thats my suggestion | 14:11 |
smoser | honestly... i *REALLY* would like to see enum.py use python-distro-info | 14:12 |
* rvba nods | 14:12 | |
smoser | whifch would take rvba probably 15 minutes to do, and then we'd never have this non-sense about maas refusing to deploy something because it doesn't hknow a string. | 14:12 |
roaksoax | rvba: yeah only revert RELEASES="precise quantal raring saucy" | 14:13 |
rvba | Okay. | 14:13 |
rvba | Next time, please test a change like this before (or right after) landing it. | 14:14 |
rvba | roaksoax: https://code.launchpad.net/~rvb/maas/fix-m-i-p-f/+merge/188610 | 14:16 |
roaksoax | rvba: done | 14:17 |
rvba | Ta. | 14:18 |
rvba | roaksoax: the change is merged. | 14:25 |
roaksoax | rvba: thanks! releasing now then | 14:25 |
smoser | rvba, http://paste.ubuntu.com/6179821/ | 14:40 |
smoser | that pretty much creates the DISTRO_SERIES and DISTRO_SERIES_CHOICES that are used in enum.py | 14:41 |
smoser | it'd be nice if distro_info gave you that dict that i have to make up. or at least something like it. | 14:42 |
rvba | Indeed. | 14:42 |
* rvba filed bug 1233713. | 14:43 | |
ubot5 | bug 1233713 in MAAS "The list of supported distro series is hardcoded in src/maasserver/enum.py" [High,Triaged] https://launchpad.net/bugs/1233713 | 14:43 |
smoser | http://paste.ubuntu.com/6179846/ | 14:47 |
smoser | nothing is simple though. | 14:48 |
mgz | juju-core uses distro-info! | 14:55 |
roaksoax | so we used to have that before I remember but the changes got reverted | 14:57 |
mgz | (we chet a little, have some hardcoded values, then repopulate if something is found, using the ubuntu.csv file) | 14:57 |
roaksoax | rvba: ok so we are good to replace maas-import-ephemerals to the py version? | 15:04 |
=== matsubara is now known as matsubara-lunch | ||
rvba | roaksoax: well, no; we haven't tested it yet. Right tych0? | 15:08 |
roaksoax | rvba: ok | 15:08 |
tych0 | well | 15:09 |
tych0 | i ran a few tests locally | 15:09 |
tych0 | i don't think anyone else has tried it, though | 15:09 |
tych0 | which would be good to do | 15:09 |
rvba | Well, I just tried it. | 15:09 |
tych0 | did it work? :-) | 15:10 |
rvba | It downloads stuff, but now I can't get my nodes to boot. | 15:10 |
tych0 | hmm | 15:10 |
=== kentb-out is now known as kentb | ||
rvba | iscsi error | 15:10 |
rvba | http://people.canonical.com/~rvb/iscsi-error.jpg | 15:11 |
tych0 | arg | 15:11 |
tych0 | i've seen that. | 15:11 |
tych0 | hmm | 15:12 |
=== freeflying is now known as freeflying_away | ||
=== freeflying_away is now known as freeflying | ||
smoser | distro_info much harder to work with than i'd like. but this gets a sane array of everything. http://paste.ubuntu.com/6180076/ | 15:45 |
byro | hi | 15:59 |
byro | need some help...any volunteer? | 15:59 |
byro | ok, I will try to post my question..... | 16:02 |
byro | I've setup two VMs in virtualbox | 16:02 |
byro | one runs the maas region/cluster controller. let's call this one "master" | 16:03 |
byro | the master vm has three interfaces: | 16:03 |
byro | lost connection | 16:04 |
byro | trying again: | 16:04 |
byro | two VMs: "master", and "node" | 16:04 |
=== matsubara-lunch is now known as matsubara | ||
byro | master has following interfaces: eth0: Internal Network - eth1: NAT - eth2: host | 16:05 |
byro | node has only eth0: on Internal network | 16:05 |
byro | I've maas region and cluster controller running on master | 16:05 |
byro | master IP addresses are: eth0:192.168.0.11, eth1=10.0.3.15, eth2=192.168.56.102 | 16:06 |
byro | the problem is: | 16:07 |
byro | node does pxeboot, but it receives a preseed with cloud-config-url=http://10.0.3.15/MAAS/metadata/latest/enlist-preseed/?op=get_enlist_preseed log_host=10.0.3.15..... | 16:08 |
byro | I see the DNS address to 10.0.3.15... | 16:09 |
byro | but node doesn't have an interface on that net | 16:09 |
byro | so, what do I do on master to make the DNS show up as 192.168.0.11 instead of 10.0.3.15? | 16:09 |
AskUbuntu | virtualbox, MAAS: wrong seeding? | http://askubuntu.com/q/352363 | 17:06 |
smoser | rvba, roaksoax we shoudl have a released ephemeral for saucy now | 17:33 |
Byro | anybody willing to take a stub at http://askubuntu.com/questions/352363/virtualbox-maas-wrong-seeding ? | 17:34 |
Byro | test | 17:37 |
roaksoax | smoser: do you need simplestreams seeded? | 17:38 |
roaksoax | smoser: i was thinking of adding a dependency to simplestreams to maas (before we oficially move to the new import epehemrals script)_ | 17:39 |
ppetraki | maas issue[tcpdump]: http://paste.ubuntu.com/6180595/ , I can bootstrap and destroy an env, but juju status always reports the env is not found | 17:59 |
smoser | roaksoax, yes. please add that dependency! | 18:15 |
smoser | infinity bothers me daily to have it seeded | 18:15 |
roaksoax | smoser: done! I just need to test and will upload to archive | 18:15 |
stokachu | anyone run into issues when installing maas from cloud-tools and postgres not starting | 19:20 |
roaksoax | smoser: great! commissioning seems to be broken! | 19:23 |
roaksoax | oh nevermind | 19:23 |
roaksoax | it just took forever | 19:23 |
roaksoax | smoser: ok so I uploaded it | 19:28 |
Byro | Hi | 19:30 |
smoser | roaksoax, you uploaded maas to saucy ? | 19:34 |
roaksoax | smoser: yes | 19:35 |
smoser | and import-ephemerals and import-pxe-files works out of the box ? | 19:35 |
smoser | did you fix import-pxe-files ? or did it get fixed due to removal of saucy | 19:35 |
Byro | Hi... I see a node booting PXE trying to access 169.254.169.254. Where is that coming from, and what should it actually be? | 19:36 |
Byro | l | 19:36 |
smoser | if its trying to get there, then it didn't find the maas datasource. | 19:36 |
smoser | how did you boot it? | 19:36 |
Byro | where is that supposed to be configured? | 19:36 |
Byro | I have a server VM with maas running... | 19:37 |
Byro | I'm booting the "node" vm in PXE | 19:37 |
Byro | it is correctly getting the IP from the dhcp server in MAAS node... | 19:37 |
Byro | until a few minutes ago I was seeing dhcp0 to be wrong, but fixed that, still that had no effect on this issue | 19:38 |
smoser | Byro, i think you need isc-dhcp-server from https://launchpad.net/~virtual-maasers/+archive/maas-updated-packages | 19:46 |
Byro | smoser: I guess my IRC timed out.... | 19:48 |
roaksoax | smoser: yes | 19:49 |
Byro | did you have some wise word for me? | 19:49 |
roaksoax | smoser: removing saucy from the list of releases in /etc/maas/import_pxe_files worked | 19:49 |
smoser | Byro, i suspect you're hitting https://bugs.launchpad.net/ubuntu/+source/maas/+bug/1069570 | 19:50 |
ubot5` | Ubuntu bug 1069570 in maas (Ubuntu Raring) "1 MAC Address, two IPs - DNS is "out of sync" with DHCP leases databases" [Undecided,Confirmed] | 19:50 |
smoser | and you can ge ta build with that fix in it at https://launchpad.net/~virtual-maasers/+archive/maas-updated-packages | 19:50 |
smoser | roaksoax, thank you. | 19:50 |
smoser | you'll also need to modify your maas dhcp.conf Byro | 19:50 |
Byro_ | the stupid client keeps timing out | 19:52 |
Byro_ | but I did see your answer.... | 19:52 |
Byro_ | I don't have a single mac though | 19:52 |
Byro_ | I have three separate eth on the server | 19:52 |
Byro_ | each one gets one IP | 19:52 |
Byro_ | do you still think that bug would affect me? | 19:52 |
byro | hi | 19:55 |
byro | my client keeps disconnecting | 19:55 |
smoser | byro, but you're surely only booting off of one of them. | 19:55 |
smoser | basically eth0 has to get an ip off of dhcp and has to be ocnnected to amas. | 19:55 |
byro | yes, I'm booting out of 192.168.0.11 | 19:55 |
byro | and correctly getting IP | 19:55 |
byro | I also was able to fix the dns0 | 19:55 |
byro | it now correctly indicates 192.168.0.11 instead of 10.0.15.3 | 19:56 |
byro | but... | 19:56 |
byro | I still get the problem | 19:56 |
smoser | byro, this is commissioning ? | 19:57 |
smoser | or enlistment? | 19:57 |
byro | uhm... | 19:58 |
byro | sorry, not entirely accustomed to jargon yet.... | 19:59 |
byro | hold on | 19:59 |
smoser | what have you done so far? | 19:59 |
byro | there are no known nodes to the controller yet | 19:59 |
smoser | yeah. | 19:59 |
smoser | so thats enlistment | 19:59 |
byro | ok | 19:59 |
smoser | the first time the node boots it shoudl pxeboot and then tell maas all of its network interfaes. | 20:00 |
smoser | then you 'accept' the node, and it will "commission" | 20:00 |
byro | right | 20:00 |
byro | ok | 20:00 |
byro | question: | 20:01 |
byro | is this ok? | 20:01 |
byro | #cloud-config datasource: MAAS: timeout : 50 max_wait : 120 # there are no default values for metadata_url or oauth credentials # If no credentials are present, non-authed attempts will be made. metadata_url: http://10.0.3.15/MAAS/metadata/enlist | 20:01 |
byro | this is the preseed | 20:01 |
byro | it also shows 10.0.3.15 | 20:01 |
byro | while the pxe-booted node can't reach that network | 20:01 |
byro | should that be 192.168.0.11 instead (address of maas server on reachable network)? | 20:02 |
smoser | yeah. | 20:03 |
byro | ok... | 20:03 |
smoser | try dpkg-reconfigure maas-provisioning-server | 20:03 |
byro | ok | 20:03 |
byro | dpkg is not installed | 20:04 |
byro | will install | 20:04 |
byro | uhm... | 20:04 |
byro | no info available to install it... | 20:04 |
byro | actually, sorry | 20:06 |
byro | maas-privisioningt server is not installed | 20:06 |
byro | maas-provisioning-server | 20:06 |
byro | ? | 20:09 |
byro | am I still in the channel? | 20:11 |
Byro | hello | 20:22 |
Byro | sorry if somebody already replied, I keep getting kicked out of the channel due to timeouts | 20:22 |
Byro | my preseed have the wrong ip address in it | 20:25 |
Byro | how do I change that? | 20:25 |
Byro | how can I change the metadata_url IP address in the preseed? | 20:32 |
Byro | anbody here? | 20:37 |
stokachu | roaksoax: whats the best way to see whats happening with a node before enlistment? i can see where the maas server offered its dhcp information but not sure where to check status of the node | 21:49 |
=== freeflying is now known as freeflying_away | ||
=== CyberJacob is now known as CyberJacob|Away | ||
tych0 | ok, i know i'm forgetting something simple; my pxe nodes can't boot because they can't find pxelinux.0 | 23:55 |
tych0 | it gets an IP, and tgtd is running on the maas machine | 23:56 |
tych0 | but /var/log/maas/pserv.log shows no signs of activity when i try to pxe boot | 23:56 |
tych0 | what am i forgetting to start? :-) | 23:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!