smoser | blackboxsw: sorry about that. dont know at what point in my history i did that. and then just assumed. | 01:00 |
---|---|---|
smoser | http://paste.ubuntu.com/25769370/ | 01:00 |
smoser | powersj: around ? | 01:10 |
smoser | understand whats going on at https://jenkins.ubuntu.com/server/job/cloud-init-ci/ ? | 01:10 |
smoser | https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/332446 has 2 failures reported, which i do not think are valid, and also 412 is still running . that seems odd. i figured it posted at end. | 01:11 |
powersj | smoser: nacc reported some odd runs at the end of the day. I think something is up with the system or Jenkins master is taking a dump | 01:26 |
powersj | 2017-10-19 12:06:38,215 - schema.py[WARNING]: Invalid config:\nntp.pools: {} is not of type \'array\'\nntp.servers: {} is not of type \'array\'\n | 14:28 |
powersj | based on this cloud-config: https://github.com/cloud-init/cloud-init/blob/master/tests/cloud_tests/testcases/modules/ntp.yaml | 14:30 |
powersj | 2017-10-19 12:06:14,762 - cc_lxd.py[WARNING]: lxd/bridge config must be a dictionary. found a \'<class \'NoneType\'>\'\n | 14:31 |
powersj | cased by this cloud-config: https://github.com/cloud-init/cloud-init/blob/master/tests/cloud_tests/testcases/modules/lxd_dir.yaml | 14:32 |
powersj | "If you are using the default CentOS configuration, and if you are not behind a proxy server, fastestmirror is highly recommended." | 16:32 |
powersj | well we are behind a proxy... soooo yeah enough of this | 16:32 |
rharper | powersj: we can inject a disable fastmirror | 16:43 |
powersj | sed -i s/enabled=1/enabled=0/ /etc/yum/pluginconf.d/fastestmirror.conf | 16:43 |
rharper | I did that in curtin a while back when we were testing | 16:43 |
rharper | I think there's a --disable-plugin <foo> to yum | 16:44 |
powersj | there is a block where I think we detect if we need to set the proxy | 16:44 |
rharper | if that's easier | 16:44 |
powersj | well if we only do it when we set the proxy, then it means others can keep using it rather than putting --disable.. everywhere we do a yum | 16:44 |
rharper | yeah | 16:44 |
powersj | I'll get an MP for that shortly | 16:44 |
* powersj is very fed up with those failures | 16:44 | |
powersj | https://code.launchpad.net/~powersj/cloud-init/+git/cloud-init/+merge/332511 | 16:51 |
powersj | blackboxsw: rharper ^ disables fastest mirror when using a proxy | 16:51 |
blackboxsw | I'll try disabling as well in my centos lxc to validatate here too | 16:52 |
blackboxsw | validate even | 16:52 |
blackboxsw | powersj: with enabled=0 on both my existing centos6 lxc and a fresh image w/ fastest mirror disabled, I can pull update and install any packages (epel included) without issues with the default mirrors provided. | 17:03 |
blackboxsw | we may have to tweak our preferred mirrors at some point if we find problems. but this is a step in the right direction. | 17:06 |
powersj | Yeah can try this first :) | 17:06 |
powersj | Thanks for checking | 17:06 |
blackboxsw | sorry meant to type that here. | 17:07 |
blackboxsw | agreed. I've +1'd your branch. ci fell over on it I'm just checking | 17:08 |
blackboxsw | Cannot contact torkoal: java.lang.InterruptedException | 17:09 |
blackboxsw | bah | 17:09 |
blackboxsw | Aborted by Joshua Powers | 17:09 |
blackboxsw | hrh | 17:09 |
blackboxsw | heh, who does that guy think he is | 17:09 |
powersj | yeah... | 17:11 |
blackboxsw | smoser: I'm about to merge powersj ci-script fix for no fastestmirror plugin on centos. I know it'll force us to cherrypick into ubuntu/artful your jsonschema branch instead of upstream-snapshot | 17:49 |
blackboxsw | https://code.launchpad.net/~powersj/cloud-init/+git/cloud-init/+merge/332511 | 17:49 |
blackboxsw | any concerns smoser ? | 17:49 |
smoser | blackboxsw: my concern is that my favorite mirror in japan might not get used | 17:52 |
blackboxsw | smoser: ha, and all your click-through ad revenue | 17:52 |
smoser | slightly better | 17:52 |
smoser | inside "$name" sed -i "s/enabled=1/enabled=0/" "/etc/yum/pluginconf.d/fastestmirror.conf" | 17:53 |
smoser | dont need shell there | 17:53 |
smoser | blackboxsw: powersj commented there. | 17:55 |
blackboxsw | smoser: ah right you mean drop the sh -c | 17:58 |
blackboxsw | agreed | 17:58 |
blackboxsw | can could do the same with the line above | 17:58 |
smoser | yeah, as i said there... milliseconds faster. for a package build and install of dozens of packages... | 17:58 |
smoser | the line above uses the shell for redirection | 17:59 |
blackboxsw | can't we still pass that w/out sh -c to lxc exec -- "stringcmd >> output" | 17:59 |
* blackboxsw checks | 17:59 | |
smoser | no. | 18:00 |
smoser | there is no shell involved. | 18:00 |
blackboxsw | sure enough. smoser correct +1 | 18:01 |
smoser | shell is what reads '>>' and opens 'output' for write append and then hands the filehandle to the child process as 1 | 18:01 |
* blackboxsw continues to be pleased w/ hackmd.io for doc writing | 18:03 | |
smoser | yeah, hackmd.io is the new gist | 18:03 |
smoser | they just need to have a 'commit' button | 18:04 |
smoser | then powersj can stop making fun of me for using gists and instead make fun of me for using hackio.md | 18:04 |
* powersj shakes head at smoser's "hacks" | 18:05 | |
powersj | I like that even better | 18:05 |
powersj | smoser: https://paste.ubuntu.com/25774147/ | 18:27 |
smoser | powersj: no quotes | 18:27 |
smoser | there, you're executing a command 'sed -i s/....' | 18:27 |
smoser | just like if you typed that (with the quotes) in your shell | 18:28 |
powersj | ah ok | 18:28 |
powersj | ahhh I see how it is used in other places now | 18:28 |
powersj | blackboxsw: updated https://code.launchpad.net/~powersj/cloud-init/+git/cloud-init/+merge/332511 | 18:29 |
blackboxsw | powersj: running it now | 18:30 |
smoser | i just want to make sure everyone realizes that i realize how tiny an improvement that is.. | 18:30 |
smoser | compared to the vast improvment that powersj made | 18:30 |
blackboxsw | I have a whole bag of 'microseconds' for you smoser | 18:31 |
powersj | :D | 18:31 |
smoser | you got them from not downloading dozens of megabytes from japan | 18:31 |
smoser | a penny saved is a penny earned | 18:31 |
blackboxsw | and mutiply that by 1000 CI runs | 18:32 |
blackboxsw | and you can buy a bag of candy | 18:32 |
rharper | mmm, candy | 18:39 |
blackboxsw | powersj: drop fastestmirror merged | 19:06 |
powersj | blackboxsw: thanks! | 19:10 |
blackboxsw | np created a card so we track things that we've actually landed in cloud-init (as I feel it's been too light given recent SRU work) | 19:10 |
blackboxsw | https://www.irccloud.com/pastebin/bsxEyHr8/ | 19:36 |
blackboxsw | oops sorry about that paste... shame on you irccloud | 19:36 |
blackboxsw | powersj: was there a bug associated with this? schema.py[WARNING]: Invalid config:\nntp.pools: {} is not of type \'array\'\nntp.servers: {} is not of type \'array\'\n | 19:37 |
* blackboxsw checks if it is still happening. (I believe it should still happen on master) | 19:39 | |
blackboxsw | and I'm writing a tiny MP for it | 19:39 |
powersj | blackboxsw: I have not put a bug in yet | 19:42 |
blackboxsw | powersj: don't worry 'bout it. I'll file one if needed | 19:43 |
blackboxsw | I think there's another ntp schema related bug smoser just filed. I can piggy back | 19:43 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!