| *** zoli__ has joined #baserock | 05:20 | |
| *** bwh_ has joined #baserock | 05:21 | |
| *** bwh has quit IRC | 05:23 | |
| *** paulw has joined #baserock | 05:41 | |
| *** paulw has quit IRC | 05:42 | |
| *** paulw has joined #baserock | 07:04 | |
| *** zoli__ has quit IRC | 07:26 | |
| *** mariaderidder has joined #baserock | 07:37 | |
| *** bashrc_ has joined #baserock | 08:02 | |
| *** mdunford has joined #baserock | 08:12 | |
| *** tiagogomes_ has joined #baserock | 08:17 | |
| *** gary_perkins has joined #baserock | 08:18 | |
| *** edcragg has joined #baserock | 08:22 | |
| *** zoli__ has joined #baserock | 08:39 | |
| *** paulw has quit IRC | 08:49 | |
| *** ssam2 has joined #baserock | 08:52 | |
| *** ChanServ sets mode: +v ssam2 | 08:52 | |
| *** zoli__ has quit IRC | 08:52 | |
| *** paulw has joined #baserock | 08:53 | |
| *** zoli__ has joined #baserock | 08:53 | |
| *** sambishop has joined #baserock | 09:11 | |
| ssam2 | what serves tarballs on a Trove system ? | 09:12 |
|---|---|---|
| ssam2 | i'm totally confused | 09:12 |
| ssam2 | as in, tarballs of Git repos | 09:12 |
| ssam2 | neither lighttpd nor morph-cache-server seem to do it | 09:12 |
| Kinnison | the git tars are served by the httpd | 09:14 |
| ssam2 | which ? | 09:14 |
| Kinnison | Or at least they used to be | 09:14 |
| ssam2 | there's morph-cache-httpd, git-httpd, or lorry-controller-webapp-httpd | 09:15 |
| Kinnison | erk, three? odd | 09:15 |
| ssam2 | and none of them have the path '/home/lorry/tarballs' anywhere in their config files in /etc/lighttpd | 09:15 |
| Kinnison | hrm | 09:15 |
| ssam2 | but tarball fetching definitely works... | 09:15 |
| ssam2 | they are served on port 80, so I guess it must be the git-httpd | 09:16 |
| ssam2 | but no idea how | 09:16 |
| ssam2 | ah... symlink in /var/www/htdocs | 09:17 |
| Kinnison | Makefile: ln -s /home/lorry/tarballs "${DESTDIR}/var/www/htdocs/tarballs" | 09:17 |
| Kinnison | looks like we got there at the same time | 09:17 |
| ssam2 | heh | 09:17 |
| ssam2 | thanks | 09:17 |
| straycat | is gbo running with https://gerrit.baserock.org/#/c/978/ now ? | 09:20 |
| ssam2 | no, didn't update Lorry yet | 09:20 |
| ssam2 | I can do it soon | 09:20 |
| ssam2 | straycat: lorry is updated now | 09:26 |
| straycat | awesome thanks | 09:33 |
| *** paulw has quit IRC | 10:00 | |
| *** paulw has joined #baserock | 10:03 | |
| *** tiagogomes_ has quit IRC | 10:03 | |
| ssam2 | does anyone mind if I take storyboard.baserock.org down for the length of time it takes to make a snapshot of the system image? | 10:07 |
| ssam2 | will be no more than 30 mins (hopefully less) | 10:08 |
| ssam2 | will do it at 11:30 if nobody objects befoe | 10:08 |
| ssam2 | *before | 10:08 |
| ssam2 | (10:30 UTC) | 10:08 |
| paulsherwood | ssam2: so given your answer on the list, what is RepoCacheServer actually doing? i'm confused | 10:15 |
| ssam2 | what's RepoCacheServer ? | 10:17 |
| paulsherwood | something in m-c-s, i thought? | 10:18 |
| perryl | yeah, its the class instance in morphcacheserver/repocache.py iirc | 10:18 |
| *** tiagogomes_ has joined #baserock | 10:18 | |
| ssam2 | oh, there's a RepoCache class in there | 10:18 |
| paulsherwood | what is it used for? i think ssam2 said it speeded up git somehow | 10:18 |
| ssam2 | it provides helper functions for querying certain things from a Git repo on disk | 10:19 |
| ssam2 | it basically wraps `git rev-parse`, `git ls-tree` and `git cat-file` | 10:19 |
| paulsherwood | so it's unrelated functionality... | 10:19 |
| paulsherwood | not cache at all? | 10:19 |
| ssam2 | it's used by the morph-cache-server process to provide certain API calls, which expose that information over HTTP | 10:20 |
| * paulsherwood wonders at other folks' views that his code is oddly coupled | 10:20 | |
| paulsherwood | :) | 10:20 |
| ssam2 | morph-cache-server provides the /sha1s, /files, and /trees methods which allow you to effectively run `git rev-parse`, `git ls-tree` and `git cat-file` in a remote Git repo, without having to clone it | 10:21 |
| ssam2 | and yes, it's a bit weird having that in the same file as the code which provides an interface over the artifact cache | 10:21 |
| Kinnison | I think it was an attempt to unify the code after the (frankly poor decision) to glom the code together into the morph repo | 10:22 |
| ssam2 | it was written pretty quicky, in the days when we had the idea of 'Trove' as 'server that does everything you need to use Baserock tools' | 10:22 |
| paulsherwood | ack | 10:22 |
| paulsherwood | so, back to the idea of a cache server... it shouldn't need that functionality? | 10:24 |
| paulsherwood | and do we know of use-cases where that functionality is still needed? ybd asks for trees, but copes if it can't get them | 10:24 |
| Kinnison | I believe morph's distbuild relies on the git cache server for efficiency | 10:26 |
| paulsherwood | i think this boils down to whether we expect a user to be able to trigger builds/deploys without having the relevant git repos locally (in the event that builds and deploys can happen from built artifacts) | 10:26 |
| paulsherwood | s/a user/a user or a server, eg distbuild server/ | 10:26 |
| ssam2 | I think that functionality will still prove to be useful | 10:26 |
| paulsherwood | ok | 10:26 |
| ssam2 | there are cases where you might want to work out cache keys, but not want to build | 10:27 |
| ssam2 | in which case, it's annoying to have to clone all the repos, just to resolve the tree SHAs | 10:27 |
| ssam2 | or we could use commit SHA1s in cache keys, as we discussed before, but then merge commits mean you have to rebuild things | 10:27 |
| ssam2 | I wonder if we could use cgit to get that info, though ... | 10:28 |
| ssam2 | probably more work than keeping the existing code around, for the moment | 10:28 |
| *** Kinnison has left #baserock | 10:29 | |
| *** straycat has left #baserock | 10:29 | |
| ssam2 | perhaps an ideal solution is for cgit to grow a query API, in addition to the HTML web interface it currently describes. but that requires writing a lot of Perl | 10:29 |
| ssam2 | i'm going to take a storyboard.baserock.org down for a moment | 10:31 |
| ssam2 | creating a snapshot of a 40GB image on this cloud still takes a long time, it seems | 10:57 |
| *** pedroalvarez has quit IRC | 10:59 | |
| *** pedroalvarez has joined #baserock | 11:00 | |
| *** ChanServ sets mode: +v pedroalvarez | 11:00 | |
| ssam2 | ok, storyboard.baserock.org is back | 11:06 |
| *** nowster has quit IRC | 12:26 | |
| paulsherwood | great... does this mean we'll now get latest and greatest storyboard on an ongoing basis? | 12:29 |
| paulsherwood | ssam2: i think your reasoning is correct... just calling it RepoCache is a bit confusing | 12:30 |
| ssam2 | no, it means we have a snapshot, so if someone wants to do that work, they have a starting point | 12:31 |
| pedroalvarez | hahah not "now" I guess | 12:31 |
| paulsherwood | ssam2: aha, ok :) | 12:31 |
| *** nowster has joined #baserock | 12:40 | |
| *** paulwaters_ has joined #baserock | 13:03 | |
| *** paulw has quit IRC | 13:07 | |
| paulsherwood | ssam2: do i take it your migrations are serious patches, for inclusion if reviewed +2 ? | 13:41 |
| ssam2 | indeed | 13:41 |
| paulsherwood | ssam2: ok, i've never reviewed anyone's python before (except in general terms)... i'm happy to try, but i expect there may be more problems with my reviews than the code... is that ok? | 13:42 |
| * paulsherwood at least can write working python now, but is still a n00b | 13:43 | |
| ssam2 | go for it | 13:43 |
| *** mdunford has quit IRC | 13:58 | |
| pedroalvarez | ouch: https://mason-x86-64.baserock.org/log/287240f4c8e0ed6eb79ecb151125cf2642c58e90--2015-07-17%2013:43:51.log | 14:08 |
| *** mdunford has joined #baserock | 14:13 | |
| paulsherwood | out of space? | 14:13 |
| SotK | nah, wget repo must have a submodule I think | 14:14 |
| SotK | indeed it does | 14:15 |
| pedroalvarez | indeed | 14:15 |
| paulsherwood | http://paste.baserock.org/egunifoped | 14:18 |
| paulsherwood | ybd is cheating | 14:19 |
| pedroalvarez | because 46 +21 is not 55? | 14:19 |
| paulsherwood | no.. because it's using chroot, letting the network calls out | 14:22 |
| *** paulwaters_ has quit IRC | 14:26 | |
| *** CTtpollard has quit IRC | 15:20 | |
| *** mariaderidder has quit IRC | 15:34 | |
| pedroalvarez | can anybody helpme to find out how ubuntu/debian/fedora install 'runit'? | 15:38 |
| rjek | What is runit? | 15:39 |
| pedroalvarez | "runit - a UNIX init scheme with service supervision" | 15:41 |
| rjek | apt-get install runit? | 15:42 |
| pedroalvarez | hah | 15:42 |
| pdar | pedroalvarez: so it is needed then? | 15:42 |
| pedroalvarez | my question was about how they generate the package | 15:42 |
| rjek | apt-get source runit | 15:42 |
| rjek | That'll get you a copy of the upstream sources and the patches/packaging instructions | 15:43 |
| pedroalvarez | pdar: don't know yet, just trying to do some research around it | 15:43 |
| pedroalvarez | rjek: brilliant, thanks! | 15:44 |
| *** ssam2 has quit IRC | 16:04 | |
| *** zoli__ has quit IRC | 16:12 | |
| persia | If one isn't running recent Debian (or a derivative), https://sources.debian.net/src/runit/2.1.2-3/ is another way to look at the code. | 16:14 |
| persia | I believe that everything there is also exposed as git, which may make life easier for some. | 16:15 |
| *** zoli__ has joined #baserock | 17:01 | |
| pedroalvarez | persia: thanks! | 17:05 |
| *** bashrc_ has quit IRC | 17:05 | |
| pedroalvarez | I should bookmark that page and stop wasting time looking for debian packages | 17:05 |
| *** mdunford has quit IRC | 17:12 | |
| pedroalvarez | hm.. if it's in git, I'm failing to find out the clone urls | 17:14 |
| *** gary_perkins has quit IRC | 17:25 | |
| *** edcragg has quit IRC | 17:30 | |
| persia | I might be wrong about the git: I thought I remembered someone referencing that, but can't find it now. Looking at the debsources source, I don't see any gitification. | 17:33 |
| *** paulw has joined #baserock | 20:34 | |
| *** paulw has quit IRC | 20:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!