| *** genii [~quassel@ubuntu/member/genii] has quit [Remote host closed the connection] | 00:27 | |
| *** rdale_ [~quassel@243.Red-2-139-7.dynamicIP.rima-tde.net] has joined #baserock | 03:31 | |
| *** rdale [~quassel@79.Red-79-145-105.dynamicIP.rima-tde.net] has quit [Read error: Connection reset by peer] | 03:31 | |
| *** persia_ [quassel@ubuntu/member/persia] has quit [Ping timeout: 244 seconds] | 04:36 | |
| *** persia_ [quassel@2400:8900::f03c:91ff:feae:3452] has joined #baserock | 04:38 | |
| *** persia_ [quassel@2400:8900::f03c:91ff:feae:3452] has quit [Changing host] | 04:38 | |
| *** persia_ [quassel@ubuntu/member/persia] has joined #baserock | 04:38 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 06:34 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 08:10 | |
| *** mariaderidder [~maria@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:12 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 08:33 | |
| *** mike [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:43 | |
| mike is now known as Guest29760 | 08:44 | |
| *** bashrc [~motters@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:58 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 09:00 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 09:18 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 09:27 | |
| *** Krin [~mikesmith@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 10:00 | |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 10:01 | |
| Mode #baserock +v ssam2 by ChanServ | 10:01 | |
| * ssam2 opens up Thunderbird to find 3 mails from 'Mason automated test tool' | 10:03 | |
| perryl | ssam2: the DC firehose machine is at capacity and apparently segfaulting too :( | 10:04 |
|---|---|---|
| perryl | it's like christmas come early! | 10:04 |
| ssam2 | that can only mean fun | 10:04 |
| ssam2 | I'll have a look | 10:04 |
| franred | it is segfaulting because of the capacity | 10:04 |
| franred | ssam2, vga only has 3 GB space which is completely used, so there are no more space for anything | 10:05 |
| ssam2 | right, so we need to delete stuff or expand its disk | 10:06 |
| perryl | expand disk seems preferable | 10:07 |
| perryl | depends on if 3G is reasonable | 10:08 |
| ssam2 | 3G seems very small.. | 10:08 |
| ssam2 | it has a 40GB disk according to Horizon though | 10:08 |
| ssam2 | I guess what happened is cloud-init failed to run so its root disk never got resized | 10:09 |
| perryl | ssam2: i think it's not expanded, franred mentioned something about requiring cloud-init in the cluster morphology | 10:09 |
| franred | ssam2, yes, but looks like when deploy your cluster morphology does not contain: "CLOUD_INIT: yes" | 10:09 |
| ssam2 | ah! CLOUD_INIT: yes | 10:09 |
| ssam2 | good spot. | 10:09 |
| ssam2 | I wonder if we can run it manually now. | 10:09 |
| ssam2 | oh, I ran 'btrfs filesystem resize max /' and it now has a 40GB disk | 10:10 |
| franred | ssam2, after a segfaulty disk, if you expand the disk do you think it would be ok? | 10:10 |
| ssam2 | don't know | 10:10 |
| perryl | i'll test | 10:10 |
| perryl | ERROR: unknown subcommand firehose. well, that's a promising start to the day | 10:11 |
| ssam2 | cache.baserock.org was powered off and none of us know why. It is back on now though. | 10:18 |
| paulsherwood | ouch | 10:19 |
| jmacs | pdar: Have you ever seen /usr/bin/ceph hang after returning the right result? Seems to only occur on Baserock. | 10:32 |
| pedroalvarez | jmacs: that was a bug in python IIRC | 10:33 |
| pedroalvarez | in the python version that baserock has | 10:33 |
| jmacs | hmm | 10:33 |
| perryl | ssam2: how long does a commit take to get to testgerrit? | 10:37 |
| mwilliams_ct | For one's own write extensions, what is the proper syntax to use in a cluster file in order to use them? | 10:38 |
| pdar | jmacs: pedroalvarez: Yep, a bug in python 2.7.8 | 10:39 |
| jmacs | pdar: pedroalvarez: Thanks, I'm trying to patch in 2.7.3 now and see if that helps | 10:40 |
| pdar | jmacs: there is a patch in python 2.7.9rc1 | 10:40 |
| ssam2 | perryl: should be pretty much instant, I think | 10:41 |
| ssam2 | once 'git commit' returns, you should see it in the web interface etc. | 10:41 |
| ssam2 | (if you refresh) | 10:41 |
| perryl | :( | 10:41 |
| ssam2 | wait, I mean once 'git push' returns ;) | 10:41 |
| pdar | jmacs: ooh, and python 2.7.9 just came out last week too | 10:42 |
| perryl | i ran firehose with no errors and got excited but there's nothing showing in gerrit | 10:42 |
| Krin | if anyone has the time, would it be possibl to have a quick eye at the lorry patch i have re-submitted? it's blocking a final touch i would like to make to the zookeeper patch series (that is altering where the zookeeper example programs should be obtained from) | 10:42 |
| ssam2 | perryl: can you see the output of 'git push' in a log file anywhere, to double check that the push was successful ? | 10:42 |
| ssam2 | if it looks like it was, I guess we can go through the logs on the Gerrit system and check | 10:43 |
| pdar | jmacs: it was a bug introduced in 2.7.8 too so previous python versions should work too. I think. | 10:43 |
| jmacs | 2.7.3 matches the ubuntu images I'm using, so I'm trying that first | 10:44 |
| perryl | ssam2: i tried again with --verbose and it doesn't seem to be committing or pushing that i can see | 10:44 |
| perryl | i'll look into it | 10:44 |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 10:48 | |
| perryl | ok, so without the change-id fix firehose attempts to push correctly. with it it doesn't push at all. so my way of adding a change id to the commit message is wrong clearly... | 11:19 |
| jmacs | Is there an easy way to get sgdisk in Baserock? It doesn't look like it's present in util-linux. | 11:39 |
| persia | jmacs: You can clone git clone git://git.code.sf.net/p/gptfdisk/code gptfdisk-code, and set up a local chunk definition. | 11:42 |
| persia | But if you want to publish something, you'd need to lorry, etc. | 11:43 |
| jmacs | OK, thanks persia. | 11:48 |
| persia | You may find some of the patches and/or integration code from http://sources.debian.net/src/gdisk/0.8.10-1/debian/ useful | 11:55 |
| pedroalvarez | so, the pxeboot.write extension doesn't work if the nfs-server.service is not running. I'm pondering 2 solutions: | 12:05 |
| pedroalvarez | 1) check if the service is running in the write extension and start/stop it if needed | 12:05 |
| pedroalvarez | 2) check if the service is running in the check extension, and suggest the user to strat it before calling morph deploy | 12:05 |
| persia | Consider the case where the NFS server is not running on the development laptop. | 12:06 |
| persia | You probably only want to check when the cluster configuration specifies local NFS, which may be tricky to determine, depending on how the cluster definition is written. | 12:07 |
| pedroalvarez | actually I think that is easy to determine when it's local and when not | 12:07 |
| pedroalvarez | if PXEBOOT_MODE is "existing-server", then it won't be local | 12:08 |
| persia | Oh, excellent. In the PXEBOOT_MODE:existing-server case, I think your idea of checking is a good one. | 12:09 |
| persia | Err, PXEBOOT_MODE != existing-server | 12:10 |
| pedroalvarez | I'll go for the option 2, given that I don't like the idea of starting and stopping services in a write extension | 12:10 |
| persia | That seems to follow the guideline of "do one thing, and do it well" | 12:14 |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Ping timeout: 260 seconds] | 12:16 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 12:17 | |
| *** Zara [~zarazaime@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 260 seconds] | 12:20 | |
| *** Zara_ [~zarazaime@access.ducie-dc1.codethink.co.uk] has joined #baserock | 12:20 | |
| jmacs | Looks like gptfdisk needs uuid-dev... this feels like a rabbit hole | 12:33 |
| persia | util-linux is in core, so it ought just be adjusting the chunk definition to expose it | 12:37 |
| jmacs | persia: Is that relevant to uuid/uuid.h or sgdisk itself? | 12:38 |
| persia | Debian's "uuid-dev" package comes from the util-linux source. | 12:39 |
| persia | Debian's util-linux source comes from http://userweb.kernel.org/~kzak/util-linux/ | 12:39 |
| jmacs | OK, I will look into that on the next build cycle. | 12:39 |
| *** br_logger [~ubuntu@85.199.252.189] has joined #baserock | 12:40 | |
| persia | Our util-linux lorry comes from git://git.kernel.org/pub/scm/utils/util-linux/util-linux.git | 12:40 |
| persia | I don't know, but I suspect that the two kernel.org sources are at least very similar. | 12:40 |
| pedroalvarez | I do have uuid/uuid.h in my system | 12:42 |
| pedroalvarez | /baserock/util-linux-devel.meta: "usr/include/uuid/uuid.h" | 12:43 |
| * pedroalvarez enjoys the combination of `grep` and the /baserock/* metadafa files | 12:43 | |
| persia | Ah good, so sgdisk should be simply buildable :) | 12:45 |
| persia | (I still want a better location for the metadata files: /baserock feels wrong, but everything else breaks something) | 12:46 |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 12:49 | |
| perryl | firehose finally pushes! | 13:02 |
| perryl | everything is wonderful at this exact point in time | 13:02 |
| ssam2 | hooray! | 13:03 |
| perryl | and it's only took me about half a day to get the change ID in the commit ;_; | 13:03 |
| perryl | ooh, that isn't good, mason has given a -1 | 13:04 |
| ssam2 | I hope this isn't like introducing two pets to each other and they immediately start fighting | 13:04 |
| persia | perryl: That is good. It indicates that the automation is doing things semi-blindly, and the other automation is catching the problems. | 13:05 |
| ssam2 | 'ERROR: Ref 4237d8e2a7cf5c60e699d9a8194f932136cb6fbd is an invalid reference for repo http://testgerrit.baserock.org:8080/baserock/baserock/definitions' | 13:05 |
| ssam2 | persia: sadly, it seems to indicate that the other automation is broken :( | 13:05 |
| persia | Indeed. | 13:05 |
| * perryl resists urge to pull hair out | 13:05 | |
| ssam2 | still, http://testgerrit.baserock.org:8080/#/c/73/ is about as close as we've got to having a visibile indicator of how close we are to our goal ... | 13:06 |
| persia | perryl: Baldness is the new chic :) | 13:06 |
| perryl | persia: patchy baldness may be less chic :D | 13:07 |
| persia | true | 13:07 |
| ssam2 | oh, I wonder if the problem is that `morph checkout` can't see branches that have been pushed to 'refs/for/master' in gerrit ? | 13:09 |
| ssam2 | I don't know what happens to those commits once they are pushed or how we actually get at them | 13:10 |
| persia | If you look at the change in the gerrit web UI, there should be a command explaining how to fetch the HEAD | 13:27 |
| *** robtaylor [~robtaylor@185.26.124.143] has joined #baserock | 13:31 | |
| persia | Or not :( | 13:31 |
| persia | I think you want `git fetch http://testgerrit.baserock.org/baserock/baserock/definitions refs/changes/73/73/1 && git checkout FETCH?HEAD` | 13:33 |
| persia | I also think that this should probably be https | 13:33 |
| *** ssam2_ [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:23 | |
| *** bashrc_ [~motters@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:23 | |
| *** ratmice_ [bosshog@nightfall.forlorn.net] has joined #baserock | 14:23 | |
| *** paulsher1ood [~paulsherw@access.ducie-dc1.codethink.co.uk] has joined #baserock | 14:23 | |
| *** robtaylor [~robtaylor@xvm-124-143.dc2.ghst.net] has joined #baserock | 14:27 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 14:50 | |
| ssam2_ | seems that the way Mason tries to check out candidates from Gerrit actually works fine | 14:53 |
| ssam2_ | the problem with http://testgerrit.baserock.org:8080/#/c/73/ is that it's got completely the wrong SHA1 from somewhere | 14:53 |
| ssam2_ | I shall have a go at comprehending the Mason logs | 14:53 |
| *** zoli__ [~zoli_@0x5e91887a.adsl.cybercity.dk] has joined #baserock | 14:54 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Ping timeout: 244 seconds] | 14:55 | |
| pedroalvarez | Sigh.. I'm stuck at trying to fix pxeboot.check. Its logic is a bit complex to understand... | 15:21 |
| paulsher1ood | ssam2_: does SotK's mason have a status page? | 15:35 |
| ssam2_ | paulsher1ood: it doesn't | 15:35 |
| * paulsher1ood wonders how it's supposed to work, then? | 15:36 | |
| ssam2_ | it tests every branch that gets submitted to Gerrit | 15:37 |
| ssam2_ | and sets Verified = -1 or Verified = +1 based on whether the tests pass | 15:37 |
| paulsher1ood | ok. no verification of anything post-merge? | 15:38 |
| persia | Mason v2 would check master, and report if it works or doesn't work. | 15:38 |
| ssam2_ | no, it's pre-merge only | 15:38 |
| persia | Mason v3 checks pre-merge, and only permits merge of things that work, so master is always working. | 15:38 |
| Kinnison | IIRC it prepares the merge and tests that | 15:38 |
| Kinnison | so it's a pre-merge test of what would be there post-merge IIRC | 15:38 |
| ssam2_ | it's super configurable, I'm sure we could get it testing master too if we wanted | 15:39 |
| ssam2_ | given enough time to learn how Zuul works :) | 15:39 |
| paulsher1ood | ok, that *sounds* safe... will be interesting to see in practice | 15:39 |
| Kinnison | paulsher1ood: It's a fairly standard pre-merge testing approach | 15:39 |
| ssam2_ | Personally, I think it makes sense to keep the old Mason going for now | 15:39 |
| Kinnison | paulsher1ood: I believe things like PQM etc do similarly | 15:39 |
| Kinnison | paulsher1ood: except they land commits which pass | 15:40 |
| ssam2_ | so we can see if 'master' really does always work | 15:40 |
| Kinnison | paulsher1ood: where mason will only vote initially | 15:40 |
| paulsher1ood | no possibility of folks merging stuff that Mason votes against?, or doesn't vote on at all? | 15:42 |
| jmacs | pedroalvarez: This bug with python 2.7.8 you mentioned earlier - do you know where I could find a description of the bug, or something to google for? | 15:42 |
| persia | It would show in the log, and we could be very annoyed at them. | 15:42 |
| persia | My preference would be for the gerrit ACL to disallow any pushes not by Zuul. | 15:42 |
| ssam2_ | I think we are a long way off that being useful | 15:43 |
| paulsher1ood | why, ssam2_ ? | 15:43 |
| ssam2_ | Mason only has one or two tests | 15:43 |
| persia | ssam2_: Are they not the same tests as Mason v2? | 15:44 |
| ssam2_ | persia: yes | 15:44 |
| franred | jmacs, I think pdar has more information about it, he was looking how to solve it 2 or 3 days ago | 15:44 |
| ssam2_ | persia: so they don't include all systems | 15:44 |
| pedroalvarez | jmacs: indeed, pdar was looking into fixing that | 15:44 |
| persia | That's OK. It gives us at least as much confidence as we have today, and we have a saner framework to add more tests. | 15:44 |
| ssam2_ | ok | 15:45 |
| persia | I still think we should disallow anyone other than Zuul merging, because that would open the possibility of merging something that would not pass even the minimal tests we have. | 15:45 |
| pedroalvarez | This version of mason will be doing verification pre-merge, which is awesome. So we don't have to build anything ourselves when we are reviewing a patch | 15:48 |
| persia | Well, it would save time to do so. If something doesn't build, waiting for Mason to tell you can be annoying. | 15:49 |
| ssam2_ | one problem I think we'll hit early on with this version of Mason is that it doesn't give any feedback until the test cycle is complete | 15:49 |
| paulsher1ood | neither does the current one? | 15:50 |
| persia | What sort of feedback could it provide before it finishes? | 15:50 |
| ssam2_ | I'd prefer it to give feedback in stages, so that as soon as you know it builds, it tells you "this built!" | 15:50 |
| paulsher1ood | ah, yes | 15:50 |
| ssam2_ | http://build.gnome.org/#/build/20141216.27 does this | 15:51 |
| ssam2_ | (when the build isn't failing :) | 15:51 |
| paulsher1ood | can't build be one test, and then tests be other tests? | 15:51 |
| ssam2_ | they kind of are | 15:51 |
| ssam2_ | but currently Zuul runs them one after the other, and doesn't give feedback till they are all done | 15:51 |
| paulsher1ood | that sounds like config | 15:52 |
| pedroalvarez | sounds configurable though | 15:52 |
| ssam2_ | sure | 15:52 |
| paulsher1ood | heh | 15:52 |
| ssam2_ | perhaps we could have the Verified label in Gerrit go +1 for builds, +2 for test passing | 15:52 |
| franred | sounds like a good plan | 15:53 |
| persia | Might make sense to work with zuul upstream about that: I doubt only we would find it useful. | 15:53 |
| *** zoli__ [~zoli_@0x5e91887a.adsl.cybercity.dk] has quit [Remote host closed the connection] | 16:20 | |
| paulsher1ood | Krin: i saw your branch, and then noticed it's 5 days old. haven't you made some changes since then??? | 16:23 |
| paulsher1ood | Krin: normal approach would be to push a -v2 branch with post-review fixes | 16:23 |
| Krin | huh, i was sure i'd pushed the changes i'v made today... O.o and i did not know that was the usual practice for this kind of thing paulsher1ood, i'll make that branch instead then | 16:25 |
| paulsher1ood | Krin: to be fair, i've force-pushed quite a few branches myself... but -v2 is probably safer :-) | 16:26 |
| Krin | git log | 16:27 |
| paulsher1ood | Krin: and now it seems to be uptodate. maybe my browser, caching | 16:27 |
| Krin | damnit >.< | 16:27 |
| * robtaylor force pushes his development branches | 16:27 | |
| Krin | no, sorry, that was me trying to make a V2, and pushing to the current... | 16:27 |
| robtaylor | Krin: but given you seem to have a unquie ability to make git go funny, i woudn't reccomend that for you ;) | 16:28 |
| paulsher1ood | Krin: 'un-nessisary' are you doing these on purpose? :-) | 16:28 |
| * Krin shrugs his shoulders as a 80's style TV show logo reading "thats just Krin" falls around him | 16:28 | |
| *** genii [~quassel@ubuntu/member/genii] has joined #baserock | 16:28 | |
| paulsher1ood | lol | 16:29 |
| Krin | no paulsher1ood i'm dyslexic and the command line does not have a spell checker that i'v found | 16:29 |
| paulsher1ood | aha. my apologies, then. i meant no offence | 16:29 |
| Krin | if you know of one, i'd be appreciative | 16:29 |
| Kinnison | Krin: You might want to try fish | 16:29 |
| Kinnison | Krin: It's a scary shell | 16:29 |
| Kinnison | Krin: it practically has spelling correction in | 16:29 |
| Kinnison | Krin: You could sort it out for Baserock if you like it | 16:30 |
| robtaylor | Kinnison: surely zsh has spelling correction? =) | 16:30 |
| Kinnison | robtaylor: naah, zsh is unforgiving | 16:30 |
| *** mariaderidder [~maria@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Ex-Chat] | 16:31 | |
| Krin | and no offence is taken paulsher1ood , long since come to laugh at my own mistakes :) admitting and not getting angry at them is a good way to start sorting them out. | 16:31 |
| robtaylor | setopt correct | 16:31 |
| Kinnison | robtaylor: You have to focus hard before you press <TAB> lest it decide to delete every third file whose name starts with a 'W' | 16:31 |
| robtaylor | oh yes | 16:31 |
| Krin | i'll have a look into fish Kinnison | 16:31 |
| * robtaylor still hasn't made the transition | 16:31 | |
| Kinnison | robtaylor: and yes, I know about setopt correct, I just try to never think of using it | 16:31 |
| Kinnison | Krin: fish tries to make useful suggestions based on your history, path, etc. It's quite fancy | 16:32 |
| * robtaylor likes that one of fishshell's highlighted selling points is "Glorious VGA Color" | 16:33 | |
| robtaylor | i wonder if it has sane history handling | 16:33 |
| pedroalvarez | Krin: re Zookeeper - do you understand what I am suggesting in the email? Do you need some help with that? | 16:34 |
| Krin | pedroalvarez, I may need some help doing that in an efficient manner, I understand what you are getting at and I do agree. it's a thought I'v had also, but at the same time I'm trying to get this out before year end for a client. and I'm not the greatest with baserock still. | 16:36 |
| Krin | though i think i misunderstood you in the original e-mail | 16:37 |
| Krin | but not your most recent follow up | 16:37 |
| Krin | pedroalvarez, ^^ | 16:37 |
| pedroalvarez | Krin: yeah, english is not my native language :/ | 16:40 |
| * robtaylor screams in horror at the yocto SDK he's been given for a dev board failing on his system because it thinks gnumake 4.1 is *too old* due to a bad regex | 16:43 | |
| Krin | pedroalvarez, it's far better than any other language i speak, so you win there :) | 16:45 |
| pedroalvarez | Krin: I've pushed my suggestion to baserock/pedroalvarez/zookeeper-fixes | 16:50 |
| ssam2_ | wow, Zuul is a hard project to search for on the web | 16:50 |
| pedroalvarez | Krin: hang on, I forgot one thing | 16:50 |
| jmacs | You should try "chef" | 16:50 |
| ssam2_ | at least there's only one software project called Chef, I think | 16:51 |
| Krin | i'll have a gander at that, i'v just tried changing it myself, which whent a lot faster than i thought it would, i'll see how your's and mine differ then cry over my lack of understanding of the system :) | 16:51 |
| ssam2_ | I've found two other Zuuls on github so far | 16:51 |
| ssam2_ | ah, https://github.com/openstack-infra/zuul | 16:51 |
| pedroalvarez | Krin: it should be easy to understand :) | 16:52 |
| pedroalvarez | Krin: anyway, I pushed what I was missing. That branch should be equivalent to yours | 16:53 |
| Krin | pedroalvarez, where is that branch? I can't find it | 17:03 |
| pedroalvarez | definitions.git I believe | 17:04 |
| paulsher1ood | Krin: not sure whether it matters, but i notice that system-version-manager does not appear to work when i try to deploy your zookeeper system as a devel-upgrade | 17:04 |
| paulsher1ood | 'env: can't execute '/tmp/tmp.FTWk3D/systems/TEST/orig/usr/bin/system-version-manager': No such file or directory' | 17:04 |
| genii is now known as ChristmasPresent | 17:04 | |
| pedroalvarez | paulsher1ood: that is because the zookeeper system doesn't have system-version-manager on it | 17:05 |
| persia | ssam2_: I think you want git://git.openstack.org/openstack-infra/zuul | 17:05 |
| pedroalvarez | which is not needed at all in a zookeeper system I believe | 17:05 |
| paulsher1ood | pedroalvarez: ack. just checking... so this means that zookeeper system would not be updateable? | 17:06 |
| pedroalvarez | Krin: here is the branch: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/log/?h=baserock/pedroalvarez/zookeeper-fixes | 17:06 |
| Krin | it's not a devel system, just base. this was one of the corrections made when people pointed out that i did not use any of the devel packages | 17:06 |
| paulsher1ood | Krin: base sysetms are updateable, afaik | 17:06 |
| ChristmasPresent is now known as genii | 17:06 | |
| Krin | huh ewell being abl to update sounds like a damn useful thing to be able to do, i just followed instructions to make it a not-devel system as best i could, i may well have misinterpretted something | 17:08 |
| pedroalvarez | paulsher1ood: that means that you can't "upgrade" a system to be a zookeeper system. | 17:09 |
| * Krin still feels like he is edging ever closer to actualy having this finished, even if he corrects one thing and has to correct another after, the number of corrections is slowly being wittled down | 17:09 | |
| pedroalvarez | Krin: yeah, you are getting there | 17:10 |
| paulsher1ood | pedroalvarez: true. but it would also not be possible to "upgrade" this zookeeper system in the normal baserock way | 17:10 |
| * paulsher1ood suggests adding foundation to both zookeeper systems | 17:10 | |
| persia | Is the base requirement for self-upgrading documented anywhere? | 17:10 |
| pedroalvarez | paulsher1ood: is this a requirement for all the systems? | 17:10 |
| paulsher1ood | i don't know. i'm asking | 17:11 |
| persia | That should probably be separated out in definitions, as there are lots of cases where folk would want to upgrade non-devel systems. | 17:11 |
| franred | Krin, you should also check which configuration-extensions you need, you have removed some strata from your strata so some configurations-extensions should disappear, i.e, cloud-init | 17:12 |
| Krin | I feel mighty happy now about my understanding of baserock growing, my own changes after hearing your suggestion are identical to yours pedroalvarez :) yay I'm learning | 17:12 |
| pedroalvarez | great! | 17:12 |
| Krin | good spot franred | 17:12 |
| pedroalvarez | Krin: note that I did 2 commits | 17:12 |
| paulsher1ood | any good reason not to have foundation in these systems? | 17:13 |
| pedroalvarez | the only good reason is that is not needed at all to work | 17:13 |
| Krin | ok, so i'll add foundation back into the zookeeper stratum, remove the excess configurations-extensions, and submit a -v3 | 17:13 |
| pedroalvarez | if we want it to be upgradeable, then we should add foundation to it | 17:13 |
| pedroalvarez | Krin: that works for me | 17:14 |
| Krin | then find a corner to curl up into and put ice on my brain | 17:14 |
| pedroalvarez | put bacon, so you can eat it later when is fried | 17:15 |
| Krin | mmm, brain bacon. | 17:15 |
| ssam2_ | It seems like the problem with Mason is that Zuul is creating a merge commit locally, then telling it to build that | 17:19 |
| ssam2_ | but it's giving it a repo URL in testgerrit, not anything local | 17:19 |
| ssam2_ | I guess in the case where the merge is fast-forward, this doesn't break anything, but when it's not a fast-forward merge, the commit is missing | 17:20 |
| ssam2_ | I don't know yet if Zuul is meant to push a merge commit to Gerrit (seems weird), or pass a local repo path to the tests... | 17:20 |
| ssam2_ | also the Mason instance is very unresponsive, but other machines in the same cloud are fine. And the IO and CPU usage on the Mason are both very low | 17:21 |
| persia | How do other Zuul users do it? | 17:23 |
| bashrc_ is now known as bashrc | 17:25 | |
| *** wdutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Quit] | 17:34 | |
| ssam2_ | persia: I'd like to find that out, but I'm not sure where to start looking | 17:35 |
| ssam2_ | at turbo-hipster I guess | 17:35 |
| persia | I'd ask in #openstack-infra, which I believe to be the upstream channel. I know it gets used for OpenStack, MediaWiki, and KDE. Be warned that you might get pointed at some puppet config repos :) | 17:36 |
| ssam2_ | thanks, I'll do that if I can come up with a question that makes sense | 17:36 |
| persia | You might just ask where to find the configuration, so that you can see what upstream Zuul is passing, and to what it is passing. | 17:38 |
| persia | I think most of those projects still use Jenkins behind Zuul, which may make a difference. | 17:39 |
| ssam2_ | right. I've found https://github.com/openstack-infra/zuul/blob/master/doc/source/merger.rst which will hopefully clear things up for me a bit | 17:40 |
| persia | Cool! | 17:40 |
| persia | But note the different git URL I posted earlier: the github is a non-production copy (I think it is refreshed regularly, but I don't know) | 17:40 |
| ssam2_ | ah, OK, thanks | 17:41 |
| Krin | hmm, i must admit, Kinnison suggested the deployment split between client and server to me, and i'm not sure i fully understood it. can i specify the location to deploy too? i'm trying to do that but just being told the the failure was "ssh username@HOSTNAME --true" | 17:52 |
| Krin | i have tied " morph deploy clusters/zookeeper.morph system.systems/zookeeper-server-x86_64.morph.my-server-system.location=kvm+ssh://mikesmith@..." | 17:53 |
| ssam2_ | HOSTNAME is a placeholder | 17:53 |
| Krin | ahh, but username will need manually replacing? | 17:53 |
| ssam2_ | both need manually replacing | 17:53 |
| Krin | i mean, can i specify it in the morph command, or do i need to edit the file? | 17:54 |
| ssam2_ | either | 17:54 |
| ssam2_ | but you don't need the 'system.systems/zookeeper-server-x86_64.morph.' part | 17:54 |
| Krin | that part is me trying to work out how to specify with the morph command. | 17:54 |
| ssam2_ | yes, and it was a good attempt, but it was wrong | 17:55 |
| Krin | XD | 17:55 |
| ssam2_ | try 'morph deploy clusters/zookeeper.morph my-server-system.location=kvm+ssh://mikesmith@...' | 17:55 |
| Krin | hmm, it still says that i'm using username@HOTSTNAME | 17:57 |
| Krin | would you mind if i IM'd you the full command i'm using ssam2_ ? | 17:57 |
| CTtpollard | ssam2, would it be possible to pass me a copy of projects.yaml which I believe gerrit uses? | 17:58 |
| CTtpollard | ^ ssam2_ | 17:59 |
| *** bashrc [~motters@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Lost terminal] | 18:01 | |
| ssam2_ | CTtpollard: I'm not sure I've seen a file like that in testgerrit.baserock.org | 18:03 |
| ssam2_ | I'll search | 18:03 |
| pedroalvarez | i have just tried `find` to locate it, but I failed | 18:04 |
| ssam2_ | me too | 18:04 |
| CTtpollard | hmm I'll get back to storyboard upstream, I got pointed to it as storyboard looks to be able to populate it's project list and groups from it | 18:06 |
| CTtpollard | cheers | 18:06 |
| ssam2_ | could depend on the Gerrit version | 18:06 |
| ssam2_ | we're using 2.9 I think | 18:06 |
| genii is now known as ChristmasPresent | 18:07 | |
| ChristmasPresent is now known as genii | 18:07 | |
| genii is now known as ChristmasPresent | 18:09 | |
| ChristmasPresent is now known as genii | 18:09 | |
| CTtpollard | ssam2: it could also be specific to openstack gerrit, searching brings up openstack results | 18:10 |
| *** Krin [~mikesmith@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection] | 18:17 | |
| *** flatmush [~flatmush@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving.] | 18:31 | |
| *** Guest29760 [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 18:46 | |
| ssam2_ | seems like something is very wrong when 'tar' takes 85% of the CPU for 30+ minutes to unpack an 800MB tar file | 19:08 |
| *** ssam2_ [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection] | 19:22 | |
| *** rdale [~quassel@80.30.109.96] has joined #baserock | 20:32 | |
| *** rdale_ [~quassel@243.Red-2-139-7.dynamicIP.rima-tde.net] has quit [Ping timeout: 272 seconds] | 20:32 | |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection] | 20:40 | |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 20:46 | |
| *** rdale_ [~quassel@89.Red-79-147-122.dynamicIP.rima-tde.net] has joined #baserock | 21:05 | |
| *** rdale [~quassel@80.30.109.96] has quit [Read error: Connection reset by peer] | 21:07 | |
| *** rdale_ [~quassel@89.Red-79-147-122.dynamicIP.rima-tde.net] has quit [Read error: Connection reset by peer] | 21:10 | |
| *** rdale [~quassel@111.Red-83-55-117.dynamicIP.rima-tde.net] has joined #baserock | 21:14 | |
| *** rdale_ [~quassel@247.Red-2-136-64.dynamicIP.rima-tde.net] has joined #baserock | 21:16 | |
| *** rdale__ [~quassel@122.Red-2-137-0.dynamicIP.rima-tde.net] has joined #baserock | 21:17 | |
| *** rdale [~quassel@111.Red-83-55-117.dynamicIP.rima-tde.net] has quit [Ping timeout: 258 seconds] | 21:18 | |
| *** rdale_ [~quassel@247.Red-2-136-64.dynamicIP.rima-tde.net] has quit [Ping timeout: 264 seconds] | 21:20 | |
| *** rdale [~quassel@166.Red-88-13-224.dynamicIP.rima-tde.net] has joined #baserock | 21:23 | |
| *** rdale__ [~quassel@122.Red-2-137-0.dynamicIP.rima-tde.net] has quit [Ping timeout: 264 seconds] | 21:26 | |
| *** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer] | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!