| *** edcragg has quit IRC | 00:51 | |
| *** persia_ has quit IRC | 01:19 | |
| *** persia_ has joined #baserock | 01:20 | |
| *** persia_ has joined #baserock | 01:20 | |
| *** zoli__ has joined #baserock | 05:07 | |
| *** petefoth has quit IRC | 05:10 | |
| *** zoli__ has quit IRC | 05:17 | |
| *** zoli__ has joined #baserock | 05:19 | |
| *** persia has quit IRC | 05:44 | |
| *** persia has joined #baserock | 05:46 | |
| *** persia has joined #baserock | 05:46 | |
| *** zoli__ has quit IRC | 06:20 | |
| *** petefoth has joined #baserock | 06:24 | |
| *** zoli__ has joined #baserock | 06:40 | |
| *** ssam2 has joined #baserock | 06:44 | |
| *** ChanServ sets mode: +v ssam2 | 06:44 | |
| ssam2 | seems login on wiki.baserock.org is broken again | 07:04 |
|---|---|---|
| ssam2 | I get 'OpenID failure: provider_error: return_to 'https://wiki.baserock.org/ikiwiki.cgi?do=postsignin&oic.time=1430117543-538e09b0e39aaa10165b' not under realm 'https://*.branchable.com/'. ' from StackOverflow OpenID, and a 500 error from openid.baserock.org | 07:05 |
| *** a1exhughe5 has joined #baserock | 07:06 | |
| *** petefoth has quit IRC | 07:07 | |
| *** petefoth has joined #baserock | 07:08 | |
| *** petefoth has quit IRC | 07:09 | |
| *** petefoth has joined #baserock | 07:09 | |
| *** perryl_ has quit IRC | 07:09 | |
| *** perryl has joined #baserock | 07:10 | |
| *** Albert_ has joined #baserock | 07:19 | |
| ssam2 | i've reported the login problem to branchable | 07:22 |
| *** fay_ has quit IRC | 07:27 | |
| *** fay_ has joined #baserock | 07:27 | |
| *** rdale has joined #baserock | 07:27 | |
| perryl | hi, i'm currently running baserock in a chroot, and it seems to have filled up my filesystem; looking at where the chroot is mounted, there's about 9 directories corresponding to my chroot and i'm not sure if i am able to delete any of them | 07:58 |
| *** persia has quit IRC | 08:00 | |
| *** persia has joined #baserock | 08:00 | |
| *** persia has joined #baserock | 08:00 | |
| *** persia has quit IRC | 08:10 | |
| *** persia has joined #baserock | 08:11 | |
| *** persia__ has joined #baserock | 08:11 | |
| *** jonathanmaw has joined #baserock | 08:12 | |
| pedroalvarez | ssam2: good morning | 08:12 |
| *** persia_ has quit IRC | 08:13 | |
| ssam2 | perryl: I've noticed weirdness with how schroot manages baserock chroots, but I don't know enough about how schroot works to give you much useful advice | 08:14 |
| ssam2 | perryl: the actual data lives in /opt/baserock/chroots/ | 08:14 |
| ssam2 | you can do `manage-baserock list` and then `manage-baserock rm` any chroots you don't need any more | 08:15 |
| *** bashrc has joined #baserock | 08:20 | |
| *** bashrc has quit IRC | 08:22 | |
| *** bashrc has joined #baserock | 08:22 | |
| perryl | that and exiting/reentering my chroot seems to have solved the issue, ta! | 08:31 |
| *** CTtpollard has joined #baserock | 08:47 | |
| *** gary_perkins has joined #baserock | 08:52 | |
| *** tpollard_ has joined #baserock | 08:52 | |
| *** tiagogomes_ has joined #baserock | 08:53 | |
| *** CTtpollard has quit IRC | 08:56 | |
| *** mauricemoss_ has joined #baserock | 09:03 | |
| *** persia__ has quit IRC | 09:08 | |
| *** persia_ has joined #baserock | 09:08 | |
| *** persia_ has joined #baserock | 09:08 | |
| *** tpollard__ has joined #baserock | 09:13 | |
| *** lachlanmackenzie has joined #baserock | 09:14 | |
| *** tpollard_ has quit IRC | 09:16 | |
| *** edcragg has joined #baserock | 09:19 | |
| ssam2 | I've tried to collate everything useful that I can think of to do with upgrades and maintenance in one page on the wiki: http://wiki.baserock.org/guides/upgrades/ | 09:28 |
| ssam2 | i also simplified the content regarding upgrades in 'quick-start' and 'devel-with' to avoid duplication | 09:29 |
| *** tpollard_ has joined #baserock | 09:33 | |
| *** tpollard__ has quit IRC | 09:37 | |
| *** tpollard_ has quit IRC | 09:46 | |
| jjardon | Hi! If someone have some Free time, it would be great if you can take a look at https://gerrit.baserock.org/#/q/status:open+project:baserock/baserock/definitions+branch:master+topic:rework_cross_systems | 09:59 |
| pedroalvarez | the topic sounds scary | 10:19 |
| rjek | IRC channel logger is watching you Baserock, pedroalvarez. | 10:20 |
| pedroalvarez | -= THIS MESSAGE NOT LOGGED =- | 10:21 |
| pedroalvarez | jjardon: they look great | 10:25 |
| rjek | I can see nothing wrong with them, but is scared of the +1 button in general. | 10:25 |
| pedroalvarez | rjek: you can always make comments without +1 ing | 10:27 |
| * rjek nods | 10:28 | |
| radiofree | jjardon: can you push your 5.1 branch somewhere? i'd like to leave it building on a jetson tonight | 10:28 |
| radiofree | also maybe someone can try it on MIPS? | 10:34 |
| jjardon | Sure, it will be ready soon | 10:36 |
| pedroalvarez | and aarch64 | 10:36 |
| *** gary_perkins has quit IRC | 10:37 | |
| pedroalvarez | I probably will be able to test it on ppc64 | 10:37 |
| jjardon | I got this warning ehrn pulling a repo: "[trove.baserock.org] WARNING: Unable to rename cgitrc into place: No such file or directory" Is this something we should be worry about? | 10:39 |
| richard_maw | er, probably. I've seen a similar message when disk space was critically low, but I don't recall ever seeing ENOENT | 10:44 |
| *** gary_perkins has joined #baserock | 10:47 | |
| *** edcragg has quit IRC | 10:49 | |
| petefoth | I was looking at http://wiki.baserock.org/guides/baserock-ERPro-8/ ('Set up an EdgeRouter Pro as a Baserock MIPS development machine') and wondering whether the text after the 'The following information may be outdated.' line can be removed now | 11:09 |
| nowster | possibly | 11:26 |
| mauricemoss_ | nowster, depending on if the kernel/bsp will be integrated into the image, this can be removed. I think it doesn't apply anymore. | 11:26 |
| petefoth | I could rephrase the question. Is any of it actually wrtong or misleading, or is it something that is 'useful historical background'? I just belive the phrase 'may be outdatedf' is not v. helpful :) | 11:27 |
| mauricemoss_ | nowster, you may also want to point to your github kernel repo with the two additional patches? | 11:28 |
| nowster | mauricemoss_: indeed yes | 11:28 |
| mauricemoss_ | The references to Cpython and binutils should not be removed, as they are not upstreamed yet. | 11:30 |
| nowster | mauricemoss_: four patches now! | 11:32 |
| pedroalvarez | Looks like git.baserock.org is not performing really well | 11:36 |
| * pedroalvarez is investigating | 11:36 | |
| pedroalvarez | I can't even log in | 11:39 |
| pedroalvarez | I assume that the people might be having problems pulling and pushing | 11:40 |
| pedroalvarez | lorry controller also looks stuck: http://git.baserock.org/lc-status.html | 11:40 |
| ssam2 | ouch | 11:41 |
| pedroalvarez | oh, I managed to log in finally | 11:41 |
| jjardon | pedroalvarez: I had | 11:41 |
| jjardon | (it took a long time to push a branch in gcc before) | 11:41 |
| ssam2 | could be IO usage. might be worth cloning iotop from source and running it | 11:45 |
| ssam2 | since we don't have iotop in baserock, but it looks useful for debugging this sort of issue | 11:45 |
| pedroalvarez | network looks ok | 11:46 |
| * pedroalvarez goes with iotop | 11:46 | |
| pedroalvarez | nothing seems to be hammering the IO | 11:49 |
| * pedroalvarez stops for lunch | 11:49 | |
| radiofree | is it ok for me to push a branch called baserock/mesa-10.5.4-jetson (for mesa) and baserock/drm-$version-jetson (for drm)? | 12:04 |
| radiofree | these will be in the repos for mesa and drm | 12:04 |
| richard_maw | we don't have gerrit on those, and we've trusted you in the past to push such branches, so +2 | 12:04 |
| radiofree | hmm, is patching in the chunk acceptable? the only thing i need for drm https://github.com/Gnurou/drm/commit/94e0b3ae46f9f12674133054d25566994a67f710 | 12:05 |
| radiofree | or is it always better to have a branch for this? | 12:05 |
| radiofree | i suppose if you do it in the chunk the sha1 won't actually point to the thing you've built | 12:06 |
| radiofree | bad idea, will branch | 12:06 |
| richard_maw | From a reproducibilty perspective, I'm ok with manually applying a patch in pre-configure-commands | 12:06 |
| radiofree | hmm, ok, i might do that then, means we can continue to use upstream drm | 12:08 |
| *** gary_perkins has quit IRC | 12:10 | |
| *** gary_perkins has joined #baserock | 12:10 | |
| *** gary_perkins has joined #baserock | 12:11 | |
| *** edcragg has joined #baserock | 12:15 | |
| *** a1exhughe5 has quit IRC | 12:26 | |
| *** edcragg has quit IRC | 12:29 | |
| *** a1exhughe5 has joined #baserock | 12:34 | |
| *** edcragg has joined #baserock | 12:35 | |
| *** paulwaters_ has quit IRC | 12:37 | |
| jonathanmaw | hi guys, I've just sent a branch to gerrit to fix systemd user sessions. It comes in 2 commits, the first one "Fix linux-pam" is unfortunately rather large, but most of that is inlining pam config into post-install-commands | 12:38 |
| * jonathanmaw looks to see if there are any open patches he can help with | 12:39 | |
| richard_maw | jonathanmaw: quick reviews are aided by posting a link to the patch you want people to review | 12:41 |
| jonathanmaw | richard_maw: noted, but I'm not pressured by enough immediacy that I want to distract people from important stuff | 12:42 |
| *** paulw has joined #baserock | 12:51 | |
| *** a1exhughe5 has quit IRC | 12:54 | |
| pedroalvarez | hm.. looks like git.b.o doesn't have IO problems, I tested ` dd if=/dev/zero of=/root/tmp/testfile bs=1G count=1` with a result of 267.5MB/s | 12:58 |
| pedroalvarez | is that ok to say that IO is ok? | 12:58 |
| pedroalvarez | s/ok/enough/ | 12:58 |
| Kinnison | pedroalvarez: well it means bulk IO is good | 13:01 |
| Kinnison | pedroalvarez: it doesn't tell you anything about the performance when you issue many smaller IOps | 13:01 |
| bjdooks | you might want to try something like bonnie | 13:01 |
| * pedroalvarez googles | 13:02 | |
| radiofree | hi, i'm trying to do http://fpaste.org/215796/14000514/ in a chunk | 13:07 |
| radiofree | but it's failing with http://fpaste.org/215797/01400701/ | 13:07 |
| radiofree | it seems to be getting as far as /a and then giving up with sh: syntax error: unterminated quoted string | 13:08 |
| radiofree | however if i do sh -c 'sed -i "/NOUVEAU_GEM_DOMAIN_MAPPABLE (1 << 3)/a #define NOUVEAU_GEM_DO | 13:08 |
| radiofree | MAIN_COHERENT (1 << 4)" include/drm/nouveau_drm.h' it works | 13:08 |
| radiofree | (manually, in the failed build directory) | 13:08 |
| radiofree | and obviously not multi-line | 13:08 |
| Kinnison | radiofree: probably YAML eating everything after the # as a comment | 13:09 |
| Kinnison | radiofree: because it doesn't honour shell quoting | 13:09 |
| *** a1exhughe5 has joined #baserock | 13:09 | |
| ssam2 | if that's true, you can do something like | 13:10 |
| ssam2 | pre-configure-commands: | 13:10 |
| ssam2 | - | | 13:10 |
| radiofree | worked if got rid of the space between /a and # | 13:10 |
| Kinnison | yamls tokenisation is "odd" | 13:11 |
| jonathanmaw | are there any repos which use git-fat I can use to test how the get-repo handles them? | 13:12 |
| rjek | Kinnison: What we need to do is come up with a new data structure language. One to rule them all. | 13:13 |
| ssam2 | jonathanmaw: good thinking, but I don't think we use that feature at all on git.baserock.org | 13:13 |
| Kinnison | rjek: Oh Gods no | 13:13 |
| rjek | >:) | 13:13 |
| ssam2 | lisp | 13:13 |
| petefoth | rjek: will it be a 'standard'? https://xkcd.com/927/ | 13:14 |
| rjek | Obv. | 13:14 |
| jonathanmaw | ssam2: hrm, get-repo has code for it, but it looks like we have no way of testing whether it's correct or works well. | 13:14 |
| jonathanmaw | given that "Add a test for `morph get-repo`" don't test git-fat behaviour, either. | 13:15 |
| jjardon | Ive built a more complete system before submit the GCC 5.1 patches (the weston-system-x86_64-generic): https://usercontent.irccloud-cdn.com/file/iDA6CgiT/Screenshot+from+2015-04-27+11-52-13.png | 13:16 |
| jjardon | patches here: https://gerrit.baserock.org/#/q/status:open+project:baserock/baserock/definitions+branch:master+topic:gcc_5_1_0 | 13:17 |
| jjardon | radiofree: ^ | 13:17 |
| radiofree | hmm i can't use tabs in my yaml, even after - | ? | 13:20 |
| ssam2 | jonathanmaw: you could set up a repo on our internal ct-mcr-1 trove, add a binary to it, and then test | 13:21 |
| ssam2 | i'm pretty confident that it'd work, given the code used is the same as in `morph edit`, which works with git-fat | 13:22 |
| ssam2 | but i am by no way certain, so I appreciate you testing :) | 13:23 |
| radiofree | i changed the tabs to "\t\t" and now i get +ttif (bo->flags & NOUVEAU_BO_COHERENT) | 13:23 |
| radiofree | it just inserts a t | 13:23 |
| Kinnison | double the backslash perhaps | 13:24 |
| Kinnison | quoting for funsies | 13:24 |
| * jonathanmaw notes that none of morph actually has tests that it's supporting git-fat correctly. | 13:24 | |
| pedroalvarez | Hm... so is 23.4MB/s bad IO? | 13:24 |
| * jonathanmaw dives into git-fat | 13:24 | |
| radiofree | double backslash ends up with +\t\tif (bo->flags & NOUVEAU_BO_COHERENT).. heh | 13:25 |
| Kinnison | :( | 13:25 |
| * rjek suddenly recalls the nightmare of Expect, tcl, sh, and sed all at the same time. The escaping becomes... convoluted. | 13:26 | |
| Kinnison | After the joys of writing makefiles which write makefiles via shell which contain quoted shell commands, I no longer fear quoting | 13:31 |
| *** fay_ has quit IRC | 13:41 | |
| ratmice__ | hmm, slow url it seems, but one of the most ridiculous cases of escaping i've ever seen | 13:50 |
| ratmice__ | http://ynform.org/w/Pub/TheTroubleWithOrigin | 13:50 |
| rjek | Escaping like a boss: LDFLAGS="${LDFLAGS//\$ORIGIN/\\\\\\\\\\\\\\\\\\\$\$\\\$\$\\\\\\\$\$\\\$\$\\\\\\\\\\\\\\\$\$\\\$\$\\\\\\\$\$\\\$\$ORIGIN}" | 13:54 |
| *** fay_ has joined #baserock | 13:58 | |
| richard_maw | rjek: uuuuuuuuuuuuuuuuuuuuuuuuuuuuugggggggggggggggggggggggggghhhhhhhhhhhhhhhhhhhhhhhhhhh | 14:18 |
| rjek | heh | 14:20 |
| richard_maw | I remember that one coming up in the libreoffice refactoring, as part of porting it to a saner build-system | 14:26 |
| tiagogomes_ | huh!? 2015-04-27 14:32:39 [systems/openstack-system-x86_64.morph][release]Renaming networkd file from systemd chunk: /src/tmp/deployments/tmp4flA1s/tmpr35wqs/etc/systemd/network/10-dhcp.network to /src/tmp/deployments/tmp4flA1s/tmpr35wqs/etc/systemd/network/10-dhcp.network.morph | 14:35 |
| tiagogomes_ | /tmp/tmpEs2zDq: line 57: : Permission denied | 14:35 |
| tiagogomes_ | is this related to the systemd upgrade/ | 14:38 |
| tiagogomes_ | s@/@? | 14:38 |
| jjardon | tiagogomes_: that looks like a bug in the simple-network extension. What version of morph are you using? | 14:39 |
| tiagogomes_ | ab92ce0343b838b336313f604ea035a60dfcb960 | 14:40 |
| tlsa | what the artifacts field for, in the system morphology? the minimal system has them for each stratum, but the build system does not | 14:44 |
| jjardon | tiagogomes_: are you using NETWORK_CONFIG in your cluster file? what it contains? | 14:45 |
| radiofree | if i do "name: mesa@vm" in the stratum, does the morph file have to be "mesa@vm" as well? | 14:46 |
| radiofree | or can it use one called "mesa" | 14:46 |
| Kinnison | radiofree: I dislike '@' in names | 14:47 |
| Kinnison | radiofree: please avoid characters which might need special escaping in various circumstances | 14:47 |
| radiofree | i dislike it as well, but it's what was decided on a while ago (e.g u-boot@jetson) | 14:47 |
| Kinnison | eww | 14:48 |
| Kinnison | Oh dear | 14:48 |
| Kinnison | Well what scheme does that follow for the .morph naming? | 14:48 |
| tiagogomes_ | jjardon, I am not using network_config | 14:48 |
| radiofree | u-boot@jetson.morph | 14:49 |
| radiofree | however i can understand that, because you want to use different configure commands there | 14:49 |
| radiofree | in this case mesa and mesa@vm can use identical morph files | 14:49 |
| Kinnison | symlink? | 14:50 |
| Kinnison | except for name: | 14:50 |
| Kinnison | hmm | 14:50 |
| Kinnison | icky | 14:50 |
| tlsa | http://wiki.baserock.org/definitions/current/ <-- does not mention artifacts in the example system | 14:50 |
| radiofree | was the @foo introduced because of morph edit? | 14:50 |
| jjardon | tiagogomes_: simple-network in your system then? | 14:50 |
| SotK_ | radiofree: I think so | 14:50 |
| SotK_ | or more because we don't want duplicate names | 14:50 |
| Kinnison | tlsa: That is a massively incomplete document with at least one known lie in it | 14:50 |
| Kinnison | tlsa: But noone has had time to update/fix it | 14:50 |
| tlsa | ok | 14:51 |
| radiofree | if i call it mesa@vm and then use mesa.morph for the morph file i guess i'll run across that "name doesn't match" bug | 14:51 |
| tlsa | well does anyone know why the minimal system specifies artifacts for each statum? | 14:51 |
| radiofree | duplicate the morph file it is then | 14:51 |
| Kinnison | tlsa: Because it's selecting only specific products from the build | 14:52 |
| Kinnison | tlsa: to produce a minimal system | 14:52 |
| tiagogomes_ | jjardon, it is | 14:52 |
| tlsa | oh, ok | 14:52 |
| SotK_ | tlsa: strata are split into multiple artifacts just like chunks are | 14:52 |
| SotK_ | normally they are all included, but the minimal system only wants some of them | 14:53 |
| *** a1exhughe5 has quit IRC | 14:55 | |
| *** petefoth has quit IRC | 14:57 | |
| jjardon | tiagogomes_: just realized /tmp/tmpEs2zDq is not the same directory as /src/tmp/deployments/tmp4flA1s/tmpr35wqs/etc/systemd/network/ . Can you add a log file in the end of the execution of morphlib/exts/simple-network.configure to check if its other extension the one is failing? | 15:00 |
| rdale | has anyone had problems with ld.so.conf not being installed in /etc. glibc.morph has a command to install it, but it isn't ending up in my final image | 15:01 |
| *** a1exhughe5 has joined #baserock | 15:04 | |
| radiofree | rdale: i haven't come across any problems, though i haven't built master in a while | 15:05 |
| rjek | Do I remember rightly that we have a CPAN importer like we do gem importer? Or am I delusional? | 15:07 |
| * rjek supposes both could be true. | 15:07 | |
| Kinnison | I know we have a 'build-system' which understands Makefile.PL | 15:08 |
| * rjek is planning to look into how much work SpamAssassin will be at some point soon. | 15:09 | |
| rjek | (As well as ClamAV and an exim4-daemon-heavy-like thing.) | 15:09 |
| Kinnison | meep | 15:09 |
| rjek | Quite. | 15:10 |
| *** edcragg has quit IRC | 15:11 | |
| *** zoli__ has quit IRC | 15:15 | |
| tlsa | why's busybox in build-essential? If you want to have busybox, you need gcc, etc? | 15:16 |
| Kinnison | It's there because it provides a bunch of utilities generally needed for building stuff | 15:16 |
| Kinnison | e.g. cp, cat, mv, rm etc | 15:16 |
| tlsa | but if you want to make a system that has busybox, but not gcc, make, etc, what do you do? | 15:17 |
| Kinnison | You build build-essential | 15:17 |
| Kinnison | and then use the artifact splitting rules | 15:17 |
| Kinnison | a'la minimal-system | 15:17 |
| tlsa | where are the splitting rules defined? | 15:18 |
| Kinnison | That is a very good question | 15:18 |
| Kinnison | there are some default rules built into morph | 15:18 |
| Kinnison | then there are rules in the chunk morphologies | 15:18 |
| Kinnison | and then there are rules in the strata | 15:18 |
| tlsa | ah, build-essential.morph has a products section | 15:19 |
| tlsa | http://paste.baserock.org/efarisawiq | 15:19 |
| tlsa | so it looks like there is already a way to do what I want :) | 15:20 |
| jjardon | pedroalvarez: are you ok with the new patch in https://gerrit.baserock.org/#/c/404/ ? | 15:21 |
| *** edcragg has joined #baserock | 15:22 | |
| pedroalvarez | jjardon: veto removed | 15:23 |
| jjardon | cheers | 15:24 |
| rdale | ah, i've found out why i don't have /etc/ld.so.conf in my minimal system. it is because it used to be part of fhs-dirs, but i moved it to glibc instead, and glibc isn't part of the build-essential-minimal products | 15:25 |
| richard_maw | rdale: some of glibc's artifacts are, if ld.so.conf is in an artifact that isn't included, this should be fixed | 15:25 |
| rdale | i'm not sure that i fully understand the artifact splitting rules | 15:27 |
| richard_maw | The design is unintuitive. There's some built-in rules to match the names of chunk artifacts, you can explicitly assign a chunk artifact to a stratum artifact, and you can define your own name matching rules. | 15:30 |
| * richard_maw wishes he had time to go back and re-do how it worked to allow split rules to be in a different file or something | 15:30 | |
| pdar | Im using a bserock openstack in onenode. When generating an instance the | 15:31 |
| pdar | eth0 ip doesnt get set | 15:31 |
| pdar | has anyone seen this? | 15:31 |
| richard_maw | pdar: by instance do you mean a booted VM or something else? | 15:32 |
| rdale | richard_maw: build-essential-minimal only has glib-nss, and i would have thought it needs a lot more than that, over and above the missing ld.so.conf | 15:32 |
| pdar | richard_maw: booted vm | 15:32 |
| richard_maw | pdar: what kind of virtual machine are you running? | 15:33 |
| richard_maw | rdale: what leads you to say that? | 15:34 |
| rdale | doesn't it need the glibc libraries, even for a minimal system | 15:34 |
| pdar | I have a bserock openstack on baremetal and Im dpeloying a baserock image to the openstack and booting that | 15:34 |
| richard_maw | rdale: I mean what leads you to say that it's only including glibc-nss? | 15:35 |
| * radiofree -1's jjardon's CI patch | 15:36 | |
| richard_maw | rdale: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/strata/build-essential.morph#n263 has some extra assignments | 15:36 |
| richard_maw | glibc-libs should be being included in the build-essential-minimal artifact | 15:36 |
| richard_maw | (I'm aware that glibc-nss appears to be assigned twice, I need to find time to resolve that confusion) | 15:37 |
| rdale | ah ok, i didn't spot that | 15:37 |
| tlsa | is the cloudinit-support stratum needed for systems that can do cloud-init, or is it just needed to be able to deploy systems that can do cloud init? | 15:37 |
| straycat | jonathanmaw, does https://gerrit.baserock.org/#/c/260/ look okay to you now? | 15:40 |
| jjardon | radiofree: thanks for the review, patch updated | 15:40 |
| richard_maw | tlsa: I think it needs to be included in systems that can do cloud-init | 15:40 |
| jonathanmaw | yep, looks fine to me. will +1 on gerritt | 15:41 |
| tlsa | richard_maw: ok, thanks | 15:42 |
| straycat | merged | 15:42 |
| tlsa | hmm, cloud-init build depends on build-seesntial, python-core, python-common, and foundation. Is there any way to determine what its runtime deps are? | 15:43 |
| richard_maw | nope | 15:43 |
| richard_maw | you're probably best off including all its build-depends as runtime depends | 15:43 |
| tlsa | that's quite a lot | 15:44 |
| Kinnison | cloud-init is a non-trivial integration tool, it likely runtime-depends on an awful lot of stuff | 15:44 |
| tlsa | ok | 15:44 |
| tlsa | maybe I can do without cloud init | 15:44 |
| Kinnison | you only want cloud-init if your systems are to be instantiated in fleets in OpenStack or similar cloud infrastructures | 15:44 |
| tlsa | right | 15:45 |
| tlsa | ok, so, looking at the build-system-x86_64, it has both strata/bsp-x86_64-generic.morph and strata/bsp-x86_both-tools.morph (lines 15 and 41) | 15:51 |
| tlsa | http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/systems/build-system-x86_64.morph | 15:52 |
| richard_maw | wtf | 15:52 |
| tlsa | but those stratum seem to have 2 of the same chunks: | 15:52 |
| tlsa | http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/strata/bsp-x86_64-generic.morph | 15:52 |
| tlsa | http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/strata/bsp-x86_both-tools.morph | 15:52 |
| richard_maw | both-tools doesn't have a kernel | 15:52 |
| tlsa | nasm even has the same ref | 15:52 |
| tlsa | yeah | 15:52 |
| richard_maw | it doesn't need bsp-x86_both-tools | 15:53 |
| tlsa | either that or strata/bsp-x86_64-generic.morph can have 2 chunks removed? | 15:53 |
| * tlsa does not know what was intended there | 15:54 | |
| * richard_maw neither | 15:54 | |
| richard_maw | the usual approach is to have a bsp stratum which includes the kernel and bootloader | 15:54 |
| richard_maw | and only depend on one of those in the system | 15:55 |
| jjardon | related to this conversation: https://gerrit.baserock.org/#/c/325/ | 15:55 |
| richard_maw | both-tools exists because we need the bootloader in our chroots if we want to deploy a system | 15:55 |
| richard_maw | it's a bit of a mess | 15:56 |
| richard_maw | jjardon: no | 15:56 |
| * radiofree thinks a "use latest morph" script would be nice | 15:57 | |
| richard_maw | jjardon: we have different config enabled for x86_64 and x86_32, so even if we're using the same refs, we still need the strata to have different morphologies | 15:57 |
| jjardon | richard_maw: AFAIK, no when I made the patches | 15:57 |
| Kinnison | ssam2: Did you ever get a reply from branchable about openid? | 15:57 |
| jjardon | let me recheck now | 15:58 |
| radiofree | "2015-04-27 16:00:21 Warning! `strata/x-common/util-keysyms.morph' referenced in `strata/x-common.morph' does not exist" | 16:00 |
| radiofree | jjardon: ^ | 16:00 |
| *** a1exhughe5 has quit IRC | 16:01 | |
| *** zoli__ has joined #baserock | 16:02 | |
| radiofree | what should i do there, remove util-keysyms from x-common? | 16:03 |
| radiofree | your patch http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/commit/?id=eb2c33af1b56b322524d058dd6f560cb21d98a0d moves util-keysysms to x-common while also deleting the util-keysyms chunk | 16:04 |
| jjardon | radiofree: yeah, patch coming | 16:05 |
| jjardon | radiofree: https://gerrit.baserock.org/466 | 16:06 |
| radiofree | +1ed | 16:06 |
| ssam2 | Kinnison: not yet, no | 16:06 |
| jjardon | radiofree: thanks, we should thanks straycat to add that warning in morph! | 16:08 |
| radiofree | yep! thanks straycat! | 16:08 |
| * radiofree begins the great fetchening | 16:09 | |
| *** edcragg has quit IRC | 16:15 | |
| radiofree | is git.baserock.org ok? | 16:15 |
| radiofree | `ssh git@git.baserock.org whoami` seems to be hanging | 16:15 |
| pedroalvarez | I've been trying today to figure out what is going on with g.b.o | 16:16 |
| pedroalvarez | (without rebooting it) | 16:16 |
| jjardon | richard_maw: yeah, they are exactly the same | 16:18 |
| richard_maw | jjardon: doffm_ has a patch that for some very sensible reasons only needs to add the config to x86_64 | 16:18 |
| * richard_maw thought it had already been merged, since he clicked the submit button days ago | 16:19 | |
| radiofree | pedroalvarez: is ssh working for you? | 16:20 |
| pedroalvarez | radiofree: nope | 16:20 |
| radiofree | :( | 16:21 |
| pedroalvarez | is not even working from console (ssh localhost) | 16:21 |
| pedroalvarez | it just hangs | 16:21 |
| pedroalvarez | the journal has some "systemd-journald.service stop sigabort timed out. Terminating" messages | 16:22 |
| pedroalvarez | over and over again | 16:23 |
| jjardon | richard_maw: reading https://gerrit.baserock.org/#/c/443/ , seems it would not be a problem to add that config to x86_32 (only 3 lines); I think its worth taking in account all the duplication we save | 16:23 |
| richard_maw | jjardon: then we differ in opinion | 16:25 |
| *** zoli__ has quit IRC | 16:26 | |
| radiofree | pedroalvarez: it was working for me before, i pushed a couple of things | 16:27 |
| radiofree | doesn't seem to be responding at all now though | 16:27 |
| pedroalvarez | :/ | 16:27 |
| pedroalvarez | anyone against a reboot? | 16:27 |
| pedroalvarez | I'd prefer to identify the problem though | 16:28 |
| radiofree | oh well i got ssh_exchange_identification: Connection closed by remote host pretty quickly that time | 16:28 |
| *** lachlanmackenzie has quit IRC | 16:29 | |
| *** gary_perkins has quit IRC | 16:32 | |
| *** mauricemoss_ has quit IRC | 16:32 | |
| pedroalvarez | ew, other systems on the same cloud have similar problems | 16:32 |
| radiofree | pedroalvarez: i'm not against rebooting it if it solves the problem, but i don't have the authority to +1 that decision | 16:32 |
| pedroalvarez | yeah I was asking for -1's though | 16:32 |
| ssam2 | open a ticket at DataCentred as well, in case it's something they aren't aware of | 16:33 |
| radiofree | oh bugger, new mesa depends on mako | 16:33 |
| jjardon | radiofree: yeah its in openstack-services | 16:33 |
| jjardon | start to think where we should put it :) | 16:34 |
| radiofree | python-common? | 16:34 |
| jjardon | (and mako depends on python-setuptools to build, that its in python-core) | 16:34 |
| radiofree | i'm sure mesa-common ends up depending on python-common somewhere... | 16:34 |
| jjardon | radiofree: python-common should be renamed, everything there looks very cloud related (I can be wrong) | 16:35 |
| radiofree | i don't particularly want to build-depend on python-common for mesa anyway | 16:35 |
| radiofree | just for mako | 16:35 |
| radiofree | and obviously i don't want to depend on openstack-common for mesa... | 16:36 |
| pedroalvarez | ssam2: will do | 16:36 |
| jjardon | radiofree: this should be enough for now: http://paste.baserock.org/icahafeway | 16:36 |
| radiofree | jjardon: i want to get these patches merged at some point... | 16:37 |
| jjardon | Id put mako in pytyhon-core | 16:37 |
| radiofree | but mesa only wants mako from there | 16:37 |
| radiofree | i'll apply your patch for now though, since it'll be enough to test mesa/mesa-vm | 16:37 |
| jjardon | mesa want mako, and mako want python-setuptools that its in python-core anyway | 16:38 |
| jjardon | I think its only a build dependency though, so no need to change the system morphologies | 16:38 |
| radiofree | why do i need x-common? | 16:39 |
| jjardon | radiofree: sorry, thats part of my testing here: ignore that | 16:39 |
| radiofree | ok, thanks for the patch, i see you came across the same problem then | 16:40 |
| jjardon | radiofree: I think this is the best approach for a final patch: http://paste.baserock.org/isepaxeson | 16:40 |
| radiofree | and remove from openstack? | 16:41 |
| jjardon | radiofree: ups, yes, of course | 16:42 |
| *** ssam2 has quit IRC | 16:42 | |
| *** Krin has quit IRC | 16:47 | |
| *** jonathanmaw has quit IRC | 16:53 | |
| * pedroalvarez sends a couple of patches for pxeboot.write https://gerrit.baserock.org/#/q/status:open+project:baserock/baserock/definitions+branch:master+topic:baserock/pedroalvarez/pxeboot-fixes | 16:59 | |
| *** bashrc has quit IRC | 17:00 | |
| *** Albert_ has quit IRC | 17:01 | |
| richard_maw | pedroalvarez: reviewing | 17:07 |
| pedroalvarez | richard_maw: ta, the second patch might not be good enough though | 17:08 |
| pedroalvarez | hm.. gerrit is also going slow | 17:09 |
| * richard_maw ponders extending simple-network.configure to be able to generate network config for the kernel command line | 17:13 | |
| rdale | i noticed that simple-network.configure is systemd specific, and yet it is included with systems that have busybox-init | 17:16 |
| richard_maw | rdale: it is *not* systemd specific | 17:19 |
| richard_maw | rdale: it also provides /etc/network/interfaces config | 17:19 |
| rdale | ah ok | 17:19 |
| *** petefoth has joined #baserock | 17:24 | |
| *** petefoth has quit IRC | 17:34 | |
| *** zoli__ has joined #baserock | 17:43 | |
| *** petefoth has joined #baserock | 17:43 | |
| *** zoli__ has quit IRC | 18:05 | |
| *** zoli__ has joined #baserock | 18:06 | |
| *** zoli__ has quit IRC | 18:14 | |
| straycat | rjek, the import tool doesn't support cpan right now | 18:21 |
| rjek | straycat: Boo! :) | 18:24 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
