17:00:27 <jaypipes> #startmeeting
17:00:28 <openstack> Meeting started Wed Feb 15 17:00:27 2012 UTC.  The chair is jaypipes. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:00:30 <donaldngo_hp> great :)
17:00:37 <jaypipes> donaldngo_hp: again, sorry for late response
17:01:19 <jaypipes> #topic update on current bug list
17:01:34 <jaypipes> JoseSwiftQA: is Daryl around?
17:02:29 <JoseSwiftQA> jaypipes:  Yes, he should pop in shortly
17:02:34 <jaypipes> david-kranz: ah, good morning sir!
17:02:40 <jaypipes> JoseSwiftQA: awesome, thx.
17:02:42 <david-kranz> Good morning
17:02:59 <jaypipes> OK, so, let us sswitch topics until Daryl gets here...
17:03:09 <jaypipes> #topic Stress tests submitted by david-kranz and colleagues
17:03:17 <dwalleck> Hey folks
17:03:18 <jaypipes> dwalleck: mornin.
17:03:30 <jaypipes> dwalleck: we'll discuss status of bugs in a sec..
17:03:47 <jaypipes> but wanted to currently discuss david-kranz' stress test submissions right now
17:04:19 <dwalleck> okay
17:04:25 <jaypipes> david-kranz: So, there's a bit of stylistic mismatch between the Python code in the rest of Tempest and your submission.
17:04:48 <jaypipes> david-kranz: those things can be addressed pretty easily -- including License headers, cleaning up imports and location of docstrings, etc
17:04:56 <david-kranz> We can change it. Can you be more specific?
17:05:22 <jaypipes> david-kranz: yes, I'll be specific in the code review inline comments -- easier then for you to just mark "Done" on the inline comment...
17:05:53 <david-kranz> OK
17:05:55 <jaypipes> david-kranz: but overall, I think it's a good start of a stress test. I'd like to discuss the terminology used though. Let me explain my thoughts...
17:06:40 <jaypipes> david-kranz: so, the "BasherChoice" https://review.openstack.org/#patch,sidebyside,3800,1,stress/basher.py
17:07:02 <jaypipes> david-kranz: is it possible to transform that into a system that can use a config file as input?
17:07:36 <jaypipes> david-kranz: Basically, read in a "payload config" that would be used to construct a Basher object?
17:08:02 <david-kranz> Yes, but the reason we didn't want to do that was that the arguments might contain code
17:08:11 <david-kranz> Of course that could be incorprated into a config file.
17:08:42 <jaypipes> david-kranz: I *think* I understand, but could you elaborate a bit on how the arguments could contain code? What kind of code?
17:09:32 <david-kranz> The constructor arguments might be the result of some expression.
17:10:36 <jaypipes> david-kranz: could you use an example?
17:10:45 <jaypipes> david-kranz: sorry, having a tough time understand right now...
17:11:26 <david-kranz> You can see one instantiation in https://review.openstack.org/#patch,sidebyside,3800,1,stress/tests/user_script_sample.py
17:12:46 <jaypipes> david-kranz: yes, I guess I'm just wondering how the arguments read from a configuration file would be interpreted as Python code...
17:13:12 <david-kranz> Well, I am not sure either which is why we didn't do that.
17:13:33 <david-kranz> It seemed that doing this with a config file would be less flexible and/or more complicated.
17:14:06 <david-kranz> I didn't think that doing it with a config file was an important consideration at the time
17:14:28 <jaypipes> david-kranz: OK, no worries. It can be added at a later time I suppose.
17:14:54 <jaypipes> david-kranz: I just didn't want to hard-code things like in the example above where CreateVM and KillVM are exactl 50% probabliyt
17:15:10 <david-kranz> It would definitley be good to have that, at least to make it easy for simple cases
17:15:29 <david-kranz> It is not really hard coded. Each one of these files is a separate test instance.
17:16:38 <jaypipes> k
17:16:47 <david-kranz> If you wanted a spread of probabilties for this test you could code that in test instance file.
17:17:07 <jaypipes> I get you...
17:18:30 <jaypipes> david-kranz: in https://review.openstack.org/#patch,sidebyside,3800,1,stress/test_server_actions.py, is there a reason for the commented-out classes?
17:19:01 <david-kranz> Yes. The original version of the test would ssh into vms to make sure they were OK.
17:19:27 <david-kranz> While I was working on integrating the test into tempest that was nontrivial so I was going to put it back
17:19:39 <jaypipes> gotcha
17:19:43 <david-kranz> once the code was stabalized in tempest
17:20:24 <jaypipes> k
17:20:54 <jaypipes> david-kranz: couple other minor things... stuff like hardcoding the log dir (in https://review.openstack.org/#patch,sidebyside,3800,1,stress/driver.py) will need to be configurable, etc
17:21:18 <jaypipes> david-kranz: but I'll comment on that in-line on the review...
17:21:27 <david-kranz> Yes. We need to make sure this can work easily with devstack.
17:21:39 <jaypipes> anybody else have any initial input for david-kranz on the stress test patch?
17:21:40 <david-kranz> I have a pending submission to make this work with essex/keystonne
17:22:25 <jaypipes> david-kranz: this is the only other pending patch I see from you? https://review.openstack.org/#change,4147
17:23:29 <david-kranz> Yeah, I have not submitted it yet because I don't yet understand how gerrit handles updates to already posted patches
17:23:41 <david-kranz> I will include this code in addition to code review changes.
17:24:14 <jaypipes> david-kranz: depends on whether you do a git commit --amend or not...
17:24:33 <david-kranz> OK, I can handle that now.
17:25:00 <jaypipes> david-kranz: if the patch is fixing something for a patch that you've already submitted, generally you want to use --amend, unless it is a very large change and should be reviewed as a separate effort...
17:25:16 <david-kranz> OK
17:25:35 <jaypipes> alrighty... well, expect a full review from me today on the stress tests. Hopefull, dwalleck and others can also add reviews.
17:25:45 <dwalleck> certainly
17:25:46 <david-kranz> Great! Thanks.
17:25:49 <jaypipes> OK, let us move on to status reports.
17:25:58 <jaypipes> #topic Status of existing bugs
17:26:23 <jaypipes> #link http://bit.ly/w4i5td
17:27:11 <jaypipes> dwalleck: want to give a quick status report on ones you are assigned to? kinda waiting to see if Ravi can show up for an update on the ones he is looking into
17:27:16 <jaypipes> and nati..
17:27:25 <dwalleck> Lets see....
17:28:07 <dwalleck> lp 893843 is blocked for now due to an agent bug. We're looking into it, and I've updated the bug
17:28:08 <uvirtbot> Launchpad bug 893843 in tempest "Create server with personality file that overwrites an existing file" [Medium,In progress] https://launchpad.net/bugs/893843
17:28:44 <dwalleck> lp 902371 is in progress, not done
17:28:45 <uvirtbot> Launchpad bug 902371 in tempest "Write Testcases for Server Fixed IP Addresses" [Medium,Triaged] https://launchpad.net/bugs/902371
17:29:12 <jaypipes> dwalleck: for 893843, is that only affecting XenServer environments?
17:29:47 <dwalleck> jaypipes: Good point. Most likely, yes
17:30:01 <jaypipes> dwalleck: in other words, if it works on libvirt/KVM, can the test case be added with a skipIf(xen_server), etc?
17:30:02 <dwalleck> We have the test. I just can't verify that it works right now
17:30:15 <dwalleck> jaypipes: We could do that
17:30:31 <jaypipes> dwalleck: feel free to do that... then others can pull the branch and run against their installs.
17:31:03 <jaypipes> dwalleck: for the server fixed IP addresses, how do we stand?
17:31:11 <dwalleck> lp 901309 is also almost done. I want to do a bit more tweaking for readability
17:31:12 <uvirtbot> Launchpad bug 901309 in tempest "Test Case: Followings next/prev links for servers/images/flavors" [Low,In progress] https://launchpad.net/bugs/901309
17:31:49 <dwalleck> jaypipes: In progress. I hoped to be done with that before this meeting, but there's still more work to be done. We're now aiming for next Wednesday
17:33:45 <jaypipes> dwalleck: OK, no worries.
17:34:23 <dwalleck> I'm going to have to duck out folks. I'll be heads down on these issues this week
17:34:35 <jaypipes> Personally, I have one bug assigned to me: https://bugs.launchpad.net/tempest/+bug/917914
17:34:36 <uvirtbot> Launchpad bug 917914 in tempest "Emphasize compute API over nova" [Medium,Triaged]
17:34:39 <jaypipes> dwalleck: k, cya
17:34:43 <dwalleck> I'll keep my eyes on gerrit and try to keep the reviews flowing
17:34:58 <jaypipes> but the bug isn't particualrly important, so I've put it on hold for now.
17:36:47 <jaypipes> SO, doesn't look like nati or Ravi have been able to make it. I will have to follow up with them on status of bug reports personally. Expect to see a post to the mailing list about status of Tempest overall later today. For those interested, we released the TryStack project today (a project formerly known as FreeCloud) that should allow people to run tests against the OpenStack APIs without actually deploying an environme
17:36:47 <jaypipes> nt themselves. More information here: https://trystack.org
17:37:52 <jaypipes> I've not gotten on jeblair or mtaylor this past week about further testing of Tempest against the CI clusters since they've had a hellish week, but hopefully I can work with them in the next week to get Tempest jobs running consistently (even if there are failures) on commits to the core project trunks
17:38:29 <mtaylor> jaypipes: I would love to run tempest jobs
17:38:34 <jaypipes> OK, unless anyone has anything else to bring up, I'm going to end the meeting and work on a reply to donaldngo_hp's email about Tempest :)
17:38:40 <jaypipes> mtaylor: yes, me too :)
17:39:02 <jaypipes> mtaylor: I'll ping you on #openstack-dev if you have time later? Haven't wanted to bother you this past week with all the issues...
17:39:32 <mtaylor> jaypipes: yes, let's talk about it then - although we've got several things we need to get finished this week, so it'll probably be next week before we can make it actually happen
17:39:55 <jaypipes> mtaylor: yes, I was aware of your tight schedule this week, which is why I backed off ;)
17:40:35 <jaypipes> alright everyone, thank you. reminder to review david-kranz' stress test work if you have some free time.
17:40:38 <jaypipes> #endmeeting