| sdake | rhallisey wouldn't selector just be "nova-controller'? | 00:00 |
|---|---|---|
| sdake | eg, if you launch nova selector, all of those other services are brought into action? | 00:00 |
| sdake | rather the nova-controller pod | 00:01 |
| rhallisey | maybe, that conf file is from packstack and I've been slowly trimming and adjusting it | 00:01 |
| sdake | I see | 00:01 |
| sdake | well i was talking about https://review.openstack.org/#/c/128138/1 | 00:02 |
| sdake | I don't think you need that at all | 00:02 |
| sdake | the only reason to make a service defn is to tell k8s where the port is | 00:02 |
| rhallisey | oh sorry | 00:02 |
| sdake | no apology necessary - we are all learning how this stuff works | 00:03 |
| rhallisey | ok so would I just need the controller then? | 00:03 |
| sdake | take a look at compute | 00:03 |
| sdake | nova-compute-service.yaml | 00:03 |
| sdake | nova-libvirt-service.yaml | 00:04 |
| sdake | see how they have selector the same value? | 00:04 |
| sdake | that tells the pod (please start these services) | 00:04 |
| sdake | it shows upin the gui | 00:04 |
| sdake | no harm in having them | 00:04 |
| sdake | but they should have the same selector imo :) | 00:04 |
| sdake | becausey ou want them part of the same pod | 00:04 |
| rhallisey | oh I see what you mean | 00:04 |
| rhallisey | gotcha, changing now | 00:05 |
| sdake | and add to "tools" directory plz :) | 00:05 |
| rhallisey | ya still learning what some of the kube stuff means | 00:05 |
| rhallisey | adding :) | 00:06 |
| * rhallisey has more coming! | 00:06 | |
| lifeless | I wonder if there is some way to express the deps etc just once | 00:07 |
| lifeless | across all the deploy mechanisms | 00:07 |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Minor fixes for the nova-controller Dockerfile https://review.openstack.org/128141 | 00:08 |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Add kube services https://review.openstack.org/128138 | 00:11 |
| sdake | larsks i have a .4.6 and .4.7, can ping between, but telnet to the port does not route or something | 00:14 |
| sdake | any tips on debugging? | 00:14 |
| *** jerryz has joined #tripleo | 00:15 | |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Add kube services https://review.openstack.org/128138 | 00:17 |
| *** ramishra has quit IRC | 00:19 | |
| *** noslzzp has joined #tripleo | 00:20 | |
| *** chuckC_ has quit IRC | 00:22 | |
| StevenK | | fault | {"message": "No valid host was found. " | 00:25 |
| StevenK | Grrr | 00:26 |
| *** ChuckC has quit IRC | 00:32 | |
| StevenK | Hmmm, how can I try and boot an instance like Heat is doing? | 00:34 |
| larsks | sdake: okay, I'm around for a bit now. You still here? | 00:36 |
| sdake | yup | 00:37 |
| sdake | want to login and have a look? | 00:37 |
| larsks | I am logged in on your system...looks like things are working right now. Within the glance container, I can run the /check.sh script without error, which talk to both keystone and glance. | 00:38 |
| sdake | try glance index after sourcing /tmp/keystonerc* | 00:38 |
| sdake | (on the master ) | 00:38 |
| larsks | I would not necessarily expect anything to work from the master...it's not linked into the network that provides connectivity between containers on the minions. | 00:39 |
| larsks | It just provides the kubernetes services. | 00:39 |
| sdake | i see | 00:39 |
| sdake | how would you access from outside the cluster? | 00:39 |
| sdake | what I really want to do is access on my host | 00:39 |
| larsks | well, as I said in some of my emails to the list, we haven't even attempted to solve the public ip question yet. I've been focused on just getting the basic containers working. | 00:40 |
| larsks | In theory, you could set PUBLIC_IP in your pods to the address of one of your minions and it might work. | 00:40 |
| larsks | I haven't tried that yet. | 00:40 |
| larsks | I would say give it a shot...it might work for now, although I worry that it might be a fragile solution (e.g., kubernetes may change so that doesn't work). | 00:41 |
| sdake | larsks yup it works on the local net | 00:42 |
| larsks | Awesome. | 00:42 |
| sdake | however logging into a minion is probably not how most people expect to use openstack ;-) | 00:43 |
| larsks | sdake: I think we're all aware of that. | 00:43 |
| larsks | But that's where we are right now. | 00:43 |
| sdake | yup | 00:43 |
| larsks | Incidentally, this works fine on the master: SERVICE_ENDPOINT=http://10.0.0.4:35357/v2.0 SERVICE_TOKEN=ADMINTOKEN keystone user-list | 00:43 |
| larsks | So that suggets that setting PUBLIC_URL to the ip of one of your minions *will* work right now. | 00:44 |
| sdake | you mean 10.0.0.4 | 00:44 |
| larsks | Yeah. | 00:44 |
| sdake | and then set the keystonerc to what 172.x? | 00:44 |
| larsks | sdake: well, set PUBLIC_IP to whatever address you expect to use from outside the k8s cluster. You should in theory be able to set it tot he floating ip address. | 00:45 |
| *** panda has quit IRC | 00:45 | |
| larsks | E.g., on your physical host, this works: SERVICE_ENDPOINT=http://172.24.4.236:35357/v2.0 SERVICE_TOKEN=ADMINTOKEN keystone user-list | 00:45 |
| *** panda has joined #tripleo | 00:45 | |
| larsks | So putting that address into the catalog by setting PUBLIC_IP in the pod config should work. | 00:45 |
| sdake | 172/24 is only visiible outside the cluster i believe | 00:46 |
| larsks | Correct. | 00:46 |
| larsks | That's fine. | 00:46 |
| sdake | doesn't glance tal kto keystone directly? | 00:46 |
| larsks | Let's find out! Hold on a sec, I'll just modify the service catalog and let's see what happens... | 00:47 |
| larsks | Okay, so, on your physical host (minime), sourt /tmp/keystonerc. | 00:50 |
| larsks | s/sourt/source/ | 00:50 |
| larsks | I can 'keystone token-get' and 'glance image-list' at least. | 00:50 |
| larsks | keystone admin access won't work (because the adminurl uses an internal address). | 00:51 |
| greghaynes | apparenly new jq supports built in 'deep-merge' of dictionaries | 00:51 |
| greghaynes | but its not on trusty yet :( | 00:51 |
| StevenK | greghaynes: Since trusty is out, and all ... | 00:53 |
| StevenK | 1.4.2 is in utopic, you could file a request for trusty-backports | 00:53 |
| greghaynes | yeaaa | 00:54 |
| *** ramishra has joined #tripleo | 00:54 | |
| sdake | larsks [sdake@minime ~(keystone_admin)]$ glance index | 00:55 |
| sdake | Authorization Failed: Unable to establish connection to http://174.24.4.236:5000/v2.0/tokens | 00:55 |
| greghaynes | or maybe we dont install it from package | 00:55 |
| StevenK | greghaynes: It depends how much we want the shiny, I guess | 00:56 |
| larsks | sdake: http://chunk.io/f/ef8c28c2428f46168a49332f68e57252 | 00:56 |
| larsks | sdake: Are you using /tmp/keystonerc? | 00:57 |
| greghaynes | alternatively, could get hacked into o-a-c since that already has a deep merge | 00:57 |
| sdake | larsks yup that works | 00:57 |
| greghaynes | basically, deep merge is something we do a lot of, I think this could clean up some code pretty well | 00:57 |
| larsks | sdake: yup. | 00:57 |
| sdake | must be my creds were wrong | 00:57 |
| sdake | why not use 10.0.0.z for address though in the keystone catalog? | 00:58 |
| sdake | so I guess we could have a load balancer on the front end or something | 01:01 |
| sdake | seems kind of clunky | 01:02 |
| sdake | would be nice if it just hopped networks magically ;-) | 01:02 |
| sdake | ok dinner time bbl | 01:02 |
| tchaypo | back | 01:03 |
| larsks | sdake: enjoy dinner. talk to you tomorrow. | 01:03 |
| *** weshay has quit IRC | 01:03 | |
| tchaypo | lifeless: still around? | 01:04 |
| tchaypo | I think I need to iterate over all the machines and check which MAC they PXE from. | 01:05 |
| *** ramishra has quit IRC | 01:05 | |
| tchaypo | It'd be nice if I could avoid doing it by hand | 01:05 |
| *** spzala has quit IRC | 01:10 | |
| StevenK | Grr. So add pxe_deploy_{kernel,ramdisk} and now I just get "Setting instance to ERROR state" | 01:11 |
| tchaypo | add to what? | 01:12 |
| greghaynes | reminds me of the good, old "Error in cloud" | 01:12 |
| StevenK | tchaypo: I was getting the error of The following parameters were not passed to ironic: ['pxe_deploy_ramdisk', 'pxe_deploy_kernel'] | 01:13 |
| StevenK | So ironic node-update <uuid> add driver_info/blah ... | 01:13 |
| openstackgerrit | A change was merged to stackforge/kolla: Minor fixes for the nova-controller Dockerfile https://review.openstack.org/128141 | 01:13 |
| *** ChuckC has joined #tripleo | 01:16 | |
| *** chuckC_ has joined #tripleo | 01:19 | |
| tchaypo | bahahah | 01:20 |
| StevenK | tchaypo: Hm? | 01:22 |
| tchaypo | laughing at your problems | 01:22 |
| tchaypo | they're so much funnier than mine | 01:22 |
| StevenK | Now I've set both ironic and nova to debug and I still can't work it out | 01:23 |
| *** ramishra has joined #tripleo | 01:23 | |
| openstackgerrit | A change was merged to stackforge/kolla: Nova-controller config setup https://review.openstack.org/128129 | 01:23 |
| tchaypo | it seems that "ionic node-list" doesn't do much | 01:23 |
| StevenK | Haha | 01:24 |
| StevenK | Command not found even? | 01:24 |
| tchaypo | can we make a new tool called ionic? | 01:24 |
| tchaypo | i even know what it should do | 01:24 |
| tchaypo | people have talked about having ironic nodes kept as pre-built warm-spares for faster failover | 01:24 |
| StevenK | tchaypo: It should print "Do you mean ironic?" and exit | 01:24 |
| tchaypo | ionic could be the service that keeps the pool "charged up" and ready to run | 01:24 |
| *** noslzzp has quit IRC | 01:25 | |
| tchaypo | okay, i can use ironic to power-on all the machines, which makes them all pxe-boot, which means I can collect the macs easily | 01:32 |
| tchaypo | but I don't know which machine each mac belongs to - they aren't providing a hostname | 01:32 |
| lifeless | tchaypo: power on, get mac, power off, goto next. | 01:33 |
| *** ramishra has quit IRC | 01:33 | |
| tchaypo | sigh. | 01:33 |
| tchaypo | sigh. | 01:33 |
| tchaypo | in fact | 01:33 |
| tchaypo | sigh x 80 | 01:33 |
| tchaypo | especially since I can't copy/paste | 01:33 |
| lifeless | you could automate it | 01:33 |
| lifeless | you should see the fails in syslog | 01:33 |
| lifeless | so power on; tail syslog | grep; power off. | 01:34 |
| tchaypo | yeah, that will probably get me the majority. | 01:35 |
| StevenK | Grrr. I want a machine booted, but I also need to buy lunch :-/ | 01:36 |
| *** rainya has joined #tripleo | 01:42 | |
| *** nosnos has joined #tripleo | 01:55 | |
| *** daneyon_ has quit IRC | 01:57 | |
| *** pradeep has joined #tripleo | 02:23 | |
| *** pradeep1 has joined #tripleo | 02:28 | |
| *** pradeep has quit IRC | 02:28 | |
| *** rainya has quit IRC | 02:29 | |
| *** rhallisey has quit IRC | 02:36 | |
| *** pradeep has joined #tripleo | 02:39 | |
| *** pradeep1 has quit IRC | 02:42 | |
| *** chuckC__ has joined #tripleo | 02:44 | |
| *** nijaba_ has joined #tripleo | 02:47 | |
| *** wendar_ has joined #tripleo | 02:49 | |
| *** sdake has quit IRC | 02:50 | |
| *** pcrews has quit IRC | 02:50 | |
| *** SpamapS_ has joined #tripleo | 02:50 | |
| *** pradeep1 has joined #tripleo | 02:51 | |
| *** pradeep has quit IRC | 02:51 | |
| *** spiffxp- has joined #tripleo | 02:51 | |
| *** sirushti_ has joined #tripleo | 02:52 | |
| *** morgan has joined #tripleo | 02:52 | |
| *** sdake has joined #tripleo | 02:52 | |
| *** d0ugal_ has joined #tripleo | 02:52 | |
| *** dtantsur has joined #tripleo | 02:53 | |
| *** chuckC_ has quit IRC | 02:54 | |
| *** Dafna has quit IRC | 02:54 | |
| *** morganfainberg has quit IRC | 02:54 | |
| *** sirushti has quit IRC | 02:54 | |
| *** jroll has quit IRC | 02:54 | |
| *** grauwulf has quit IRC | 02:54 | |
| *** jrist has quit IRC | 02:54 | |
| *** greghaynes has quit IRC | 02:54 | |
| *** wendar has quit IRC | 02:54 | |
| *** stevebaker has quit IRC | 02:54 | |
| *** adam_g has quit IRC | 02:54 | |
| *** russell_h has quit IRC | 02:54 | |
| *** nijaba has quit IRC | 02:54 | |
| *** dshulyak has quit IRC | 02:54 | |
| *** d0ugal has quit IRC | 02:54 | |
| *** spiffxp_ has quit IRC | 02:54 | |
| *** dtantsur|afk has quit IRC | 02:54 | |
| *** SpamapS has quit IRC | 02:54 | |
| *** d0ugal_ is now known as d0ugal | 02:54 | |
| *** sirushti_ is now known as sirushti | 02:54 | |
| *** morgan is now known as morganfainberg | 02:54 | |
| *** chuckC__ is now known as chuckC_ | 02:54 | |
| *** d0ugal is now known as Guest85887 | 02:54 | |
| *** stevebaker has joined #tripleo | 02:57 | |
| *** adam_g has joined #tripleo | 02:57 | |
| *** adam_g has quit IRC | 02:58 | |
| *** adam_g has joined #tripleo | 02:58 | |
| *** Dafna has joined #tripleo | 03:00 | |
| *** dshulyak has joined #tripleo | 03:00 | |
| *** untriaged-bot has joined #tripleo | 03:00 | |
| untriaged-bot | Untriaged bugs so far: | 03:00 |
| untriaged-bot | https://bugs.launchpad.net/tripleo/+bug/1380698 | 03:00 |
| untriaged-bot | https://bugs.launchpad.net/tripleo/+bug/1380608 | 03:00 |
| uvirtbot | Launchpad bug 1380698 in tripleo "swift-proxy doesn't install keystonemiddleware" [Undecided,New] | 03:00 |
| uvirtbot | Launchpad bug 1380608 in tripleo "Convert block-storage-nfs.yaml to StructuredDeployment" [Undecided,Confirmed] | 03:00 |
| *** untriaged-bot has quit IRC | 03:00 | |
| *** yamahata has joined #tripleo | 03:01 | |
| *** greghaynes has joined #tripleo | 03:01 | |
| *** jrist has joined #tripleo | 03:02 | |
| *** rainya has joined #tripleo | 03:04 | |
| *** rainya has quit IRC | 03:04 | |
| *** ramishra has joined #tripleo | 03:05 | |
| *** grauwulf has joined #tripleo | 03:07 | |
| *** noslzzp has joined #tripleo | 03:15 | |
| *** russell_h has joined #tripleo | 03:21 | |
| *** jroll has joined #tripleo | 03:21 | |
| *** Azher has joined #tripleo | 03:22 | |
| Azher | hi there, i have a question related to diskimage-builder | 03:22 |
| tchaypo | hi azher | 03:22 |
| Azher | hello tchaypo! | 03:23 |
| Azher | I was trying to build a disk image on my Ubuntu 14.04 server | 03:23 |
| Azher | and i got the below error | 03:24 |
| Azher | mke2fs 1.42.9 (4-Feb-2014) Discarding device blocks: 4096/393168 done Filesystem label=cloudimg-rootfs OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) Stride=0 blocks, Stripe width=0 blocks 393216 inodes, 393168 blocks 19658 blocks (5.00%) reserved for the super user First data block=0 Maximum filesystem blocks=402653184 12 block groups 32768 blocks | 03:24 |
| Azher | mv: cannot remove '/tmp/image.ZmAXOSQD/built/dev/pts/ptmx': Operation not permitted mv: error reading '/tmp/image.ZmAXOSQD/built/proc/sysrq-trigger': Input/output error mv: failed to extend '/tmp/image.ZmAXOSQD/mnt/proc/sysrq-trigger': Input/output error | 03:24 |
| Azher | However the build goes through on my Ubuntu 13.10 | 03:25 |
| *** nosnos has quit IRC | 03:25 | |
| Azher | Could you kindly opine ? | 03:26 |
| *** gus has joined #tripleo | 03:26 | |
| *** nosnos has joined #tripleo | 03:26 | |
| greghaynes | error reading? | 03:26 |
| greghaynes | is your fs hosed? | 03:27 |
| *** pradeep1 has quit IRC | 03:27 | |
| gus | I'd heard a rumour that tripleo sets up the (overcloud?) keystone catalog once after everything else is finished, rather than as part of setting up each service like devstack does. Is that correct, and can someone point me to a relevant bit of code? | 03:27 |
| gus | (I'm lazy and failed to find it after a minute of looking) | 03:27 |
| Azher | yes it is | 03:28 |
| tchaypo | Azher: I'm not sure why, but it smells to me as though it's trying to use a partition that isn't big enough for the images it's making. | 03:28 |
| tchaypo | I'm wondering if perhaps it's trying to build on a ramdisk, and perhaps your 13.10 is 32-bit and 14.04 is 64-bit, which might mean less ram available for the ramdisk on the latter | 03:29 |
| greghaynes | gus: tripleo-incubator/scripts/devtest_overcloud.sh is pretty readable | 03:29 |
| Azher | I checked that as well but i do have enough space on my fs | 03:29 |
| tchaypo | but that's a big chain of speculation, I'd start by checking to see if you're running out of space somewhere | 03:29 |
| tchaypo | and you've already done that | 03:29 |
| tchaypo | whee | 03:29 |
| greghaynes | gus: init-keystone is the line you are probably interested in | 03:29 |
| greghaynes | which is a script in /scripts, also | 03:30 |
| Azher | Checked the RAM too and it is 32GB | 03:30 |
| Azher | from the H/W front there is no constraints | 03:30 |
| gus | greghaynes: haha! I wasn't even looking in the right git repo ;) thanks. | 03:31 |
| greghaynes | np | 03:31 |
| *** nosnos has quit IRC | 03:31 | |
| tchaypo | gus: well done sir | 03:31 |
| Azher | I am assuming this has got to do something with the mke2fs version (For Ubuntu 14.04 it is 1.42.9 (4-Feb-2014) ) | 03:31 |
| tchaypo | it isn't something I've seen | 03:32 |
| *** noslzzp has quit IRC | 03:32 | |
| Azher | The build goes through smoothly with the mke2fs version of 1.42.8 (20-Jun-2013) for Ubuntu 13.10 | 03:32 |
| gus | tchaypo: public displays of ineptitude is my super power. | 03:32 |
| SpamapS_ | hm | 03:34 |
| *** SpamapS_ is now known as SpamapS | 03:34 | |
| *** SpamapS has quit IRC | 03:34 | |
| *** SpamapS has joined #tripleo | 03:34 | |
| SpamapS | Azher: can you check dmesg | 03:36 |
| SpamapS | Azher: and /var/log/syslog | 03:36 |
| SpamapS | Azher: I'm wondering if there's something strange preventing the mv | 03:36 |
| Azher | I have tried multiple ways of getting out of this issue from past 2 weeks and have come here to the experts for help :) | 03:36 |
| SpamapS | Azher: it's failing to move a device node. That might be something a security module would block. | 03:37 |
| SpamapS | Azher: also can you try adding 'break=after-error' before your image build? | 03:37 |
| SpamapS | though I'm not sure if that will do anything given where it is failing | 03:38 |
| Azher | actually the build systems comes to a complete hault | 03:39 |
| Azher | at this point and i would have to restart the machine to get it back | 03:39 |
| Azher | at this point i would have to restart the system to even access ssh | 03:40 |
| *** pradeep has joined #tripleo | 03:43 | |
| SpamapS | whhhaaaa? | 03:44 |
| SpamapS | Azher: that sounds very not right. | 03:44 |
| SpamapS | Azher: before a build, can you paste here the output of 'mount' ? | 03:44 |
| Azher | It seems to me that though the image creation is being executed under chroot environment, somehow a process in the build probably mounted /dev from the real file system into /dev in the chroot then attempted to remove it rather than unmount it. | 03:44 |
| SpamapS | Azher: we had that problem a long, long time ago, and have some guards in place to prevent it. | 03:45 |
| * SpamapS has fond memories of wiping everybody's /dev out at the first day of the Havana mid-cycle | 03:46 | |
| Azher | could you kindly share that | 03:46 |
| SpamapS | Azher: they're in the code | 03:47 |
| SpamapS | hmmmmmm now | 03:48 |
| SpamapS | I had thought we had used --one-file-system but apparently not | 03:48 |
| SpamapS | Azher: can you do a build using -x over SSH, and show us the last 200 lines or so in a pastebin? | 03:50 |
| openstackgerrit | A change was merged to openstack/diskimage-builder: Echo that qemu-img convert is running https://review.openstack.org/127870 | 03:54 |
| *** pradeep1 has joined #tripleo | 03:54 | |
| *** pradeep has quit IRC | 03:56 | |
| openstackgerrit | greghaynes proposed a change to openstack/tripleo-image-elements: Add ability to enable debug logging on seed https://review.openstack.org/128164 | 03:59 |
| *** ChuckC has quit IRC | 04:00 | |
| *** ChuckC has joined #tripleo | 04:01 | |
| *** bvivek has joined #tripleo | 04:01 | |
| openstackgerrit | greghaynes proposed a change to openstack/tripleo-incubator: Add debug logging option for seed https://review.openstack.org/128165 | 04:01 |
| StevenK | greghaynes: I want your babies | 04:03 |
| greghaynes | Reviews on those two would be awesome to help debug a spurious CI fail going on right now | 04:03 |
| greghaynes | StevenK: :) | 04:03 |
| greghaynes | StevenK: Thats why I wanted the deep merge | 04:03 |
| StevenK | greghaynes: Aren't you missing ironic? | 04:03 |
| greghaynes | yep | 04:04 |
| Azher | Sure Clint! | 04:04 |
| greghaynes | I just copied what we currently turn on in t-h-t | 04:04 |
| StevenK | greghaynes: set -o pipefail in your install.d too | 04:04 |
| openstackgerrit | greghaynes proposed a change to openstack/tripleo-image-elements: Add ability to enable debug logging on seed https://review.openstack.org/128164 | 04:05 |
| greghaynes | ah! right after I pushed | 04:05 |
| *** jpeeler has quit IRC | 04:05 | |
| openstackgerrit | greghaynes proposed a change to openstack/tripleo-image-elements: Add ability to enable debug logging on seed https://review.openstack.org/128164 | 04:06 |
| greghaynes | bnemec: ^ reviews wanted plz :) | 04:06 |
| *** jpeeler has joined #tripleo | 04:07 | |
| *** jpeeler has joined #tripleo | 04:07 | |
| *** pradeep1 has quit IRC | 04:08 | |
| tchaypo | okay, progress is being made - turning nodes on one at a time and looking in syslog to find the mac addresses is working. It's probably slower, but I'm leaving it to run in a for loop while I do other work. Will revisit and pull data from logs tomorrow | 04:10 |
| StevenK | steven 13029 0.0 0.0 12468 1536 pts/2 SN+ 14:58 0:00 /bin/bash /home/steven/.cache/tripleo/tripleo-incubator/scripts/devtest_overcloud.sh | 04:12 |
| StevenK | Oh my | 04:12 |
| StevenK | I think that will fail hard | 04:12 |
| StevenK | But it means I have a working undercloud | 04:12 |
| tchaypo | SpamapS: I managed to trigger that on my laptop on my second or third day | 04:14 |
| tchaypo | I think i interrupted the build at the wrong point, and in trying to clean up it ended up deleting /dev | 04:15 |
| Azher | Clint: Please take a look at the log http://paste.openstack.org/show/120882/ | 04:15 |
| *** bvivek has quit IRC | 04:15 | |
| tchaypo | SpamapS: and when you have a chance, if you're not going away for the night yet, I'm still waiting for feedback ;) | 04:15 |
| *** daneyon has joined #tripleo | 04:16 | |
| *** rushiagr_away is now known as rushiagr | 04:17 | |
| *** yamahata has quit IRC | 04:19 | |
| *** yamahata has joined #tripleo | 04:19 | |
| *** nosnos has joined #tripleo | 04:20 | |
| *** nosnos has quit IRC | 04:27 | |
| *** nosnos has joined #tripleo | 04:27 | |
| *** ramishra has quit IRC | 04:31 | |
| *** ramishra has joined #tripleo | 04:32 | |
| *** ramishra_ has joined #tripleo | 04:35 | |
| *** rushiagr is now known as rushiagr_away | 04:36 | |
| *** ramishra has quit IRC | 04:36 | |
| *** pensu has joined #tripleo | 04:46 | |
| *** killer_prince is now known as lazy_prince | 04:50 | |
| *** rameshg87 has joined #tripleo | 04:59 | |
| *** tzumainn has quit IRC | 05:19 | |
| *** russell_h has quit IRC | 05:27 | |
| *** russell_h has joined #tripleo | 05:27 | |
| *** Azher has quit IRC | 05:34 | |
| *** lazy_prince has quit IRC | 05:40 | |
| *** ccrouch has quit IRC | 05:40 | |
| *** lxsli has quit IRC | 05:40 | |
| *** sbadia has quit IRC | 05:40 | |
| *** vipul has quit IRC | 05:40 | |
| *** dividehex has quit IRC | 05:40 | |
| *** rbrady has quit IRC | 05:40 | |
| *** mrunge has joined #tripleo | 05:45 | |
| *** dshulyak_ has joined #tripleo | 05:45 | |
| *** bvivek has joined #tripleo | 05:46 | |
| *** lazy_prince has joined #tripleo | 05:46 | |
| *** ccrouch has joined #tripleo | 05:46 | |
| *** sbadia has joined #tripleo | 05:46 | |
| *** lxsli has joined #tripleo | 05:46 | |
| *** vipul has joined #tripleo | 05:46 | |
| *** dividehex has joined #tripleo | 05:46 | |
| *** rbrady has joined #tripleo | 05:46 | |
| *** bvivek has quit IRC | 05:46 | |
| *** bvivek has joined #tripleo | 05:46 | |
| openstackgerrit | A change was merged to openstack/tuskar-ui: Updated from global requirements https://review.openstack.org/128048 | 05:47 |
| *** jerryz has quit IRC | 05:50 | |
| *** Azher has joined #tripleo | 05:51 | |
| *** rameshg87_ has joined #tripleo | 06:04 | |
| *** rameshg87 has quit IRC | 06:05 | |
| StevenK | Huh | 06:05 |
| *** rushiagr_away is now known as rushiagr | 06:05 | |
| StevenK | Sourcing overcloudrc and then running nova list gives no results, but 192.0.2.46 is running | 06:05 |
| lifeless | whats that | 06:09 |
| StevenK | 192.0.2.46 is a running demo instance | 06:10 |
| *** dshulyak_ has quit IRC | 06:10 | |
| StevenK | I'm ssh'd to it, neutron has ports for 10.0.0.2 and 192.0.2.46, but nova seems to know nothing about it | 06:10 |
| *** lazy_prince is now known as killer_prince | 06:12 | |
| openstackgerrit | Steve Kowalik proposed a change to openstack/os-cloud-config: Add support for creating baremetal flavors https://review.openstack.org/115872 | 06:12 |
| *** killer_prince is now known as lazy_prince | 06:13 | |
| openstackgerrit | Steve Kowalik proposed a change to openstack/os-cloud-config: Add support for injecting extra-specs into flavors https://review.openstack.org/123345 | 06:14 |
| *** pradeep has joined #tripleo | 06:18 | |
| *** rameshg87_ has quit IRC | 06:18 | |
| *** k4n0 has joined #tripleo | 06:24 | |
| *** pradeep has quit IRC | 06:26 | |
| *** pradeep has joined #tripleo | 06:26 | |
| *** jcoufal has joined #tripleo | 06:30 | |
| *** jprovazn has joined #tripleo | 06:34 | |
| *** pradeep has quit IRC | 06:36 | |
| lifeless | StevenK: nova list --all-users? | 06:37 |
| StevenK | lifeless: Hm, I shall keep that mind, but my test has been dead for about 15 minutes. :-( | 06:42 |
| *** pradeep has joined #tripleo | 06:53 | |
| *** gfidente has joined #tripleo | 07:00 | |
| *** ifarkas has joined #tripleo | 07:01 | |
| *** ramishra_ has quit IRC | 07:02 | |
| *** cwolferh has quit IRC | 07:03 | |
| *** jcoufal has quit IRC | 07:06 | |
| *** jcoufal has joined #tripleo | 07:06 | |
| *** shardy_z is now known as shardy | 07:09 | |
| *** pradeep has quit IRC | 07:10 | |
| *** pradeep has joined #tripleo | 07:13 | |
| *** jtomasek has joined #tripleo | 07:21 | |
| *** Guest85887 has quit IRC | 07:32 | |
| *** d0ugal has joined #tripleo | 07:32 | |
| *** d0ugal is now known as Guest50275 | 07:32 | |
| *** jp_at_hp has joined #tripleo | 07:34 | |
| *** Guest50275 has quit IRC | 07:35 | |
| openstackgerrit | Tomas Sedovic proposed a change to openstack/tripleo-heat-templates: Compute and controller templates without merge.py https://review.openstack.org/123100 | 07:35 |
| *** dmatthews__ has joined #tripleo | 07:36 | |
| *** pensu has quit IRC | 07:40 | |
| *** dmatthews__ has quit IRC | 07:46 | |
| *** dmatthews__ has joined #tripleo | 07:47 | |
| *** dmatthews__ has quit IRC | 07:52 | |
| *** d0ugal has joined #tripleo | 07:53 | |
| *** d0ugal is now known as Guest56922 | 07:53 | |
| *** Guest56922 is now known as d0ugal | 07:54 | |
| *** d0ugal has joined #tripleo | 07:54 | |
| *** jistr has joined #tripleo | 07:57 | |
| *** akrivoka has joined #tripleo | 07:57 | |
| *** rdopieralski has joined #tripleo | 08:00 | |
| *** regebro has joined #tripleo | 08:06 | |
| *** shadower has quit IRC | 08:11 | |
| *** shadower has joined #tripleo | 08:13 | |
| openstackgerrit | A change was merged to openstack/tuskar-ui: New node registration and upload dialogs https://review.openstack.org/126576 | 08:16 |
| *** jp_at_hp has quit IRC | 08:22 | |
| *** jp_at_hp has joined #tripleo | 08:22 | |
| *** pensu has joined #tripleo | 08:23 | |
| *** regebro has quit IRC | 08:27 | |
| *** lucasagomes has joined #tripleo | 08:34 | |
| *** rushiagr is now known as rushiagr_away | 08:41 | |
| *** lazy_prince has quit IRC | 08:42 | |
| *** rushiagr_away is now known as rushiagr | 08:43 | |
| openstackgerrit | A change was merged to openstack/tripleo-image-elements: Support Debian distro for iptables https://review.openstack.org/111369 | 08:46 |
| *** mrunge has quit IRC | 08:50 | |
| openstackgerrit | A change was merged to openstack/tripleo-image-elements: Add ability to enable debug logging on seed https://review.openstack.org/128164 | 08:50 |
| *** killer_prince has joined #tripleo | 08:53 | |
| openstackgerrit | A change was merged to openstack/tripleo-image-elements: Fix os-svc-install https://review.openstack.org/127363 | 08:53 |
| *** killer_prince is now known as lazy_prince | 08:53 | |
| *** mrunge has joined #tripleo | 08:54 | |
| *** rushiagr is now known as rushiagr_away | 08:54 | |
| *** pradeep has quit IRC | 08:57 | |
| jp_at_hp | StevenK: you should recheck https://review.openstack.org/#/c/128165/ | 08:59 |
| *** k4n0 has quit IRC | 09:01 | |
| *** rushiagr_away is now known as rushiagr | 09:02 | |
| *** lazy_prince has quit IRC | 09:07 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/diskimage-builder: disable all interfaces on eni systems https://review.openstack.org/128213 | 09:07 |
| *** rlandy has joined #tripleo | 09:10 | |
| *** k4n0 has joined #tripleo | 09:13 | |
| *** yuanying has quit IRC | 09:15 | |
| *** yuanying has joined #tripleo | 09:15 | |
| *** killer_prince has joined #tripleo | 09:17 | |
| *** killer_prince is now known as lazy_prince | 09:17 | |
| *** yuanying has quit IRC | 09:20 | |
| *** rushiagr is now known as rushiagr_away | 09:21 | |
| *** rushiagr_away is now known as rushiagr | 09:23 | |
| openstackgerrit | Giulio Fidente proposed a change to openstack/tripleo-heat-templates: Passthrough{Specific} and allNodesConfig for BlockStorage nodes https://review.openstack.org/112075 | 09:27 |
| *** jprovazn has quit IRC | 09:27 | |
| *** regebro has joined #tripleo | 09:28 | |
| *** regebro has quit IRC | 09:29 | |
| *** pelix1 has joined #tripleo | 09:32 | |
| *** rushiagr has quit IRC | 09:36 | |
| *** regebro has joined #tripleo | 09:40 | |
| *** lazy_prince has quit IRC | 09:41 | |
| *** funzo has joined #tripleo | 09:44 | |
| *** funzo_ has quit IRC | 09:44 | |
| *** yuanying has joined #tripleo | 09:45 | |
| *** killer_prince has joined #tripleo | 09:49 | |
| *** killer_prince is now known as lazy_prince | 09:49 | |
| *** Azher has quit IRC | 09:49 | |
| *** mrunge has quit IRC | 09:50 | |
| *** rushiagr_away has joined #tripleo | 09:50 | |
| rdopieralski | Anybody has an idea why am I not allowed to upload patches to tuskar-ui? | 09:50 |
| rdopieralski | To ssh://thesheep@review.openstack.org:29418/openstack/tuskar-ui.git ! [remote rejected] HEAD -> refs/publish/master/fix-required-upload (you are not allowed to upload merges) | 09:50 |
| *** mrunge has joined #tripleo | 09:51 | |
| rdopieralski | aah, it's a merge | 09:51 |
| rdopieralski | sorry | 09:51 |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/tuskar-ui: Fix the plan validation https://review.openstack.org/128228 | 10:01 |
| *** lazy_prince has quit IRC | 10:09 | |
| *** lazy_prince has joined #tripleo | 10:09 | |
| *** AzherKhan has joined #tripleo | 10:13 | |
| *** Azher has joined #tripleo | 10:22 | |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/tuskar-ui: Add password reveal icon to new nodes in node register https://review.openstack.org/128231 | 10:24 |
| *** pensu has quit IRC | 10:29 | |
| *** k4n0 has quit IRC | 10:33 | |
| *** yamahata has quit IRC | 10:37 | |
| *** shadower has quit IRC | 10:44 | |
| *** shadower has joined #tripleo | 10:45 | |
| *** shadower has quit IRC | 10:46 | |
| *** pblaho has joined #tripleo | 10:47 | |
| *** pblaho has quit IRC | 10:47 | |
| *** shadower has joined #tripleo | 10:47 | |
| *** pblaho has joined #tripleo | 10:48 | |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/tuskar-ui: Add support for min, max and step in the NumberPicker https://review.openstack.org/128235 | 10:48 |
| *** dtantsur is now known as dtantsur|lunch | 10:55 | |
| openstackgerrit | Lennart Regebro proposed a change to openstack/tuskar-ui: Catch missing resource error when deploying https://review.openstack.org/128238 | 10:59 |
| *** derekh has joined #tripleo | 11:06 | |
| *** jerryz has joined #tripleo | 11:08 | |
| *** Azher has quit IRC | 11:09 | |
| *** AzherKhan has quit IRC | 11:10 | |
| openstackgerrit | Tomas Sedovic proposed a change to openstack/tuskar: Don't limit the count parameter for role scaling https://review.openstack.org/128244 | 11:24 |
| *** jerryz has quit IRC | 11:24 | |
| *** cody-somerville has joined #tripleo | 11:26 | |
| *** AzherKhan has joined #tripleo | 11:27 | |
| *** AzherKhan has quit IRC | 11:29 | |
| *** jistr is now known as jistr|english | 11:29 | |
| *** rushiagr_away is now known as rushiagr | 11:29 | |
| *** pradeep has joined #tripleo | 11:35 | |
| openstackgerrit | A change was merged to openstack/tripleo-incubator: Add debug logging option for seed https://review.openstack.org/128165 | 11:36 |
| openstackgerrit | A change was merged to openstack/tripleo-image-elements: SELinux: swift denied access to /var/lib/swift/.local https://review.openstack.org/127104 | 11:37 |
| *** bvivek has quit IRC | 11:39 | |
| *** lucasagomes is now known as lucas-hungry | 11:40 | |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/tuskar-ui: Add support for min, max and step in the NumberPicker https://review.openstack.org/128235 | 11:40 |
| *** cody-somerville has quit IRC | 11:46 | |
| *** ramishra has joined #tripleo | 11:53 | |
| *** yamahata has joined #tripleo | 11:56 | |
| openstackgerrit | A change was merged to openstack/tripleo-image-elements: Configure logging for keepalived https://review.openstack.org/125668 | 11:58 |
| *** shadower has quit IRC | 12:00 | |
| *** shadower has joined #tripleo | 12:00 | |
| *** _crobertsrh is now known as crobertsrh | 12:05 | |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/tuskar-ui: Names for flavor suggestions https://review.openstack.org/128252 | 12:05 |
| *** noslzzp has joined #tripleo | 12:11 | |
| *** k4n0 has joined #tripleo | 12:14 | |
| *** athomas has quit IRC | 12:17 | |
| *** jkraj has joined #tripleo | 12:20 | |
| *** jerryz has joined #tripleo | 12:21 | |
| *** athomas has joined #tripleo | 12:25 | |
| *** ramishra has quit IRC | 12:26 | |
| *** jprovazn has joined #tripleo | 12:28 | |
| *** lucas-hungry is now known as lucasagomes | 12:29 | |
| openstackgerrit | Peter Belanyi proposed a change to openstack/tripleo-heat-templates: Add converted version of block and object storage https://review.openstack.org/123713 | 12:30 |
| *** jdob has joined #tripleo | 12:31 | |
| *** prad has quit IRC | 12:31 | |
| *** lazy_prince is now known as killer_prince | 12:32 | |
| openstackgerrit | Radomir Dopieralski proposed a change to openstack/tuskar-ui: Remove the buttons from modal templates https://review.openstack.org/128256 | 12:34 |
| *** ramishra has joined #tripleo | 12:36 | |
| *** dprince has joined #tripleo | 12:37 | |
| openstackgerrit | A change was merged to stackforge/kolla: batch operations in nova-compute dockerfile https://review.openstack.org/128099 | 12:38 |
| openstackgerrit | A change was merged to stackforge/kolla: Nova-api should not set up the database https://review.openstack.org/128132 | 12:40 |
| *** ramishra has quit IRC | 12:52 | |
| *** jistr|english is now known as jistr | 12:54 | |
| *** tzumainn has joined #tripleo | 12:56 | |
| *** noslzzp has quit IRC | 12:58 | |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Config will be done by the controller https://review.openstack.org/128261 | 12:58 |
| *** morazi has joined #tripleo | 13:00 | |
| *** rhallisey has joined #tripleo | 13:00 | |
| *** noslzzp has joined #tripleo | 13:02 | |
| *** sseago has quit IRC | 13:03 | |
| *** sseago has joined #tripleo | 13:04 | |
| openstackgerrit | Tomas Sedovic proposed a change to openstack/tuskar: Don't limit the count parameter for role scaling https://review.openstack.org/128244 | 13:05 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tripleo-image-elements: Add svc-map to cinder-api https://review.openstack.org/104198 | 13:07 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tripleo-image-elements: Add svc-map to cinder-api https://review.openstack.org/104198 | 13:08 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tripleo-image-elements: Add svc-map to cinder-api https://review.openstack.org/104198 | 13:10 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tripleo-image-elements: Add svc-map to cinder-api https://review.openstack.org/104198 | 13:12 |
| *** Shrews has quit IRC | 13:12 | |
| *** Shrews has joined #tripleo | 13:14 | |
| *** julim has joined #tripleo | 13:18 | |
| *** jtomasek has quit IRC | 13:18 | |
| *** jroll has quit IRC | 13:19 | |
| *** jroll has joined #tripleo | 13:19 | |
| openstackgerrit | therese-mchale proposed a change to openstack/tripleo-image-elements: Allow configurable garp_master_refresh https://review.openstack.org/126622 | 13:20 |
| *** sseago has quit IRC | 13:21 | |
| *** sseago has joined #tripleo | 13:22 | |
| openstackgerrit | A change was merged to openstack/tuskar: Fixed count constraint https://review.openstack.org/128083 | 13:23 |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Nova-controller needs a start script https://review.openstack.org/128273 | 13:25 |
| *** pblaho has quit IRC | 13:31 | |
| *** jehb has joined #tripleo | 13:32 | |
| openstackgerrit | A change was merged to openstack/os-net-config: Update child_members to use a Set https://review.openstack.org/122507 | 13:33 |
| *** nijaba_ is now known as nijaba | 13:34 | |
| *** pblaho has joined #tripleo | 13:34 | |
| openstackgerrit | A change was merged to openstack/os-net-config: Add a .gitreview file for os-net-config https://review.openstack.org/122192 | 13:34 |
| *** nijaba has quit IRC | 13:34 | |
| *** nijaba has joined #tripleo | 13:34 | |
| *** wfoster has quit IRC | 13:35 | |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/diskimage-builder: 'iso' element to build bootable ISO images. https://review.openstack.org/123642 | 13:39 |
| *** wfoster has joined #tripleo | 13:39 | |
| *** nijaba has quit IRC | 13:40 | |
| *** andrearosa has joined #tripleo | 13:41 | |
| rhallisey | larsks, morning | 13:42 |
| rhallisey | larsks, doesn't nova-compute use port 8774? | 13:43 |
| larsks | rhallisey: nova-api uses port 8774. Nova-compute doesn't have a REST interface; it talks to the controllers via the RPC bus. | 13:43 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tuskar-ui: Enable ExtraConfig Parameter in Tuskar UI https://review.openstack.org/127621 | 13:44 |
| rhallisey | larsks, do I only need 8774 then? | 13:45 |
| *** andreaf has joined #tripleo | 13:45 | |
| larsks | Well, you need 8775 for the metadata api. | 13:45 |
| rhallisey | ok | 13:45 |
| *** nijaba has joined #tripleo | 13:45 | |
| *** nijaba has quit IRC | 13:45 | |
| *** nijaba has joined #tripleo | 13:45 | |
| openstackgerrit | Ryan Brady proposed a change to openstack/tuskar-ui: Enable ExtraConfig Parameter in Tuskar UI https://review.openstack.org/127621 | 13:46 |
| rhallisey | larsks, I have another question.. what container port should I use for the nova-controller-pod? | 13:46 |
| larsks | rhallisey: and 8773 for the ec2 api. | 13:46 |
| larsks | rhallisey: You would specify the ports for the services it is running. | 13:47 |
| larsks | If your controller is running nova-api, that probably means 8773, 8774, 8775. | 13:47 |
| larsks | If this is for a service configuration, that probably means three seperate services. | 13:47 |
| openstackgerrit | A change was merged to openstack/tuskar-ui: Catch missing resource error when deploying https://review.openstack.org/128238 | 13:48 |
| larsks | I don't think you can register a multi-port service at this time. | 13:48 |
| rhallisey | larsks, ok thanks, I'll add these in coming patches | 13:48 |
| larsks | Awesome. | 13:48 |
| *** nijaba has quit IRC | 13:55 | |
| *** rushiagr is now known as rushiagr_away | 13:55 | |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Update nova-controller's container ports https://review.openstack.org/128294 | 13:55 |
| *** dtantsur|lunch is now known as dtantsur | 13:55 | |
| *** nijaba has joined #tripleo | 14:00 | |
| *** nijaba has quit IRC | 14:00 | |
| *** nijaba has joined #tripleo | 14:00 | |
| openstackgerrit | Lennart Regebro proposed a change to openstack/tripleo-image-elements: Default to the "scale" deployment mode https://review.openstack.org/128296 | 14:01 |
| *** yamahata has quit IRC | 14:03 | |
| *** yamahata has joined #tripleo | 14:04 | |
| *** pensu has joined #tripleo | 14:07 | |
| *** rdopieralski has quit IRC | 14:07 | |
| *** radez_g0n3 is now known as radez | 14:07 | |
| sdake | morning | 14:14 |
| larsks | sdake: morning! | 14:19 |
| sdake | morning larsks | 14:20 |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Add kube services https://review.openstack.org/128302 | 14:20 |
| rhallisey | morning :) | 14:20 |
| sdake | larsks did you mean to only +1 this change? https://review.openstack.org/#/c/128294/ | 14:21 |
| sdake | I generally use +1 when I'm not sure and want other reviwers to take a critcal look | 14:22 |
| larsks | Nope, I meant +2. Fixed. | 14:22 |
| openstackgerrit | A change was merged to stackforge/kolla: Update nova-controller's container ports https://review.openstack.org/128294 | 14:22 |
| openstackgerrit | A change was merged to stackforge/kolla: Add kube services https://review.openstack.org/128302 | 14:23 |
| *** noslzzp has quit IRC | 14:23 | |
| openstackgerrit | Ryan Brady proposed a change to openstack/diskimage-builder: Enable RHEL Registration https://review.openstack.org/123145 | 14:25 |
| *** noslzzp has joined #tripleo | 14:26 | |
| *** pblaho has quit IRC | 14:26 | |
| openstackgerrit | Tom Cammann proposed a change to openstack/tripleo-heat-templates: Refactor Heat dependencies for nova compute https://review.openstack.org/122384 | 14:29 |
| *** noslzzp has quit IRC | 14:31 | |
| *** pblaho has joined #tripleo | 14:34 | |
| *** pblaho has quit IRC | 14:36 | |
| *** pblaho has joined #tripleo | 14:36 | |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Expose ports needed by the nova-controller https://review.openstack.org/128308 | 14:38 |
| *** nosnos has quit IRC | 14:40 | |
| *** nosnos has joined #tripleo | 14:40 | |
| *** k4n0 has quit IRC | 14:42 | |
| *** Dafna has quit IRC | 14:42 | |
| *** nosnos has quit IRC | 14:45 | |
| *** noslzzp has joined #tripleo | 14:46 | |
| *** pblaho has quit IRC | 14:46 | |
| *** k4n0 has joined #tripleo | 14:49 | |
| *** k4n0 has quit IRC | 14:50 | |
| *** k4n0 has joined #tripleo | 14:50 | |
| sdake | larsks does this pod start an image at all? | 14:50 |
| sdake | https://github.com/anguslees/kube-openstack/blob/master/kubernetes-in/nova-db-sync-pod.yaml | 14:51 |
| larsks | Sure. It says: image: localhost:4999/nova | 14:51 |
| larsks | The container runs some commands and then exits. | 14:51 |
| larsks | See my reply to Angus on the ML. | 14:51 |
| larsks | This is probably a good model to pursue. | 14:51 |
| sdake | I didn't see it | 14:51 |
| larsks | I just sent < 5 minutes ago. | 14:51 |
| sdake | ya me too | 14:51 |
| sdake | I created bp for his first idea | 14:52 |
| larsks | I think is second idea is more valuable. | 14:52 |
| larsks | I think that will dramatically simplify the design of things like replicationcontrollers. | 14:52 |
| *** ramishra has joined #tripleo | 14:52 | |
| sdake | the replication controller I implemented is pretty braindead simple | 14:53 |
| sdake | whether it works who knows :-) | 14:53 |
| *** pcrews has joined #tripleo | 14:53 | |
| *** ramishra has quit IRC | 14:54 | |
| larsks | To make service and database initialization happen sanely you need to make things not simple. Angus' idea would let you make things simple again. | 14:54 |
| *** ramishra has joined #tripleo | 14:54 | |
| *** Dafna has joined #tripleo | 14:55 | |
| sdake | i tihnk it is simple enough just to spin until db is active | 14:55 |
| larsks | It's not that simple. E.g., if you are setting up a keystone replication controller, who sets up the database schema? | 14:55 |
| larsks | How do you know which container has started "first"? They can't all spin on the database. | 14:55 |
| sdake | you mean for horizontal scaling? | 14:56 |
| larsks | THat is what replicationcontrollers are for. | 14:56 |
| sdake | yes | 14:56 |
| sdake | actually I use them differently in libvirt | 14:56 |
| sdake | to ensure libvirt + compute runs on each node | 14:56 |
| larsks | In anycase, that is the problem: synchronizing one-time service setup, like database schema initialization, across many pods. | 14:56 |
| sdake | I didn't comprehend you wer etlaking about hs on non-compute services - in that case if they export a port, yes 1 service to create db makes sense | 14:56 |
| sdake | rhallisey how close is your nova code to ready to be integrated? | 14:57 |
| rhallisey | I'm testing now | 14:57 |
| sdake | it may not work with current compute | 14:58 |
| sdake | curreent compute creates a db iirc | 14:58 |
| rhallisey | I may have a few questions about whether I'm doing it correct | 14:58 |
| larsks | Compute should not be creating a database...the compute does not talk to the database (only the conductor does). | 14:58 |
| *** noslzzp has quit IRC | 14:58 | |
| rhallisey | larsks, do I just want to see if nova-api works? | 14:59 |
| larsks | I think that's a good start, yeah. | 14:59 |
| *** ramishra has quit IRC | 14:59 | |
| larsks | Can we get a successful repsonse from "nova list"? | 14:59 |
| sdake | larsks yes I know it shouldn't create the db | 14:59 |
| larsks | Cool. | 14:59 |
| sdake | we already had this discussion with rhallisey | 14:59 |
| larsks | Okay. It sounded like it hadn't been fixed yet, which is why I mentioned. | 14:59 |
| sdake | it probably hasnt | 15:00 |
| sdake | last night was the first time i got glance to work | 15:00 |
| sdake | and then i ptfo'ed | 15:00 |
| *** untriaged-bot has joined #tripleo | 15:00 | |
| untriaged-bot | Untriaged bugs so far: | 15:00 |
| untriaged-bot | https://bugs.launchpad.net/tripleo/+bug/1380608 | 15:00 |
| uvirtbot | Launchpad bug 1380608 in tripleo "Convert block-storage-nfs.yaml to StructuredDeployment" [Undecided,Confirmed] | 15:00 |
| *** untriaged-bot has quit IRC | 15:00 | |
| *** dkehn has quit IRC | 15:00 | |
| *** noslzzp has joined #tripleo | 15:01 | |
| *** dkehn has joined #tripleo | 15:02 | |
| *** jpeeler has quit IRC | 15:04 | |
| *** jpeeler has joined #tripleo | 15:04 | |
| *** jpeeler has joined #tripleo | 15:04 | |
| *** jerryz has quit IRC | 15:04 | |
| *** jtomasek has joined #tripleo | 15:06 | |
| *** jkraj has quit IRC | 15:07 | |
| *** marun has joined #tripleo | 15:07 | |
| *** andrearosa has quit IRC | 15:10 | |
| *** andrearosa has joined #tripleo | 15:10 | |
| *** pblaho has joined #tripleo | 15:12 | |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: do not perform database or keystone setup in nova-compute https://review.openstack.org/128330 | 15:14 |
| *** pradk has joined #tripleo | 15:15 | |
| greghaynes | derekh: on https://review.openstack.org/#/c/71837/2/toci_devtest.sh - where were you seeing the pip ignoring packages due to pre-release? | 15:19 |
| greghaynes | I would expect to see that, but cant find it in the logs | 15:19 |
| sdake | why are we multilining crudini? | 15:20 |
| sdake | my text editor has 80 chars wide ;-) | 15:21 |
| greghaynes | oh, I see | 15:22 |
| greghaynes | someone went in an changed those packages to be installed as --pre | 15:22 |
| larsks | I think it's more readable, but I don't really care. | 15:22 |
| derekh | greghaynes glad you found it, I was struggling to explain it | 15:22 |
| greghaynes | yea | 15:22 |
| greghaynes | I am not very fond of this fix though | 15:23 |
| openstackgerrit | A change was merged to stackforge/kolla: do not perform database or keystone setup in nova-compute https://review.openstack.org/128330 | 15:23 |
| greghaynes | I think you can set pre=true in the pip config, then we could turn it on for just CI | 15:24 |
| *** pblaho has quit IRC | 15:28 | |
| *** yamahata has quit IRC | 15:31 | |
| *** rushiagr_away is now known as rushiagr | 15:32 | |
| *** cwolferh has joined #tripleo | 15:33 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/diskimage-builder: disable all interfaces on eni systems https://review.openstack.org/128213 | 15:35 |
| openstackgerrit | Steven Dake proposed a change to stackforge/kolla: Have qemu uri point to correct place https://review.openstack.org/128337 | 15:36 |
| greghaynes | clarkb: are the ZUUL_* env vars documented somewhere? | 15:38 |
| clarkb | greghaynes: yup one sec | 15:38 |
| greghaynes | trying to figure out the ZUUL_CHANGES var | 15:38 |
| clarkb | greghaynes: http://ci.openstack.org/zuul/launchers.html#zuul-parameters | 15:39 |
| greghaynes | tyty | 15:39 |
| greghaynes | derekh: I notice you have a test for ZUUL_CHANGES and if its empty it installs all packages from HEAD... any idea why? | 15:40 |
| greghaynes | derekh: maybe you meant ZUUL_CHANGE_IDS? | 15:40 |
| greghaynes | still unsure what the case where its empty is about, though | 15:41 |
| *** andreaf has quit IRC | 15:41 | |
| clarkb | greghaynes: if you open your test in jenkins you can look at the parameters a given job ran with too | 15:41 |
| derekh | greghaynes: iirc it tested if it was empty incase we ever had a periodic job to test HEAD of everything | 15:41 |
| clarkb | greghaynes: often helpful if ou want concrete data | 15:41 |
| greghaynes | derekh: ah, I might break that out into its own param | 15:42 |
| derekh | greghaynes: ok | 15:43 |
| *** rameshg87 has joined #tripleo | 15:46 | |
| *** crobertsrh has left #tripleo | 15:46 | |
| rameshg87 | SpamapS: dprince: please have a look at https://review.openstack.org/#/c/123642/ . i have fixed the CI failure. :) | 15:47 |
| *** rameshg87 has quit IRC | 15:51 | |
| *** k4n0 has quit IRC | 15:51 | |
| *** ramishra has joined #tripleo | 15:55 | |
| sdake | we ain't going out like that | 15:57 |
| sdake | cypress hill @ 8am = win | 15:57 |
| *** ramishra has quit IRC | 16:00 | |
| sdake | rhallisey nova-controller has all 3 of those ports? | 16:00 |
| sdake | I thought some were in nova-api etc | 16:00 |
| rhallisey | sdake, I wasn't sure where I should expose them | 16:01 |
| rhallisey | if that's what you mean | 16:01 |
| *** spzala has joined #tripleo | 16:01 | |
| sdake | Nova-api 8773 (for EC2 API) | 16:02 |
| sdake | 8774 (for openstack API) | 16:02 |
| sdake | 8775 (metadata port) | 16:02 |
| sdake | probably all from nova-api rather then =nova-controller image | 16:02 |
| sdake | nova-novncproxy 6080 | 16:05 |
| sdake | 5800/5900 (VNC) | 16:05 |
| sdake | rhallisey I'd say expose em from the api image and we can go from there | 16:06 |
| rhallisey | ok | 16:06 |
| sdake | i'll see if rbryant knows for certain which bins create which ports | 16:06 |
| *** jang2 has joined #tripleo | 16:08 | |
| rhallisey | larsks, my keystone pod doesn't startup and I get nothing from journalctl | 16:08 |
| larsks | rhallisey: "docker logs <containerid>" | 16:08 |
| sdake | is it not entering the running state? | 16:09 |
| sdake | or still in waiting | 16:09 |
| larsks | sdake: nova-api creates 8773, 8774, 8775. the vnc services only need to be opened on the compute side. vncproxy opens 6080. | 16:09 |
| rhallisey | it waited then went to terminated | 16:09 |
| sdake | sudo docker -ps -a | 16:09 |
| sdake | this will show you old processes | 16:09 |
| sdake | find one that is bust | 16:09 |
| larsks | Just: docker ps -a (no -ps) | 16:09 |
| sdake | and run docker logs id | 16:09 |
| rhallisey | larsks, do I expose them in the nova-api image? | 16:10 |
| sdake | right no -ps typo :) | 16:10 |
| larsks | rhallisey: If that's where you're running the services, sure. | 16:10 |
| sdake | rhallisey I believe so | 16:10 |
| larsks | rhallisey: you can use the glance pod as an example, maybe, since it runs two containers, each with a different service. | 16:11 |
| larsks | Oh, hey, that reminds me -- I have run into some selinux related quirks. You may want to ensure you "setenforce 0" on your minions. | 16:12 |
| larsks | Need to file some selinux-policy bugs, I think, or have the templates install a local policy. | 16:12 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tuskar-ui: Enable ExtraConfig Parameter in Tuskar UI https://review.openstack.org/127621 | 16:12 |
| sdake | rhallisey you need to split out nova-api into 3 pod blobs | 16:13 |
| sdake | nova-api | 16:13 |
| sdake | nova-api-ec2 | 16:13 |
| sdake | nova-api-metadata | 16:13 |
| rhallisey | err ok maybe not ready to test yet | 16:14 |
| sdake | and probably nova-api-os-compute | 16:14 |
| larsks | I am less certain about that last one. I think that nova-api by itself will start up the metadata service. Try it, I guess. | 16:14 |
| openstackgerrit | Ghe Rivero proposed a change to openstack/diskimage-builder: disable all interfaces on eni systems https://review.openstack.org/128213 | 16:14 |
| sdake | these are the binaries that are installed by packstack atleaset | 16:14 |
| larsks | Looking at my system right now, nova-api is the things that has ports 8773, 8774, and 8775 open. There are no separate metadata or ec2 services. | 16:15 |
| sdake | yup thats right larsks | 16:15 |
| sdake | rhallisey ignore me ;) | 16:15 |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: add support for pre-commit validation https://review.openstack.org/128345 | 16:18 |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: add support for pre-commit validation https://review.openstack.org/128345 | 16:19 |
| larsks | curious if people find that useful ^^^^ | 16:20 |
| *** pblaho has joined #tripleo | 16:21 | |
| *** pblaho has quit IRC | 16:25 | |
| *** pblaho has joined #tripleo | 16:26 | |
| sdake | larsks that would be better implemented as a check | 16:29 |
| larsks | sdake: I think the scripts can be used as a check. But I think it's a good idea to not commit broken code, too :) | 16:29 |
| larsks | So I think it solves both problems. | 16:29 |
| larsks | We can use the validation scripts in the gate, but if people *want* they can also run them locally as pre-commit hooks. | 16:30 |
| larsks | But it is not necessary to do so. | 16:30 |
| larsks | Nor is it even enforceable :) | 16:30 |
| sdake | ya wfm | 16:30 |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: add support for pre-commit validation https://review.openstack.org/128345 | 16:35 |
| *** rushiagr is now known as rushiagr_away | 16:36 | |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: add better validation checks (part 1) https://review.openstack.org/128345 | 16:40 |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: modify tox config to use new validation code https://review.openstack.org/128355 | 16:42 |
| openstackgerrit | Ryan Brady proposed a change to openstack/tuskar-ui: Enable ExtraConfig Parameter in Tuskar UI https://review.openstack.org/127621 | 16:43 |
| *** derekh has quit IRC | 16:45 | |
| jpeeler | i just attempted to do start and keystone is complaining about the database not being available, but it seems to be looking at the wrong ip (scheduled on the other minion) | 16:46 |
| larsks | jpeeler: it should be looking at the *local* minion ip. | 16:47 |
| larsks | That is where the service proxy lives. | 16:47 |
| larsks | But I am actually looking into exactly that error right now for rhallisey. | 16:48 |
| sdake | jpeeler ya it doesn't make sense to me either , but it works ;) | 16:48 |
| larsks | jpeeler: pods talk to the local service proxy, which is responsible for routing connections to the right endpoint. | 16:48 |
| jpeeler | larsks: i keep forgetting about the proxy magic, so yes it's looking at the right IP then | 16:48 |
| sdake | I don't understand why the local proxy should be involved | 16:48 |
| sdake | seems like extra pian for no gain | 16:48 |
| larsks | sdake: Lots of gain. | 16:49 |
| larsks | sdake: Ability to locate a service even if the pods move. | 16:49 |
| sdake | ya that is true | 16:49 |
| sdake | ok arch makes sense in my head now thanks for that point | 16:49 |
| larsks | jpeeler: hold on a few minutes; I am looking into something rightnow. | 16:50 |
| openstackgerrit | greghaynes proposed a change to openstack-infra/tripleo-ci: WIP: build a local pip mirror of python packages being tested https://review.openstack.org/71837 | 16:52 |
| *** jistr has quit IRC | 16:53 | |
| *** athomas has quit IRC | 16:54 | |
| *** ramishra has joined #tripleo | 16:56 | |
| larsks | jpeeler: okay, rhallisey's problem was simply kernel-version-too-old. | 16:56 |
| larsks | Any chance I can log into your environment? | 16:56 |
| jpeeler | sure, but that shouldn't be the case here | 16:56 |
| larsks | Right. | 16:56 |
| jpeeler | one sec | 16:56 |
| sdake | too bad one of our use cases is not deploying a whole bunch of openstack in one org | 16:58 |
| sdake | because this tool rocks for that ;) | 16:58 |
| *** ramishra has quit IRC | 17:00 | |
| openstackgerrit | Steven Dake proposed a change to stackforge/kolla: Start nova-controller services and pods https://review.openstack.org/128363 | 17:02 |
| *** lucasagomes is now known as lucas-dinner | 17:04 | |
| *** jcoufal has quit IRC | 17:04 | |
| openstackgerrit | Ryan Brady proposed a change to openstack/tuskar-ui: Enable ExtraConfig Parameter in Tuskar UI https://review.openstack.org/127621 | 17:07 |
| *** jp_at_hp has quit IRC | 17:07 | |
| openstackgerrit | Tomas Sedovic proposed a change to openstack/tripleo-specs: Remove merge.py from TripleO Heat Templates https://review.openstack.org/97939 | 17:08 |
| openstackgerrit | Steven Dake proposed a change to stackforge/kolla: nova-compute-service doesn't expose a port https://review.openstack.org/128366 | 17:08 |
| *** dsneddon_ has quit IRC | 17:21 | |
| *** dsneddon has joined #tripleo | 17:21 | |
| *** ifarkas has quit IRC | 17:21 | |
| *** jtomasek has quit IRC | 17:22 | |
| *** athomas has joined #tripleo | 17:23 | |
| openstackgerrit | Ryan Hallisey proposed a change to stackforge/kolla: Expose ports needed by the nova-controller https://review.openstack.org/128308 | 17:26 |
| *** pcrews has quit IRC | 17:26 | |
| *** pelix1 has quit IRC | 17:27 | |
| *** pcrews has joined #tripleo | 17:28 | |
| *** jtomasek has joined #tripleo | 17:38 | |
| openstackgerrit | Tzu-Mainn Chen proposed a change to openstack/tuskar-ui: Fix deployment check to only look at active nodes https://review.openstack.org/128374 | 17:38 |
| *** dtantsur is now known as dtantsur|afk | 17:39 | |
| jpeeler | larsks: reproduced list minions not working again, restart of kube-apiserver on master fixed it | 17:40 |
| larsks | Interesting. Sounds like a race somewhere. | 17:41 |
| larsks | I will try to deploy a few times and see if I can trigger it. | 17:41 |
| jpeeler | since i haven't done anything yet, i could try again if that's helpful | 17:42 |
| daneyon | rhallisey: what are your plans for nova vnc? will this be in a separate container from api? | 17:45 |
| rhallisey | yes | 17:46 |
| rhallisey | daneyon, I'm going to put it into a separate container | 17:46 |
| larsks | jpeeler: Nah, I am too distracted right now. BUt thanks. | 17:46 |
| daneyon | rhallisey: OK. Thanks! | 17:47 |
| *** jtomasek has quit IRC | 17:48 | |
| *** pcrews has quit IRC | 17:54 | |
| *** ChuckC has quit IRC | 17:56 | |
| *** athomas has quit IRC | 17:56 | |
| *** pcrews has joined #tripleo | 17:57 | |
| *** ramishra has joined #tripleo | 17:57 | |
| *** wendar_ is now known as wendar | 18:00 | |
| *** ramishra has quit IRC | 18:01 | |
| *** d0ugal has quit IRC | 18:04 | |
| sdake | jpeeler I have that problem as well | 18:04 |
| sdake | you also need to restart kube-scheduler iirc | 18:04 |
| *** d0ugal has joined #tripleo | 18:06 | |
| *** d0ugal is now known as Guest86578 | 18:06 | |
| sdake | how does one convert json to yaml easily? | 18:09 |
| sdake | the web converter no worky for me | 18:09 |
| clarkb | sdake: SpamapS has a tool for that somewhere | 18:09 |
| mordred | flamel | 18:11 |
| mordred | sdake: "pip install flaeml" | 18:11 |
| mordred | gah | 18:11 |
| mordred | sdake: "pip install flamel" | 18:11 |
| sdake | thanks mordred | 18:11 |
| mordred | sdake: then "flamel foo.json > foo.yaml" or "flamel foo.yaml > foo.json" | 18:11 |
| sdake | i manged to find an online tool that get er done tho :) | 18:11 |
| *** radez is now known as radez_g0n3 | 18:11 | |
| mordred | well, I'll say having flamel around has been really useful since I learned about it | 18:11 |
| openstackgerrit | Steven Dake proposed a change to stackforge/kolla: Convert nova-compute-replication to yaml https://review.openstack.org/128386 | 18:13 |
| *** jcoufal has joined #tripleo | 18:13 | |
| larsks | sdake: I just wrote a python script that does json.load() followed by yaml.dump(). | 18:13 |
| sdake | ya I was going to do that but ELAZY | 18:13 |
| *** athomas has joined #tripleo | 18:15 | |
| *** athomas has quit IRC | 18:15 | |
| *** athomas has joined #tripleo | 18:16 | |
| SpamapS | larsks: that is exactly what flamel does | 18:16 |
| SpamapS | larsks: but pip installable. ;) | 18:16 |
| larsks | SpamapS: But...so...short.... :) | 18:17 |
| SpamapS | clint@clint-HP:~/src/flamel$ wc -l flamel | 18:17 |
| SpamapS | 81 flamel | 18:17 |
| sdake | original machine name ;-) | 18:17 |
| SpamapS | 54 without comments | 18:17 |
| SpamapS | sdake: I gave up naming my computers years ago | 18:18 |
| SpamapS | makes one too attached | 18:18 |
| SpamapS | laptops are cattle too | 18:18 |
| SpamapS | I just brand it with the clint-ranch | 18:18 |
| SpamapS | larsks: note that this also does some nice things to make the yaml pretty :) | 18:19 |
| SpamapS | thanks to mordred | 18:19 |
| mordred | yeah - it keeps key sort order, for one | 18:20 |
| mordred | so if the document was one that people actually interact with, like default_data.json in the stackalytics repo - a dump(load()) will make the result pretty unreadable | 18:20 |
| mordred | you know "undefined sort order" blah blah | 18:20 |
| openstackgerrit | Steven Dake proposed a change to stackforge/kolla: Change nova compute replication to from json to yaml https://review.openstack.org/128390 | 18:21 |
| *** michchap has quit IRC | 18:21 | |
| openstackgerrit | A change was merged to stackforge/kolla: Convert nova-compute-replication to yaml https://review.openstack.org/128386 | 18:23 |
| *** morazi has quit IRC | 18:28 | |
| *** jprovazn has quit IRC | 18:28 | |
| *** morazi has joined #tripleo | 18:28 | |
| *** akrivoka has quit IRC | 18:28 | |
| *** jtomasek has joined #tripleo | 18:31 | |
| *** akrivoka has joined #tripleo | 18:33 | |
| SpamapS | mordred: flamel 0.0.3 uploaded.. can operate on HOT templates now :) | 18:36 |
| mordred | nice! | 18:36 |
| *** sdake_ has joined #tripleo | 18:40 | |
| *** sdake_ has quit IRC | 18:40 | |
| *** sdake_ has joined #tripleo | 18:40 | |
| openstackgerrit | Daneyon Hansen proposed a change to stackforge/kolla: Fixes Keystone Public URL https://review.openstack.org/128402 | 18:41 |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: add better validation checks (part 1) https://review.openstack.org/128345 | 18:43 |
| sdake_ | larsks can you start a thread re daneyon's change above and what the meaning of the 3 keystone urls are | 18:49 |
| sdake_ | and tag it [keystone] so we can get to th ebottom of the correct answer | 18:49 |
| *** radez_g0n3 is now known as radez | 18:49 | |
| larsks | daneyon: Actually, are you around here right now? | 18:50 |
| sdake_ | i'm still not convinced public_IP is correct | 18:51 |
| larsks | Here's the thing: | 18:51 |
| larsks | *none* of the values are correct right now for use outside of the cluster. | 18:51 |
| sdake_ | agree | 18:51 |
| larsks | Not a single one. It doesn't matter what we put there. | 18:51 |
| sdake_ | ya - that part sucks | 18:52 |
| sdake_ | but we accept that as our initial milestone 1 prototype | 18:52 |
| larsks | We need an external visible VIP, and internally visible ip that is moves from host host, and a way to reliably connect that to a particular service. | 18:52 |
| larsks | In any case, fixing this here in the keystone start.sh is the wrong place. | 18:52 |
| larsks | If it needs to be changed there, then also for all other services. | 18:52 |
| sdake_ | i have seen where the ips are x.y.3.z and x.y.4.z and they dont route | 18:53 |
| sdake_ | although I could hav ebeen in error in my analsysi at that time | 18:53 |
| larsks | If someone can present to me a system that doesn't work I would be curious to take a look. So far, all my tests have been successful. | 18:54 |
| *** akrivoka has quit IRC | 18:54 | |
| sdake_ | ya my system last night was rebooted after the .3 and .4 scenario | 18:54 |
| sdake_ | before you studied it | 18:54 |
| sdake_ | and we got to th ebottom of the problem, that keystone crds must be used on the minions | 18:54 |
| sdake_ | and the clients must be used on the minions | 18:54 |
| larsks | sdake_: Well, we demonstrated successfully that you can run the clients anywhere...ie. we were able to run them on "minimi" and have them work correctly. | 18:55 |
| larsks | Err, "minime" | 18:55 |
| sdake_ | its a minidell, had to call it something mini :) | 18:55 |
| sdake_ | versus my hp z820 which is "bigiron" ;-) | 18:55 |
| larsks | Yeah. Point being, we were able to access openstack services from outside of kubernetes without errors. | 18:56 |
| sdake_ | yes with same public ip subnet | 18:56 |
| sdake_ | not sure about different subnet tho | 18:56 |
| *** rhallisey has quit IRC | 18:56 | |
| *** cody-somerville has joined #tripleo | 18:57 | |
| sdake_ | spamaps do they give you hp z820s to work on? | 18:57 |
| *** ramishra has joined #tripleo | 18:57 | |
| openstackgerrit | A change was merged to openstack/tuskar-ui: Added arch info https://review.openstack.org/127986 | 18:58 |
| SpamapS | sdake_: I can get one if I want, yes. I prefer to keep things lightweight.. have an EliteBook and thats it. | 19:00 |
| openstackgerrit | A change was merged to openstack/tuskar-ui: Fix the plan validation https://review.openstack.org/128228 | 19:01 |
| openstackgerrit | A change was merged to openstack/tuskar-ui: Remove the buttons from modal templates https://review.openstack.org/128256 | 19:01 |
| *** sdake_ has quit IRC | 19:02 | |
| *** ramishra has quit IRC | 19:03 | |
| openstackgerrit | patrick-crews proposed a change to openstack/tripleo-image-elements: Add missing packages to tempest element https://review.openstack.org/127055 | 19:03 |
| *** jtomasek has quit IRC | 19:04 | |
| sdake | z820 rocks | 19:04 |
| sdake | wish I had more then 16gb ram | 19:05 |
| sdake | fortunately I have 4 slots left for memory | 19:05 |
| sdake | and 8 slots on my other cpu that I dont have :( | 19:05 |
| lifeless | mornink | 19:08 |
| openstackgerrit | A change was merged to stackforge/kolla: Config will be done by the controller https://review.openstack.org/128261 | 19:20 |
| lifeless | ah, its the off week for meetings ? | 19:20 |
| openstackgerrit | Lars Kellogg-Stedman proposed a change to stackforge/kolla: add better validation checks (part 1) https://review.openstack.org/128345 | 19:21 |
| openstackgerrit | A change was merged to stackforge/kolla: Nova-controller needs a start script https://review.openstack.org/128273 | 19:21 |
| bnemec | I hope so. Otherwise I'm totally blowing it off right now. :-) | 19:21 |
| *** rhallisey has joined #tripleo | 19:29 | |
| jpeeler | sdake: getting errors on mariadb. to debug, docker run -t kollaglue/fedora-rdo-mariadb /bin/bash ? | 19:29 |
| sdake | jpeeler I have never shelled into a docker container - what d o the logs say? | 19:30 |
| sdake | sudo docker logs XYZ | 19:30 |
| jpeeler | ls: cannot open directory /var/lib/mysql: Permission denied | 19:31 |
| larsks | jpeeler: did you see my message earlier about selinux? | 19:31 |
| sdake | ya that is selinux | 19:31 |
| jpeeler | larsks: no | 19:31 |
| sdake | turn that off imo :) | 19:31 |
| larsks | setenforce 0 all the things! :) | 19:31 |
| sdake | I am certain that is selinux, I have seen it and debugged it prior | 19:31 |
| jpeeler | where are we talking about, on the host? | 19:31 |
| *** mrunge has quit IRC | 19:31 | |
| larsks | Yes, on the host. | 19:32 |
| sdake | on the minions | 19:32 |
| larsks | What sdake said. | 19:32 |
| sdake | too many hosts ;-) | 19:32 |
| *** jlabocki has quit IRC | 19:32 | |
| larsks | Yeah. | 19:32 |
| sdake | we have like 150 levels of hosts ftw | 19:32 |
| *** jlabocki has joined #tripleo | 19:32 | |
| daneyon | larsks: here is where neutron is using the keystone publicurl: https://github.com/danehans/kolla/blob/danehans_keystone_neutron/docker/neutron/neutron-base/config-neutron.sh#L52-L55 | 19:32 |
| jpeeler | ok let me try that | 19:32 |
| larsks | daneyon: But it's not...it's explicitly using KEYSTONE_PUBLIC_PORT_5000_TCP_ADDR. | 19:33 |
| larsks | So it's not using the service catalog at all in that example. | 19:33 |
| *** cody-somerville has quit IRC | 19:33 | |
| larsks | That variable is set by kubernetes and will point at the local service proxy for the given service (the keystone "public" api in this example). | 19:33 |
| daneyon | larsks: i am following the openstack docs for config params. I have not looked at the code to see if this should actually be the arminurl. If so, I'm happy to change. I know where you're going with the seperate IP for the publicurl. | 19:33 |
| larsks | daneyon: it should be pointing at the public *api*, yes, and you are doing it. | 19:34 |
| larsks | My point is that it isn't using the service catalog to get that address. | 19:34 |
| larsks | That line hardcodes an address into the neutron configuration. | 19:34 |
| larsks | And it's correct in its current form. | 19:34 |
| larsks | The glance images work the same way. | 19:35 |
| larsks | The service catalog, though, will ultimately need to be populated with addresses that are accesible from outside of kubernetes so that people can, e.g., run clients remotely. | 19:36 |
| larsks | In this case, both the publicURL and adminURL endpoints would use something like PUBLIC_IP, where that would probably be some load-balancer maintained vip passed in as a pod environment variable. | 19:36 |
| daneyon | larsks: so, should the public url use the KEYSTONE_PUBLIC_PORT_5000_TCP_ADDR instead of PUBLIC_IP or should I just map PUBLIC_IP to KEYSTONE_PUBLIC_PORT_5000_TCP_ADDR (similar to what I did for Glance)? | 19:37 |
| larsks | daneyon: I guess I am saying that you should stop worrying about the service catalog entries right now. | 19:37 |
| larsks | I am defaulting PUBLIC_IP to a pod address so that is accesible from elsewhere in the k8s cluster while we are developing things. | 19:38 |
| larsks | Let's juse use that, and then put PUBLIC_IP into the service catalog. | 19:38 |
| larsks | When we have a solution for actual externally visible addresses, we will pass that in as a pod environment variable and everything will be correct. | 19:38 |
| daneyon | larsks: let me try again, but I think my neutron-server container was failing because it was trying to talk to Keystone on the publicurl which was mapped to its container ip. The neutron-server container is on a different host and can not communicate with keystone's container ip. | 19:38 |
| larsks | The value you have there is correct, and has nothing to do with the "publicurl" endpoint in the keystone service catalog. There is a chance you have another networking problem that is preventing the communication from working correctly. | 19:39 |
| larsks | It is also possible that I am wrong! | 19:39 |
| larsks | Is is possible for me to log into your system and take a look? | 19:40 |
| larsks | Or are your changes published somewhere so that I can try them out locally? | 19:40 |
| larsks | sdake: https://review.openstack.org/#/c/128345/ is passing jenkins now, if you can restore your earlier +2. | 19:41 |
| openstackgerrit | A change was merged to stackforge/kolla: Start nova-controller services and pods https://review.openstack.org/128363 | 19:42 |
| jpeeler | larsks: is there a way to force scheduling on one minion? kubernetes just scheduled everything on the other minion and now i'm redownloading all the images again | 19:43 |
| *** cody-somerville has joined #tripleo | 19:43 | |
| *** cody-somerville has joined #tripleo | 19:43 | |
| larsks | jpeeler: That's not something I've look into, so not sure. | 19:44 |
| openstackgerrit | A change was merged to stackforge/kolla: add better validation checks (part 1) https://review.openstack.org/128345 | 19:44 |
| larsks | Maybe just run a single-minion kubernetes install? Might work. | 19:45 |
| *** dprince has quit IRC | 19:45 | |
| openstackgerrit | A change was merged to stackforge/kolla: nova-compute-service doesn't expose a port https://review.openstack.org/128366 | 19:45 |
| sdake | jpeeler you will need multiminion for compute likely | 19:45 |
| sdake | maybe not need, but want ;) | 19:46 |
| daneyon | larsks: Here is my neutron-server work: https://github.com/danehans/kolla/tree/neutron_server | 19:47 |
| openstackgerrit | A change was merged to stackforge/kolla: Have qemu uri point to correct place https://review.openstack.org/128337 | 19:47 |
| daneyon | larsks: I am in the process of rebuilding keystone. I'll et you know how it goes. | 19:48 |
| *** athomas has quit IRC | 19:52 | |
| daneyon | kolla dev should i be able to communicate with any of the services using any of the kube-proxy ip's? | 19:54 |
| sdake | larsks any luck on getting https://review.openstack.org/#/c/121139/2/heat/engine/cfn/functions.py merged? | 19:55 |
| larsks | sdake: It got hit by the icehouse freeze so I kind of forgot about it. I will poke it this afternoon. | 19:55 |
| larsks | daneyon: you *should* only be able to use the service proxy identifyed in the variables visible to your container. Right now, you can talk to proxies on other minions...but that will stop working very soon now. | 19:56 |
| larsks | The upstream goal is to make proxies pod-local. | 19:57 |
| *** ramishra has joined #tripleo | 19:58 | |
| daneyon | larsks: for example, i have mariadb running on host 1 and keystone running on host 2. Keystone's mariadb ENV points to the ip of host 2 and not host 1. keystone can not connect to the DB because it's using host 2's proxy IP instead of Host 1 where the mariadb container resides. | 19:59 |
| larsks | daneyon: That is exactly the situation that is expected to work, leading me to believe you may have a local environment problem. | 19:59 |
| daneyon | larsks: when i use mysql -h -u -p i can connect to mariadb on host 1's IP but not host 2's. | 19:59 |
| larsks | You expect to see the services pointing at the *local hosts* service proxy. | 19:59 |
| daneyon | larsks: OK | 20:00 |
| larsks | Right, that will work for now, but (a) that's not how you're supposed to access services, and (b) that will soon stop working. | 20:00 |
| *** killer_prince has quit IRC | 20:01 | |
| daneyon | larsks: So the local kube proxy, proxy's the traffic to the correct IP (host 1)? | 20:01 |
| larsks | daneyon: The local kube proxy proxies traffic to the correct *pod*, actually. | 20:02 |
| larsks | It connects directly to the target pods ip address. | 20:02 |
| daneyon | larsks: OK | 20:02 |
| sdake | well if this openstack thing doesn't work out ,we can all get a job doing k8s dev ;) | 20:03 |
| *** ramishra has quit IRC | 20:03 | |
| *** morazi has quit IRC | 20:05 | |
| *** morazi has joined #tripleo | 20:05 | |
| *** cody-somerville has quit IRC | 20:07 | |
| jlabocki | ping sdake and larks … I've been asking you about the mariadb stuff. Here is a summary and question in pastebin - http://fpaste.org/141893/13317285/ | 20:08 |
| *** cody-somerville has joined #tripleo | 20:08 | |
| openstackgerrit | Ben Nemec proposed a change to openstack/diskimage-builder: Consolidate lsb_release source-repositories https://review.openstack.org/128434 | 20:09 |
| larsks | jlabocki: can you get me access to that atomic host? | 20:10 |
| daneyon | larsks: Here is a gist of the issue I'm seeing: https://gist.github.com/danehans/377d1d504a07391518da When keystone is deployed to the same host as mariadb, it works. | 20:10 |
| jlabocki | larsks: I sure can | 20:11 |
| *** killer_prince has joined #tripleo | 20:11 | |
| jlabocki | I'll PM you | 20:11 |
| larsks | jlabocki: thanks. | 20:11 |
| *** killer_prince is now known as lazy_prince | 20:11 | |
| jlabocki | thnk you! | 20:12 |
| *** jcoufal has quit IRC | 20:17 | |
| *** jcoufal has joined #tripleo | 20:18 | |
| openstackgerrit | Ryan Brady proposed a change to openstack/tuskar-ui: Enable ExtraConfig Parameter in Tuskar UI https://review.openstack.org/127621 | 20:20 |
| *** jprovazn has joined #tripleo | 20:24 | |
| sdake | daneyon and when different hosts, it doesn't work unless they have the same subnet for PUBLIC_IP correct? | 20:25 |
| sdake | bbiaf time f or a break | 20:26 |
| *** TheJulia has quit IRC | 20:30 | |
| daneyon | sdake: I'm still trying to figure things out, but it appears that i can only get keystone to work when it's on the same host as mariadb. It's like the kube-proxy is not proxy'ing correctly. If I understand it correctly, I should be able to hit the host IP on 1 host and have it proxy me to the correct host where the service (ie 3306) is listening... correct? | 20:30 |
| *** jehb has quit IRC | 20:30 | |
| *** pblaho has quit IRC | 20:35 | |
| *** openstackgerrit has quit IRC | 20:36 | |
| larsks | daneyon: that's correct, that should work (and does work, locally). | 20:38 |
| larsks | https://github.com/GoogleCloudPlatform/kubernetes/blob/master/docs/networking.md | 20:38 |
| larsks | Notes about kubernetes network design ^^^^^ | 20:38 |
| *** julim has quit IRC | 20:39 | |
| daneyon | larsks: it's been a while, so let me look at the kube net design again. i updated my gist, so take a look starting at L125. I can connect to db using remote proxy but not local proxy from a functional rabbit container. | 20:42 |
| *** jprovazn has quit IRC | 20:42 | |
| larsks | daneyon: understood. Again, I think you have an environment config problem. I would like to help debug it, if you can get me access. Right now I need to be afk for a few hours (picking up kids and &c), but I will be back around later this evening. | 20:43 |
| *** jtomasek has joined #tripleo | 20:44 | |
| *** andreaf has joined #tripleo | 20:46 | |
| *** cwolferh has quit IRC | 20:49 | |
| sdake | daneyon are your public ips on different subnets? | 20:50 |
| sdake | when it works for me they are on the same subnet | 20:50 |
| *** jtomasek has quit IRC | 20:50 | |
| sdake | I'm curious is there a wayt o create a local cache of the images to speed things up a notch? | 20:51 |
| jpeeler | me too ^ | 20:53 |
| jpeeler | sdake: i think a local docker registry would have to configured | 20:53 |
| *** cwolferh has joined #tripleo | 20:53 | |
| *** ramishra has joined #tripleo | 20:59 | |
| *** untriaged-bot has joined #tripleo | 21:00 | |
| untriaged-bot | Untriaged bugs so far: | 21:00 |
| untriaged-bot | https://bugs.launchpad.net/tripleo/+bug/1380608 | 21:00 |
| uvirtbot | Launchpad bug 1380608 in tripleo "Convert block-storage-nfs.yaml to StructuredDeployment" [Undecided,Confirmed] | 21:00 |
| *** untriaged-bot has quit IRC | 21:00 | |
| *** sdake_ has joined #tripleo | 21:02 | |
| *** ramishra has quit IRC | 21:03 | |
| *** pradeep has quit IRC | 21:07 | |
| *** andreaf has quit IRC | 21:08 | |
| *** pradeep has joined #tripleo | 21:08 | |
| *** rlandy has quit IRC | 21:09 | |
| *** sdake_ has quit IRC | 21:15 | |
| *** zigo has quit IRC | 21:16 | |
| *** ChuckC has joined #tripleo | 21:19 | |
| *** pensu has quit IRC | 21:22 | |
| *** andreaf has joined #tripleo | 21:22 | |
| sdake | jpeeler - yup would need to be on the same private net as the rest of the stuff too - probably as a sparate template | 21:22 |
| sdake | that way you could kill one with affecting the other | 21:22 |
| *** morazi has quit IRC | 21:31 | |
| *** jcoufal has quit IRC | 21:32 | |
| *** noslzzp has quit IRC | 21:37 | |
| sdake | jpeeler try this http://paste.fedoraproject.org/141931/32285614 | 21:41 |
| sdake | just fired one up | 21:41 |
| sdake | jpeeler I dont think its quite right - it is getting a 10.0.0.2 address which is already in the other heat template | 21:42 |
| *** spzala has quit IRC | 21:42 | |
| jpeeler | i will at some point, devstack is broken for me currently. too many layers | 21:43 |
| sdake | I am using rdo atm | 21:43 |
| sdake | seems to work | 21:43 |
| *** jdob has quit IRC | 21:46 | |
| *** jang2 has quit IRC | 21:54 | |
| *** gfidente has quit IRC | 21:57 | |
| *** sdake_ has joined #tripleo | 21:59 | |
| *** ramishra has joined #tripleo | 22:00 | |
| *** ccrouch has quit IRC | 22:00 | |
| *** ccrouch has joined #tripleo | 22:01 | |
| *** ramishra has quit IRC | 22:05 | |
| *** ccrouch has quit IRC | 22:05 | |
| *** pensu has joined #tripleo | 22:06 | |
| *** pensu has quit IRC | 22:06 | |
| *** ccrouch has joined #tripleo | 22:12 | |
| * SpamapS builds a tripleo cloud in anger | 22:13 | |
| lifeless | I thought you were in LA? | 22:15 |
| SpamapS | lifeless: Los Angeres | 22:15 |
| cinerama | hmmm....i want to do some bashate fixups on a script i'm poking at, but how much can i do without it being considered too much on top of the goal of the patch? | 22:18 |
| lifeless | cinerama: do it as a separate changeset, then you can do as much as you want | 22:20 |
| SpamapS | cinerama: git add -p is fantastic for that. | 22:23 |
| SpamapS | one of those "OMG where has this been all my life?" moments when sdague pointed it out to me. | 22:24 |
| daneyon | sdake: I'm back, public IP's are on the same subnet. | 22:25 |
| daneyon | larsks: I can get you access when you have a moment to help. | 22:26 |
| sdake_ | daneyon hmm that works for me, your tlaking about the 251.x.y.z addresses? | 22:26 |
| sdake_ | already tuesday and no wokring base system with glance :( | 22:27 |
| sdake_ | i got it to boot earlier | 22:27 |
| sdake_ | now it bombs out | 22:27 |
| *** andreaf has quit IRC | 22:27 | |
| *** andreaf has joined #tripleo | 22:28 | |
| daneyon | sdake: I thought you were referring to neutron public's IP assigned to the kube hosts. The kbr0 IP's are different on my 2 minions (10.251.4/5.1). But this is expected though... correct? | 22:32 |
| sdake_ | when i had that setup (4/5).1 it was bust | 22:33 |
| sdake_ | when I have (4/4).1/2 it works | 22:33 |
| sdake_ | dont know why | 22:33 |
| sdake_ | i do know that the 5 and the 4 don't route to each other | 22:33 |
| sdake_ | because you can't telnet to a port on one to the other | 22:34 |
| sdake_ | give the telnet test a try | 22:34 |
| sdake_ | pretty crazy kolla is < 1500 loc yet does all it does | 22:37 |
| sdake_ | 214 commits | 22:39 |
| *** pradk has quit IRC | 22:48 | |
| SpamapS | sigh | 22:51 |
| SpamapS | Timeout: (<pip._vendor.requests.packages.urllib3.connection.VerifiedHTTPSConnection object at 0x7f0e4cf26290>, 'Connection to pypi.python.org timed out. (connect timeout=15)') | 22:51 |
| SpamapS | some day I'll learn that you just have to have a local pypi mirror | 22:51 |
| *** ramishra has joined #tripleo | 23:00 | |
| *** openstackgerrit has joined #tripleo | 23:02 | |
| *** lucas-dinner has quit IRC | 23:02 | |
| *** ramishra has quit IRC | 23:05 | |
| daneyon | larsks: I figured out the networking issue. For some reason linkmanager was running but in a corrupted state. Refresh my gist if you want to see the details. | 23:07 |
| larsks | daneyon: I suspected something like that. I thought maybe linkmanager had stopped running; I pushed an update yesterday that sets the restart policy in the systemd unit to "always". Do you have the gist url handy? I'm on a different system right now and my scrollback doesn't go far enough back. | 23:14 |
| daneyon | larsks: https://gist.github.com/danehans/377d1d504a07391518da | 23:16 |
| larsks | daneyon: so linkmanager had stopped running? | 23:17 |
| daneyon | larsks: sorry, do a refresh on the gist. It was running, but the logs looked suspicious. After restarting the service, linkmanager again showed running but the logs looked good. | 23:18 |
| larsks | daneyon: interesting. It looks like the code raised an IncompleteRead exception, which I would have expected to kill the process. Thanks for the logs. Is there any chance you can get the full traceback? | 23:21 |
| daneyon | larsks: how can i do that? | 23:21 |
| larsks | The tail of "journalctl -u linkmanager" should get it for you (so, "journalctl -u linkmanager | tail -<somenumber>", where <somenumber> is whatever you need to get to the top of the traceback) | 23:22 |
| sdake_ | what is linkmanager | 23:24 |
| larsks | sdake_: that is the component in my heat templates that maintains the vxlan overlay network | 23:25 |
| sdake_ | who is the author? | 23:25 |
| larsks | me. | 23:26 |
| sdake_ | is a vxlan overlay network not needed on bare metal? | 23:26 |
| * sdake_ a network dummy | 23:26 | |
| sdake_ | larsks I think the minions are ooming btw | 23:27 |
| larsks | *something* is needed on baremetal. Recall the kubernetes network design -- you are expected to be able to connect to any pod from any other pod, regardless of which minion you're on. | 23:27 |
| larsks | sdake_: you can select a different flavor in your local environment file. | 23:27 |
| sdake_ | ya, I'll send you a git pull - ibroke out the minions separately from the master | 23:27 |
| larsks | sdake_: Specifically, kube-proxy needs to be able to connecto pods wherever they are. THis is what the overlay network permits. | 23:27 |
| sdake_ | so what is needed on bare metal - just a router + hub? | 23:28 |
| sdake_ | hub/switch | 23:28 |
| * sdake_ living in 90s networking ;) | 23:28 | |
| SpamapS | So you're layering a single L2 segment? | 23:28 |
| larsks | No, I think you would need a similar overlay network. You could use the same tool, or something like flannel (https://coreos.com/blog/introducing-rudder/), or some other tool that is built for this purpose. | 23:29 |
| daneyon | larsks: pls refresh gist. | 23:29 |
| SpamapS | Is this because it's running on cloud vms and thus might be on separate networks? | 23:29 |
| larsks | daneyon: Fabulous! Thanks. | 23:29 |
| SpamapS | larsks: is this just to keep from needing to bridge the containers onto whatever physical network exists in the machines? | 23:30 |
| * SpamapS keeps understanding more each dumb question he asks. | 23:30 | |
| larsks | SpamapS: So, kubernetes requires that any pod can contact any other pod on any other minion. Docker doesn't set this up by default. | 23:31 |
| larsks | There are various ways to get to that point. | 23:31 |
| larsks | Doing it via an overlay network happens to be simple and not place constraints on the local networking environment. | 23:31 |
| SpamapS | Yeah. It's funny how this starts to look a lot like Neutron and nova.. ;) | 23:32 |
| larsks | Totally. | 23:32 |
| larsks | Here's the thin: | 23:32 |
| larsks | thing: | 23:32 |
| larsks | The more I work with kubernetes, the more I wish it was just the openstack container scheduling service, and could use neutron, cinder, etc. | 23:32 |
| openstackgerrit | Steve Baker proposed a change to openstack/os-collect-config: Collector which does a simple HTTP GET https://review.openstack.org/128479 | 23:32 |
| SpamapS | larsks: you probably could teach it to do that. | 23:32 |
| larsks | Because openstack has already solved many of the problems that kubernetes is still trying to iron out. | 23:32 |
| larsks | SpamapS: Maybe! :) | 23:33 |
| larsks | Or maybe the openstack container folks will come up with something. | 23:33 |
| SpamapS | I mean really, you just have to tell neutron-openvswitch-agent where your container is. | 23:33 |
| sdake_ | larsks we could shift efforts onto magnum once we get kolla in solid shape | 23:33 |
| SpamapS | Which is basically what Nova does. | 23:33 |
| * stevebaker *really* wishes gantt was a thing | 23:33 | |
| larsks | sdake_: And once somebody gets magnum in good shape, possibly. I haven't looked to see where things are in that project. | 23:34 |
| daneyon | larsks: i added the linkmanager logs for my other minion to the gist. | 23:34 |
| sdake_ | ya i have no idea if it runs or what its status is | 23:34 |
| SpamapS | stevebaker: https://www.youtube.com/watch?v=RV-p51fvYLc | 23:34 |
| SpamapS | ;) | 23:34 |
| larsks | daneyon: Super. | 23:34 |
| * SpamapS is only trying to be funny. :) | 23:34 | |
| stevebaker | heh, indeed | 23:35 |
| sdake_ | did he say wish in one hand crap in the other? | 23:35 |
| * sdake_ audio sucks | 23:35 | |
| SpamapS | with that I'll go get my kids | 23:35 |
| SpamapS | sdake_: yes | 23:36 |
| *** michchap has joined #tripleo | 23:41 | |
| *** andreaf has quit IRC | 23:41 | |
| *** andreaf has joined #tripleo | 23:42 | |
| openstackgerrit | Stephanie Miller proposed a change to openstack/tripleo-incubator: Clean up seed image file https://review.openstack.org/128481 | 23:43 |
| devananda | jang: ping | 23:46 |
| *** sdake_ has quit IRC | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!