Thursday, 2017-02-02

« Wednesday, 2017-02-01 Index Friday, 2017-02-03 »
*** cody-somerville has joined #openstack-lbaas00:12
*** fnaval has quit IRC00:20
*** ducttape_ has quit IRC00:39
*** ducttape_ has joined #openstack-lbaas00:39
*** ducttape_ has quit IRC00:41
*** _ducttape_ has joined #openstack-lbaas00:41
*** _ducttape_ has quit IRC00:42
*** ducttape_ has joined #openstack-lbaas00:42
*** yuanying has quit IRC00:55
*** openstackgerrit has quit IRC01:02
*** cody-somerville has quit IRC01:23
*** yuanying has joined #openstack-lbaas02:08
rm_workjohnsom: well, fixed that bug, discovered a new bug02:14
rm_workas is the way of things02:15
rm_workif a spares-pool amp fails over, there's a decent chance the housekeeping process will see too few spares and create one, so you end up with an extra spare02:15
rm_worknot a HUGE deal02:15
rm_workjust... annoying02:15
*** openstackgerrit has joined #openstack-lbaas02:21
openstackgerritAdam Harwell proposed openstack/octavia master: Fix failover for unallocated amps (spares-pool)  https://review.openstack.org/42799002:21
rm_workjohnsom: hopefully that's the right approach02:23
rm_workseems really straightforward02:23
johnsomYeah, the excess spare or two was a "do we care" choice02:29
johnsomI will take a look when I get home02:30
rm_worknp02:31
openstackgerritAdam Harwell proposed openstack/octavia master: Implement sorting and pagination for octavia  https://review.openstack.org/38214702:40
johnsomFyi, there is an actual bug in launchpad for the extra spares02:57
*** sanfern has joined #openstack-lbaas03:00
rm_workAh, well, good to know it's tracked :P03:02
rm_workand that we don't really care03:02
*** cody-somerville has joined #openstack-lbaas03:35
*** armax has quit IRC03:40
*** dileepr has quit IRC03:49
*** armax has joined #openstack-lbaas03:56
*** armax has quit IRC04:10
*** armax has joined #openstack-lbaas04:22
*** armax has quit IRC04:34
*** ducttape_ has quit IRC04:36
*** cody-somerville has quit IRC04:43
*** cody-somerville has joined #openstack-lbaas05:02
*** dileepr has joined #openstack-lbaas05:19
*** amotoki has joined #openstack-lbaas05:28
*** ducttape_ has joined #openstack-lbaas05:37
*** ducttape_ has quit IRC05:42
*** yuanying has quit IRC05:57
*** yuanying has joined #openstack-lbaas06:11
*** bedis has quit IRC06:12
*** bedis has joined #openstack-lbaas06:13
*** anilvenkata has joined #openstack-lbaas06:32
*** rcernin has joined #openstack-lbaas07:06
*** ducttape_ has joined #openstack-lbaas07:08
*** ducttape_ has quit IRC07:13
*** tesseract has joined #openstack-lbaas07:16
*** yamamoto has quit IRC07:25
*** eezhova has joined #openstack-lbaas07:38
*** pcaruana has joined #openstack-lbaas07:39
*** pcaruana has quit IRC07:42
*** pcaruana has joined #openstack-lbaas07:43
*** yamamoto has joined #openstack-lbaas07:56
*** yamamoto_ has joined #openstack-lbaas07:59
*** yamamoto has quit IRC08:03
*** sanfern has quit IRC08:03
*** sanfern has joined #openstack-lbaas08:07
*** yuanying_ has joined #openstack-lbaas08:08
*** eezhova has quit IRC08:10
*** yuanying has quit IRC08:11
*** rcernin has quit IRC08:28
openstackgerritAdam Harwell proposed openstack/octavia master: Fix failover for unallocated amps (spares-pool)  https://review.openstack.org/42799008:33
*** ducttape_ has joined #openstack-lbaas08:38
*** ducttape_ has quit IRC08:43
*** sanfern has quit IRC08:43
*** sanfern has joined #openstack-lbaas08:44
*** Alex_Staf has joined #openstack-lbaas08:46
*** eezhova has joined #openstack-lbaas08:51
*** kobis has joined #openstack-lbaas08:52
*** rcernin has joined #openstack-lbaas08:52
openstackgerritMerged openstack/octavia master: Remove outdated kernel tuning for haproxy  https://review.openstack.org/42795709:17
*** yamamoto_ has quit IRC09:20
*** yamamoto has joined #openstack-lbaas09:20
*** cody-somerville has quit IRC09:21
*** yamamoto has quit IRC09:25
openstackgerritChristian Zunker proposed openstack/neutron-lbaas master: Prevent LBaaSv2 agent from filling up logs.  https://review.openstack.org/30528409:31
openstackgerritMerged openstack/neutron-lbaas master: Replace six.iteritems with dict.items()  https://review.openstack.org/40105109:37
*** yamamoto has joined #openstack-lbaas09:53
*** cody-somerville has joined #openstack-lbaas09:55
*** ducttape_ has joined #openstack-lbaas10:09
*** ducttape_ has quit IRC10:14
*** yamamoto has quit IRC10:15
*** yamamoto has joined #openstack-lbaas10:20
*** yamamoto has quit IRC10:27
*** kobis has quit IRC10:32
*** kobis has joined #openstack-lbaas10:33
*** yamamoto has joined #openstack-lbaas10:51
*** yamamoto has quit IRC10:57
*** ducttape_ has joined #openstack-lbaas11:10
*** ducttape_ has quit IRC11:14
*** sanfern has quit IRC11:16
*** yuanying_ has quit IRC11:20
*** ducttape_ has joined #openstack-lbaas11:30
*** dileepr has quit IRC11:39
*** catintheroof has joined #openstack-lbaas11:47
*** ducttape_ has quit IRC11:57
*** yamamoto has joined #openstack-lbaas11:57
*** ducttape_ has joined #openstack-lbaas12:00
*** openstackgerrit has quit IRC12:03
*** ducttape_ has quit IRC12:22
*** yamamoto has quit IRC12:31
*** yamamoto has joined #openstack-lbaas12:39
*** ducttape_ has joined #openstack-lbaas12:42
*** sanfern has joined #openstack-lbaas12:43
*** yamamoto has quit IRC12:47
*** ducttape_ has quit IRC12:53
*** ducttape_ has joined #openstack-lbaas13:02
*** yamamoto has joined #openstack-lbaas13:47
*** yamamoto has quit IRC13:53
*** ducttape_ has quit IRC13:57
*** beardedeagle has joined #openstack-lbaas14:12
*** ducttape_ has joined #openstack-lbaas14:16
*** beardedeagle has quit IRC14:18
*** eezhova has quit IRC14:18
*** beardedeagle has joined #openstack-lbaas14:26
*** ducttape_ has quit IRC14:27
*** eezhova has joined #openstack-lbaas14:48
*** anilvenkata has quit IRC15:06
*** anilvenkata has joined #openstack-lbaas15:07
*** anilvenkata has quit IRC15:12
*** ducttape_ has joined #openstack-lbaas15:14
*** anilvenkata has joined #openstack-lbaas15:20
*** anilvenkata has quit IRC15:25
*** beardedeagle has quit IRC15:32
*** beardedeagle has joined #openstack-lbaas15:32
*** anilvenkata has joined #openstack-lbaas15:36
*** fnaval has joined #openstack-lbaas15:59
*** armax has joined #openstack-lbaas15:59
*** _ducttape_ has joined #openstack-lbaas16:00
*** _ducttape_ has quit IRC16:01
*** ducttape_ has quit IRC16:03
*** ducttape_ has joined #openstack-lbaas16:06
xgermanI have16:06
xgermanhttps://www.irccloud.com/pastebin/SQyT2yrl/16:06
johnsomWhat is "amp_ssh_allowed_access"?16:07
johnsomHa, didn't know we had that setting.16:07
xgermanwell, that would explain why it doesn’t do anyhting16:07
xgermanbut I see that setting in my devstack as well16:08
*** openstackgerrit has joined #openstack-lbaas16:08
openstackgerritSindhu Devale proposed openstack/octavia master: Octavia v2 API for listeners  https://review.openstack.org/42474416:08
johnsomYeah, it looks like something Phil added:16:08
johnsomhttps://github.com/openstack/octavia/commit/0d0853f4b9d2d7a7debfe5e293cc3f77b15cd97d16:08
johnsomNot sure why we needed that16:08
xgermanwell, it doesn’t work16:09
xgermanbut Phil’s code looked kosher16:11
johnsomYeah, oddness16:16
xgermanok, figure out the problem16:16
xgermanfix forthcoming16:16
johnsomBTW, I have tracked down the current issue facing our Py3x gates....16:16
johnsomSo, this change: https://github.com/openstack-dev/devstack/commit/7682ea88a6ab8693b215646f16748dbbc2476cc416:16
johnsomMeans that configdrive is not the default in devstack anymore.  Which means the instances booted via nova use the metadata service (ugh, I have issues, but whatever).16:17
xgermanYuck16:17
johnsomProblem is, the neutron metadata proxy is crashing on python3:16:17
johnsomhttp://logs.openstack.org/99/407099/25/check/gate-rally-dsvm-py35-neutron-neutron-ubuntu-xenial/4741e0d/logs/screen-q-l3.txt.gz?level=ERROR#_2017-02-02_11_41_52_02916:17
johnsomKind enough to not have any errors, it just dies after dumping it's debug config.16:18
johnsomSo Cirros boots up, but doesn't get configured and ends up with no ssh keys, so our gates bomb trying to log in and setup the web server16:19
openstackgerritGerman Eichberger proposed openstack/octavia master: Fixes misspelled amp_ssh_access_allowed config  https://review.openstack.org/42824016:19
xgermanjohnsom and if we used telnet everyhting would be fine16:19
johnsomUhhh, yeah, I guess......  It would certainly support the "Opens" by making every instance very open....16:21
*** kberger has joined #openstack-lbaas16:25
openstackgerritAnkur proposed openstack/octavia master: Add v2 load balancer type and controllers  https://review.openstack.org/42354316:27
*** rcernin has quit IRC16:29
openstackgerritSindhu Devale proposed openstack/octavia master: Octavia v2 API for listeners  https://review.openstack.org/42474416:29
*** kobis has quit IRC16:29
*** eezhova has quit IRC16:30
nmagnezio/16:36
nmagnezixgerman, telnet.. LOL16:36
*** eezhova has joined #openstack-lbaas16:38
*** ducttape_ has quit IRC16:38
*** ducttape_ has joined #openstack-lbaas16:39
johnsomrm_work Online?16:40
*** tesseract has quit IRC16:41
*** Alex_Staf has quit IRC17:01
*** kobis has joined #openstack-lbaas17:09
*** kobis has quit IRC17:10
*** beardedeagle has quit IRC17:16
*** beardedeagle has joined #openstack-lbaas17:17
rm_workjohnsom: now17:18
rm_workjohnsom: yeah I wasn't testing with ACTIVE/STANDBY mode17:18
xgermanI have some sec-group question?17:19
xgermanwhy is 0.0.0.0/0 not the same as any?17:19
rm_workjohnsom: I wasn't sure if I had access to the amp data there17:19
rm_workjohnsom: all I saw I had access to was role17:19
rm_workand i was having trouble getting pdb to work inside flows >_>17:19
rm_workjohnsom: do you have a trick to make that work?17:19
rm_workxgerman: it should be? for an ipv4 rule? is it not?17:20
xgermannope, it was blocking17:20
rm_workpretty sure that's what my SGs say17:20
rm_workhuh.17:20
xgermanhttps://www.irccloud.com/pastebin/AYpLSUUr/17:20
xgermanonly after I added any it stopped blocking17:21
xgermanWTF?17:21
rm_workhuh.17:21
xgermanyep,  no idea17:22
rm_workjohnsom: ah nm it's covered, i had thought of that already17:23
rm_workjohnsom: see my comments17:23
rm_workjohnsom: I was looking at using a decider, but I'm not sure exactly what's in "history", would I have access to the amphora object in there so I could check the status field?17:24
*** anilvenkata has quit IRC17:25
openstackgerritMerged openstack/octavia master: Fix a typo in the release notes  https://review.openstack.org/42749617:29
*** chlong has joined #openstack-lbaas17:39
*** ducttape_ has quit IRC17:39
*** dileepr has joined #openstack-lbaas18:01
johnsomOk, sorry, had a call with some folks looking to use Octavia.18:03
johnsomrm_work Ok, yeah, I see how that will work correctly.  Not a huge fan of overloading "role"'s meaning, but it will work correctly based on the current code.18:14
rm_workYeah, so... How do we get access to the status in here?18:15
rm_workor better yet, do you know how to make pdb work18:15
rm_worklooking at turning off threading18:15
rm_workharlowja is saying just remove the executor so taskflow is single threaded and that might help? so I'm giving that a shot18:18
rm_workthough I'm in like a 5 hour training right now <_<18:18
johnsomYeah, I'm pretty sure your pdb issues are with the threading, you would have to attach to the right thread.18:22
johnsomWe could pass in the status like we do the role with the get call, just adding it here: https://github.com/openstack/octavia/blob/master/octavia/controller/worker/controller_worker.py#L62218:25
johnsomOr changing it to pass in the amp instead of the role18:26
johnsomAll get_failover_flow is doing is building the flow, it's not actually running taskflow yet.18:26
johnsomBut, that said, it's heading towards merge, so....18:27
johnsomThe get_*_flow methods are just lining up the tasks into a flow that could be run, i.e. defining the sequences18:28
*** diltram has quit IRC18:29
*** diltram has joined #openstack-lbaas18:34
*** ducttape_ has joined #openstack-lbaas18:40
*** ducttape_ has quit IRC18:44
openstackgerritMerged openstack/octavia master: Fix failover for unallocated amps (spares-pool)  https://review.openstack.org/42799019:01
*** saju_m has joined #openstack-lbaas19:04
*** anilvenkata has joined #openstack-lbaas19:18
*** ipsecguy has quit IRC19:18
*** ipsecguy has joined #openstack-lbaas19:19
*** eezhova has quit IRC19:20
*** ducttape_ has joined #openstack-lbaas19:32
*** sanfern has quit IRC20:01
*** sanfern has joined #openstack-lbaas20:02
*** sanfern has quit IRC20:02
*** anilvenkata has quit IRC20:08
*** kobis has joined #openstack-lbaas20:47
*** harlowja has quit IRC21:00
*** _ducttape_ has joined #openstack-lbaas21:06
openstackgerritMerged openstack/octavia master: Fixes misspelled amp_ssh_access_allowed config  https://review.openstack.org/42824021:09
*** ducttape_ has quit IRC21:10
*** catintheroof has quit IRC21:14
*** catintheroof has joined #openstack-lbaas21:15
*** yamamoto has joined #openstack-lbaas21:19
*** catintheroof has quit IRC21:19
*** _ducttape_ has quit IRC21:23
*** ducttape_ has joined #openstack-lbaas21:24
rm_workjohnsom: OK, so I worked up the patch to do it the way you mentioned using status21:38
rm_workjohnsom: should I not bother? I FEEL like role should be fine here21:39
rm_workbut21:39
rm_workI'd just need to type "git review" here21:39
johnsomOh, let's go for it21:39
johnsomKeep things nice and clean.  Let me sleep with a clear conscious....21:40
johnsomgrin21:40
johnsomI am eating lunch then will work on the stable/rc121:40
rm_workactually21:42
rm_workthis fixes the other bug too21:42
rm_worksec let me look up that ID21:42
johnsomhttps://bugs.launchpad.net/octavia/+bug/161550221:43
openstackLaunchpad bug 1615502 in octavia "LBAAS - housekeeping serive does not cleanup stale amphora VMs" [Low,In progress] - Assigned to Ravikumar (ravikumar-vallabhu)21:43
rm_worknah it wasn't that i don't think?21:44
rm_workbecause that's different21:44
johnsomThat is the one I was thinking of21:44
rm_workI was looking at that just now21:44
rm_workok21:44
rm_workthen it's more like21:44
rm_workit fixes randomly getting an extra one to begin with21:44
rm_work<_<21:44
rm_workk21:46
openstackgerritShashank Kumar Shankar proposed openstack/octavia master: Introduce Octavia v2 API for pools  https://review.openstack.org/40592221:48
rm_workugh21:54
rm_workthe tests that came in from the RHAmp patch broke stuff on OSX21:54
rm_worksome of them look at real files :/21:54
rm_workFile "octavia/amphorae/backends/agent/api_server/amphora_info.py", line 98, in _get_meminfo21:54
rm_work        with open('/proc/meminfo', 'r') as meminfo:21:54
rm_work    IOError: [Errno 2] No such file or directory: '/proc/meminfo'21:54
rm_workfrom octavia.tests.unit.amphorae.backends.agent.api_server.test_amphora_info.TestAmphoraInfo.test__get_meminfo21:55
johnsom??? I wrote that.  I'm using the mock_open for that21:56
rm_workhmm21:56
rm_workwell on OSX it's breaking21:56
rm_workbecause ... that isn't there21:57
johnsomhttps://github.com/openstack/octavia/blob/master/octavia/tests/unit/amphorae/backends/agent/api_server/test_amphora_info.py#L12921:57
johnsomOh!21:57
johnsomRight, I did put in a "live data" test21:57
johnsomDarn, probably should not have done that...21:57
rm_workthere's a few more21:57
rm_workunrelated, but same patch21:58
johnsomI was thinking we should catch if the kernel changed the terms, etc.21:58
rm_worklots of: octavia.common.exceptions.InvalidAmphoraOperatingSystem: Invalid amphora operating system:21:58
johnsomProbably a bad idea21:58
rm_workI'll track these fails down in a minute21:58
rm_worknot related to my patch21:58
johnsomYeah, for meminfo we can just nuke: https://github.com/openstack/octavia/blob/master/octavia/tests/unit/amphorae/backends/agent/api_server/test_amphora_info.py#L7021:59
johnsomand the asserts below it22:00
rm_workah yeah so22:04
rm_workduring the tests22:04
rm_workwe don't mock which OS22:04
rm_workit is ACTUALLY LOOKING to see which OS22:04
rm_workin the osutils in test_plug.py22:04
openstackgerritAdam Harwell proposed openstack/octavia master: Improve the failover handling for spares-pool amps  https://review.openstack.org/42846522:06
rm_workjohnsom: ^^ that should do it22:10
rm_workjust short-circuit before any of the replacement code22:10
rm_worka bit simpler even22:10
johnsomYep, just looked it over.  Looks ok to me22:10
johnsomAre you spinning a test fix or should I look at it?22:18
rm_workAH almost done22:21
rm_workwith the test fix22:21
rm_workone last issue...22:21
johnsomThanks22:21
rm_workit needs a config param22:21
rm_workerg and afk for a little bit22:25
*** harlowja has joined #openstack-lbaas22:25
rm_workbut i'll have it up in a bit22:25
*** ducttape_ has quit IRC22:26
*** ducttape_ has joined #openstack-lbaas22:31
*** harlowja has quit IRC22:32
*** harlowja has joined #openstack-lbaas22:33
openstackgerritMichael Johnson proposed openstack/octavia master: Fix py3x gates  https://review.openstack.org/42489022:38
*** ducttape_ has quit IRC22:43
-openstackstatus- NOTICE: Restarting gerrit due to performance problems22:51
johnsomrm_work FYI, pagination proposal: https://review.openstack.org/#/c/39097322:52
*** yuanying has joined #openstack-lbaas23:08
armaxjohnsom: ping23:19
johnsomarmax Yo23:19
armaxoyoyoy23:19
nmagnezirm_work, we can always declare Octavia supports OSX and add it to osutils :D23:28
*** _beardedeagle has joined #openstack-lbaas23:29
rm_workorlol23:30
rm_work*lol23:30
rm_worki've got the fix23:30
rm_workworking on the last set23:30
*** _beardedeagle has quit IRC23:31
*** beardedeagle has quit IRC23:33
openstackgerritAdam Harwell proposed openstack/octavia master: Fix unit tests for OSX  https://review.openstack.org/42848423:35
rm_worknmagnezi / johnsom ^^23:35
rm_workrunning tests one more time to double-check i got them all23:35
johnsomAwesome, thanks!23:35
rm_workyeah seems good23:38
rm_workbrb23:46
*** amotoki has quit IRC23:48
johnsomUgh, the one test to see if I can work around the neutron py3x issue happened to hit the broken OVH hosts23:53
« Wednesday, 2017-02-01 Index Friday, 2017-02-03 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!