Tuesday, 2014-08-26

*** sballe has joined #openstack-lbaas00:04
*** sballe__ has quit IRC00:07
*** crc32 has quit IRC00:12
*** sballe has quit IRC00:12
*** HenryG_ is now known as HenryG00:16
*** markmcclain1 has quit IRC00:36
*** xgerman has quit IRC00:40
*** sbfox has quit IRC00:45
*** Zebra has joined #openstack-lbaas01:28
*** mlavalle has quit IRC01:53
*** woodster_ has quit IRC02:15
sbalukoffxgerman: Do you mean there's no major difference from before, or no major difference between running one and two processes?02:40
sbalukoff(Sorry--  if you hadn't noticed, I have a tendency to get pulled off onto other things for hours at a time. Yay for asynchronous communication. XD )02:41
openstackgerritBrandon Logan proposed a change to stackforge/octavia: Initial migration for database structure  https://review.openstack.org/11467103:34
*** ptoohill-oo has joined #openstack-lbaas04:06
*** sbfox has joined #openstack-lbaas04:13
*** amotoki has joined #openstack-lbaas04:19
*** fnaval has quit IRC05:21
*** Zebra has quit IRC05:39
*** Zebra has joined #openstack-lbaas05:40
*** fnaval has joined #openstack-lbaas05:58
*** fnaval has quit IRC06:45
*** jschwarz has joined #openstack-lbaas07:29
*** sbfox has quit IRC08:18
*** Zebra has quit IRC09:47
*** sqonkey has joined #openstack-lbaas11:02
*** ptoohill-oo has quit IRC11:02
*** sqonkey has quit IRC12:10
*** ptoohill-oo has joined #openstack-lbaas12:10
*** ptoohill-oo has quit IRC12:34
*** ptoohill-oo has joined #openstack-lbaas12:36
*** sballe has joined #openstack-lbaas13:39
*** markmcclain has joined #openstack-lbaas14:05
*** markmcclain1 has joined #openstack-lbaas14:09
*** markmcclain has quit IRC14:12
*** markmcclain1 has quit IRC14:13
*** markmcclain has joined #openstack-lbaas14:14
*** markmcclain1 has joined #openstack-lbaas14:15
*** markmcclain has quit IRC14:19
*** markmcclain1 has quit IRC14:22
*** markmcclain has joined #openstack-lbaas14:23
*** woodster_ has joined #openstack-lbaas14:31
*** amotoki has quit IRC14:36
*** xgerman has joined #openstack-lbaas14:58
*** mlavalle has joined #openstack-lbaas15:04
*** markmcclain has quit IRC15:33
*** jorgem has joined #openstack-lbaas15:35
blogan_hello to all16:07
*** jschwarz has left #openstack-lbaas16:12
blogan_xgerman: you wanted an ERROR operating status?16:13
xgermanyes, you never know what causes irregular operations16:14
*** fnaval has joined #openstack-lbaas16:15
blogan_wouldn't the degraded status be similar?16:15
xgermanDEGRADED is that the nodes are not reponding; ERROR is that the LB has issues16:16
xgermanor Listener, or16:16
blogan_wouldn't that be what the provisioning status is for?16:16
xgermanwell, if you provision am emtity to be ACTIVE, then after a while being active encounters an error...16:17
blogan_im sure you're right, I just can't see the case where setting the operating status to ERROR would make much sense16:17
xgermanwell, in Libra's case haproxy sometimes crashes, then we try to fail over, that fails for some reason, and then we set  it to ERROR16:18
blogan_i guess im not dileneating provisioning and operating status in my head properly16:18
blogan_so provisioning status is a status that can only be changed by a user doing an action on an entity, and operating status will be a status that can be updated by the system16:19
xgermanthat is how I look at it16:19
xgermanbut I  might be wrong16:19
blogan_well if that is the case, then, load balancer should have an operating status16:19
blogan_no I can definitely see how that makes sense16:20
xgermanyeah, but I am happy as long as there is a way to express ERROR16:20
blogan_well it could be through the provisioning status field16:20
blogan_but I'd like to get more people's opinions on it, i'm indifferent towards either16:21
xgermanyeah, more people is good -- I am leaning towards operational status since for me it feels ambiguious to have it provisioned correctly but getting an error because the CPU explodes16:22
xgermanERROR in provisioning would indicate that I (the user have to do something) whereas in OPERATION the system might still self heal16:22
xgermanor even worse I the user did something wrong16:23
blogan_yeah I think you're right16:24
blogan_you've convinced me16:24
*** sballe has quit IRC16:24
blogan_dont ever try that again16:25
xgermanlet's wait for sbalukoff's contrarian view16:31
*** sballe_ has joined #openstack-lbaas16:31
blogan_xgerman: he may just agree with you16:31
*** sballe_ has quit IRC16:37
*** markmcclain has joined #openstack-lbaas16:38
*** markmcclain has quit IRC16:38
*** balles has joined #openstack-lbaas16:40
sbalukoffI like the dileneation that provisioning status is triggered based on user actions (usually having to do with configuration changes), but operational status is triggered based on events within the system probably beyond the user's control.16:42
sbalukoffDo y'all ever see the potential for an provisional status to be ERROR, but the operational status to be UP?16:42
sbalukoffxgerman: Did you see my question to you last night about the benchmark data you've collected?16:44
sbalukoffWhen you said "there's no major difference" did you mean between the results with tuning and the result without tuning, or between the 1 process / 2 listeners approach and the 2 processes / 1 listener approach?16:45
blogan_sbalukoff: I could definitely see a provisional status being ERROR and operational being UP if it isn't handled correctly16:46
sbalukoffblogan_: Say... in the case where the user tries to make a change, and this doesn't succeed but prior to pushing any changes out to the Octavia VMs?16:47
blogan_as in haproxy starts on the VM, but for whatever reason, there is an error in the communication back to the controller16:47
*** sballe_ has joined #openstack-lbaas16:47
sbalukoffblogan_: Aah. True.16:48
xgermansbalukoff, no major difference between 2 processes/2 listeners -- thought he later has a lower error rate16:48
sbalukoffxgerman: Yeah, that's a little odd to me.  It would be interesting to find out what the nature of those errors was.16:48
blogan_sbalukoff: the case you just mentioned is much more likely16:48
sbalukoffAlso, finding out how long you waited between runs can be helpful--  it's possible one of the machines in there was still working on cleaning up TIME_WAIT sockets from a previous run or something, which can lead to errors under high load.16:49
xgermanyeah, I figured it was nothing to be too worried about16:50
xgermanI also looked at CPU/MEM during the runs and that was far away from the max16:50
*** markmcclain has joined #openstack-lbaas16:51
sbalukoffxgerman: Ok. I assume you're going to be at tomorrow's Octavia meeting, right? It would be great to go over the results of your benchmarks and try for consensus around the 1 listener per haproxy vs. 1 loadbalancer per haproxy direction, eh.16:52
xgermanwe are also under new management (Marc I.) so we are rethinking our position on that question...16:52
sbalukoffOh, ok.16:52
xgermanyep, will be there and we can go over it16:52
sbalukoffxgerman: Did that happen in this last week?16:52
xgermanyeah, he just started :-)16:53
sbalukoffOk, well, cool. I hope he's good to work with, eh!16:53
xgermanhe is three levels above me but very passionate about LBaaS16:53
sballe_He is great! he knows jorgem very well :) is used to be at Rackspace16:53
xgermanyeah, +1 he is great and knows LBaaS inside/out16:54
sballe_I am just sitting and listening in the background. I am still ctaching up from my vacation in Denmark16:54
sbalukoffsballe_: Yeah, I was out for the last few days too--  trying to make sure I have my action items taken care of prior to tomorrow's meeting. XD16:55
blogan_are we doing irc meeting tomorrow?16:55
sbalukoffSpeaking of which, if y'all haven't started reviewing the v0.5 design, or blogan's proposed DB migrations, please do so today!16:55
sbalukoffblogan_: Yes, I'll send an agenda out in a few hours.16:56
blogan_I need to make one more change, per xgerman's suggestion16:56
sbalukoffblogan_: Sounds good.16:56
sbalukoffOh! Does anyone know if we need to do anything special in this IRC channel to use the openstack bots here to run a meeting?16:58
sbalukoffI guess I can look into that today, eh.16:59
blogan_i have no idea on that one17:01
blogan_i assume the first thing we need is an open time slot17:01
blogan_and maybe thats all we need17:01
blogan_seems like they would have some kind of central registrar to lock in time slots though17:01
sbalukoffYeah...  I'll be trying to find documentation on this somewhere. I'm hoping just having the IRC bots here means they're ready for running meetings.17:03
sbalukoffHeck, let me try something17:03
sbalukoff#startmeeting Octavia17:03
openstackMeeting started Tue Aug 26 17:03:41 2014 UTC and is due to finish in 60 minutes.  The chair is sbalukoff. Information about MeetBot at http://wiki.debian.org/MeetBot.17:03
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.17:03
openstackThe meeting name has been set to 'octavia'17:03
openstackMeeting ended Tue Aug 26 17:03:48 2014 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)17:03
openstackMinutes:        http://eavesdrop.openstack.org/meetings/octavia/2014/octavia.2014-08-26-17.03.html17:03
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/octavia/2014/octavia.2014-08-26-17.03.txt17:03
openstackLog:            http://eavesdrop.openstack.org/meetings/octavia/2014/octavia.2014-08-26-17.03.log.html17:03
sbalukoffAnd now, my research is done, and I've saved myself probably an hour or two. XD17:04
sbalukoffOk, since this channel was created at the same time the Octavia project was birthed (at the mid-cycle summit), and no other meetings happen here... I think having the meeting here probably makes the most sense.17:05
sbalukoffAnyway, I'll be sending out notifications about it to the mailing list in a few hours.17:05
sbalukoffIn the mean time, I've got to go AFK for a couple hours.17:05
*** crc32 has joined #openstack-lbaas17:48
dougwigthere goes sbalukoff again, finding new ways to make octavia meetings exclusive.  keeping them open for less than a second doesn't allow anyone to type.  i think we need a ML thread about this.18:14
*** sballe__ has joined #openstack-lbaas18:17
*** sballe_ has quit IRC18:19
*** xgerman has quit IRC18:22
*** xgerman has joined #openstack-lbaas18:23
ballesdougwig, lol18:23
*** balles has quit IRC18:25
*** balles has joined #openstack-lbaas18:25
*** xgerman has quit IRC18:49
*** sbfox has joined #openstack-lbaas19:02
rm_worksbalukoff: sounds good to me20:00
*** sbfox has quit IRC20:03
*** sbfox has joined #openstack-lbaas20:03
*** sbfox1 has joined #openstack-lbaas20:51
*** sbfox has quit IRC20:51
*** sbfox1 has quit IRC20:55
*** sbfox has joined #openstack-lbaas20:55
*** markmcclain has quit IRC21:12
*** barclaac has joined #openstack-lbaas21:15
*** barclaac has quit IRC21:30
*** barclaac has joined #openstack-lbaas21:31
barclaacjoin #openstack-neutron21:31
barclaacwrong window :-D21:31
*** sbfox has quit IRC21:36
*** xgerman has joined #openstack-lbaas21:42
*** sbfox has joined #openstack-lbaas21:42
*** sbfox1 has joined #openstack-lbaas21:44
*** sbfox has quit IRC21:46
*** markmcclain has joined #openstack-lbaas21:54
*** markmcclain has quit IRC21:54
*** markmcclain has joined #openstack-lbaas21:56
rm_workwhoops, think I missed saying hi to Dustin and Craig >_>22:05
sbalukoffOh, did they head out already?22:10
*** fnaval_ has joined #openstack-lbaas22:20
*** fnaval has quit IRC22:23
rm_workguessing so, thing ended at 5pm22:26
blogan_they didn't even try to come see us :(22:27
blogan_sbalukoff why do they hate us?22:27
*** sbfox1 has quit IRC22:30
*** sballe_ has joined #openstack-lbaas22:32
*** barclaac|2 has joined #openstack-lbaas22:35
*** sballe__ has quit IRC22:36
*** sbfox1 has joined #openstack-lbaas22:38
*** sballe_ has quit IRC22:41
*** sbfox2 has joined #openstack-lbaas22:41
*** sbfox1 has quit IRC22:42
*** barclaac has quit IRC22:43
*** mlavalle has quit IRC22:43
*** openstackgerrit has quit IRC22:43
*** sballe has joined #openstack-lbaas22:44
*** HenryG has quit IRC22:46
*** sballe has quit IRC22:54
*** barclaac|2 has quit IRC22:55
*** barclaac|2 has joined #openstack-lbaas22:55
sbalukoffblogan_: Well, it's all I can do to stand y'all. ;)22:59
sbalukoffSeriously though-- that is odd. I guess they must have had to get to the airport in a hurry.22:59
sbalukoffI know Dustin really wanted to talk with your network people about how Octavia is going to plug in there.23:00
xgermansbalukoff, for the agenda tomorrow: You know our disagreement about rendering haproxy config files in the controller vs. on the haproxy machine23:06
xgermanI am wondering if we can discuss that at the meeting also23:06
*** markmcclain has quit IRC23:07
crc32So was the meeting going to be on IRC or WebEX?23:10
blogan_crc32: IRC for the next couple of weeks, then re-evaluate23:15
xgermanyeah, the irc party one23:19
*** sbfox2 has quit IRC23:21
*** fnaval_ has quit IRC23:34
*** Zebra has joined #openstack-lbaas23:43
*** jorgem has quit IRC23:49

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