[00:12] <_mup_> ensemble/expose-provision-service-hierarchy r237 committed by jim.baker@canonical.com [00:12] <_mup_> Robust async testing for service, service unit watching in provisioning agent actions [00:14] brb, but this is nice, getting at some of the same goodness as hazmat's waiting for godot redux [00:15] no more poke_zk, no more sleep, and definitely no superpoke ;) [00:15] and the code is clean too [13:24] hmm I wonder if someone is visiting OSCON, any chance of presenting Ensemble there [14:26] Hello all === marrusl_ is now known as marrusl [15:20] <_mup_> ensemble/cli-beautification r203 committed by kapil.thangavelu@canonical.com [15:20] <_mup_> doc string cleanup [15:27] <_mup_> ensemble/trunk r229 committed by kapil.thangavelu@canonical.com [15:27] <_mup_> merge cli-beautification [r=niemeyer][f=767955] [15:27] <_mup_> Improve ensemble command output, via argparse customization, and doc string changes. [15:45] jimbaker: Your watch functions do not return *any* deferreds.. I'm not surprised you need sleeps in the code [15:48] hazmat: Morning [15:50] jimbaker: Both branches look good overall, though [15:50] jimbaker: Thanks for splitting that stuff up [15:52] SpamapS: ping [15:57] <_mup_> ensemble/expose-machine-service-units r234 committed by jim.baker@canonical.com [15:57] <_mup_> Merged trunk [16:28] niemeyer: pong! [16:32] SpamapS: Hey! [16:32] SpamapS: I'll be doing some work on the PPA today.. was just wondering if you had any input on that area [16:37] niemeyer: not really. I've been updating it whenever you guys break the AMI with whats in there now. ;) [16:37] niemeyer: let me know when you change it so I can re-run my demo/test script (hmm.. maybe we should make that part of the build..) [16:39] SpamapS: Sounds good, thanks [16:39] SpamapS: I'll send an email after I finish whatever ends up happening there [16:42] niemeyer: btw, I have an ensemble packaging branch that you can just do a 'bzr merge-upstream lp:ensemble' to update. [16:43] SpamapS: Hmm.. tell me more about it [16:43] niemeyer: lp:~clint-fewbar/ensemble/packaging [16:43] niemeyer: maybe we should put that under ~ensemble ? [16:43] SpamapS: Yeah! [16:43] SpamapS: That's awesome [16:43] niemeyer: I'll update the Vcs-Bzr: in debian/control [16:44] SpamapS: What's the plugin behind merge-upstream? [16:44] niemeyer: I think bzr-builddeb [16:46] <_mup_> ensemble/trunk r230 committed by jim.baker@canonical.com [16:46] <_mup_> merges expose-machine-service-units [r=niemeyer][f=785344] [16:46] <_mup_> Implements MachineState.get_all_service_unit_states [16:46] niemeyer: its quite clever really.. adds the ~bzr### version portion and everything [16:46] though.. I can see a problem coming where we'll have to rev the pretend upstream version (0.0.2 right now) because you guys go over 1000 commits ;) [16:47] niemeyer: ok, pushed to lp:~ensemble/ensemble/packaging [16:47] Hmm.. :) [16:47] SpamapS: Super, thanks [16:48] SpamapS: Let me research a bit about bzr-builddep... not familiar with it at all [16:48] <_mup_> ensemble/expose-provision-service-hierarchy r238 committed by jim.baker@canonical.com [16:48] <_mup_> Merged trunk [16:48] niemeyer: definitely the easiest way to maintain packaging IMO [16:48] SpamapS: Sweet.. eager to learn more. === niemeyer is now known as niemeyer_lunch [17:15] <_mup_> ensemble/expose-service-watch r234 committed by jim.baker@canonical.com [17:15] <_mup_> Addressed review comments - return result of _watch_topology, enabling removal of most pokes === deryck is now known as deryck[lunch] [17:56] <_mup_> ensemble/trunk r231 committed by jim.baker@canonical.com [17:56] <_mup_> merges expose-service-watch [r=niemeyer][f=785343] [17:56] <_mup_> Implements watches for service states and service unit states. === niemeyer_lunch is now known as niemeyer [17:59] <_mup_> ensemble/expose-provision-service-hierarchy r239 committed by jim.baker@canonical.com [17:59] <_mup_> Merged trunk [18:17] <_mup_> ensemble/expose-provision-service-hierarchy r240 committed by jim.baker@canonical.com [18:17] <_mup_> Cleanup, including removing attempt to catch InternalTopologyError at a high level, to be fixed in a separate bug fix branch [18:27] <_mup_> Bug #787701 was filed: Certain service state methods leak InternalTopologyError to using code < https://launchpad.net/bugs/787701 > [18:32] <_mup_> ensemble/expose-service-internal-topology-error r232 committed by jim.baker@canonical.com [18:32] <_mup_> Guard ServiceState.get_all_unit_states against the possibility the service is no longer present [18:53] <_mup_> ensemble/expose-service-internal-topology-error r233 committed by jim.baker@canonical.com [18:53] <_mup_> Testing outside of external branch [18:53] hazmat, any tips on txzookeeper.client.NotConnectedException issues? [18:53] jimbaker, context? [18:54] jimbaker, they come up in tests, if there is activity after the test connection closes [18:54] hazmat, sure. i'm just rerunning a large number of tests with -u against the provisioning agent. eventually i see it fail, in StateBase._read_topology [18:54] jimbaker, a particular test? [18:54] er. the same tests over again [18:55] hazmat, ahh, i see. so i just to wait on the agent shutdown properly, iwould think [18:55] that should work === deryck[lunch] is now known as deryck [18:55] it's a new set of tests for service exposing, specifically setting up and tearing down the corresponding watches [18:55] thanks, makes a lot of sense [18:59] jimbaker, as long as you have the shutdown for the watcher api protocol in stop it should be okay. [19:00] SpamapS: While going through the docs, I'm wondering how bzr-builddeb is related to the branch+packaging integration that Launchpad has been pushing [19:00] * hazmat looks over expose in trunk [19:00] hazmat, not certain what you mean by that - i have the watchers being stopped by StopWatch exceptions, and that normally works. but occasionally the connection is lost while this is proceeding it would seem [19:01] in any event, if i just add a yield on agent shutdown, it should be good [19:01] jimbaker, cool, i was thinking about the refactoring, it doesn't apply to the current expose watch api [19:02] for the state request protocol to encapsulate these apis, i think i'll tackle that next [19:03] hazmat, sounds good [19:03] all of expose's needs for watches are now in trunk [19:04] although i tried to follow the redux work, but not on the exposed flag itself since that was earlier [19:07] <_mup_> ensemble/watching-godot-redux r247 committed by kapil.thangavelu@canonical.com [19:07] <_mup_> doc string cleanup [19:11] jimbaker, the expose watch apis look good wrt to the godot stuff, thanks [19:11] hazmat, cool [19:12] hazmat, definitely nice we can run without sleeps, so much better to not have to do these annoying waits [19:12] (as i watch a -u test scroll by :) ) [19:28] <_mup_> ensemble/watching-godot-redux r248 committed by kapil.thangavelu@canonical.com [19:28] <_mup_> merge trunk and resolve conflict [19:36] bcsaller, jimbaker, hazmat: Standup? [19:37] niemeyer, sounds good [19:37] sure [19:37] sounds good, logging into skype [19:37] Ok, grabbing phones [19:39] I'm up [19:40] bcsaller, i don't see you online re skype [19:40] I can relog [19:41] bcsaller, maybe kill from the terminal.. still don't see you [19:42] hazmat: after the kill its now saying I'm already logged in, gie it a sec [19:43] SpamapS: Do you wanna join us today? [19:43] hazmat: You've held me off [19:43] niemeyer, re-dialing [19:54] http://www.10gen.com/conferences/mongodc2011 === negronjl_ is now known as negronjl [20:44] SpamapS: Have you used bzr-builder? [20:45] niemeyer: minimally [20:46] niemeyer: sorry I was at lunch [20:46] SpamapS: No worries [20:46] niemeyer: can join tomorrow tho [20:46] SpamapS: Super [21:13] hazmat: You mentioned zk 3.3.4.. is there an important fix we need from it? [21:14] (the latest stable is still 3.3.3, which is why I wonder) [21:14] niemeyer, we need a newer version than packaged 3.3.1 afaik.. i haven't bisected failures for a release, the last published one works fine though. [21:14] hazmat: Ok, the latest release stable is 3.3.3.. I'll start with that then [21:15] But first, a sandwich [21:36] % find -name '*.jar' | wc -l [21:36] 13 [21:36] *shrug* [21:39] (that's the *source* tarball, btw)