Wednesday, 2017-02-01

« Tuesday, 2017-01-31 Index Thursday, 2017-02-02 »
*** ducttape_ has joined #openstack-lbaas00:01
*** amotoki has joined #openstack-lbaas00:04
openstackgerritMichael Johnson proposed openstack/octavia: Fix a typo in the release notes  https://review.openstack.org/42749600:12
*** mugsie has quit IRC00:29
openstackgerritMichael Johnson proposed openstack/octavia: Fix py3x gates  https://review.openstack.org/42489000:38
johnsomDon't get excited, I just added some additional logging to see what is up00:39
openstackgerritMichael Johnson proposed openstack/octavia: Fix py3x gates  https://review.openstack.org/42489000:50
*** beardedeagle has quit IRC00:51
*** beardedeagle has joined #openstack-lbaas00:52
*** ducttape_ has quit IRC00:55
*** beardedeagle has quit IRC00:56
*** ducttape_ has joined #openstack-lbaas01:06
*** kevo has quit IRC01:11
rm_workjohnsom: uhh, issue with health checking on spares pool...01:11
rm_workhttps://gist.github.com/rm-you/4e91110b008213b818f22150f17ef61201:11
rm_worki stopped the agent on an amp in the spares pool, looks like it got MOST of the way through the recycle operation and failed because it thought it should have a VIP :/01:11
rm_workwhich ... it wouldn't yet01:11
rm_workguess that's a bug01:13
johnsomJoy, so a bug with the spares pool.01:13
rm_workwhich I'm honestly not sure why people don't use01:13
johnsomProbably not related to RH patch?01:13
rm_workno01:13
rm_workabsolutely not related01:13
rm_worktrying to multitask01:13
johnsomWell, we can get a fix into ocata.01:13
rm_workI'm going to see if I can patch it up01:14
johnsomI can't look at it until tomorrow though01:14
rm_workyeah no worries01:14
rm_workI'll tell you when I have a fix up01:14
rm_workjust an FYI01:14
johnsomOk01:14
rm_workI am honestly not sure though how much ACTIVE/PASSIVE buys us over using a spares pool01:14
rm_workthe time difference is basically the time it takes to send a config via REST01:15
rm_workwhich is ... <100ms?01:15
johnsomFaster01:15
rm_workso I guess we trade ~100ms for 2n instead of 1n+x01:15
johnsomWell, spares failover has to plug ports which is slow with neutron01:15
rm_workah.01:16
rm_workinteresting01:16
rm_workso I guess the way I'm implementing it, that doesn't matter01:16
rm_workI'm doing one port per amphora for VIP, and one FLIP per LB, and directing the FLIP to the active amp-port01:16
johnsomCould be.  Last time I checked, failover can take up to a minute on stock01:17
rm_workand exposing the FLIP address as the "VIP"01:17
rm_workwhich is kinda shitty because the user can't own/control the FLIP01:17
rm_workbut01:17
rm_workit "works" :/01:17
johnsomOk, I am on dinner duty tonight.  We are trying blue apron...01:19
rm_workaah lol01:27
rm_worktell me how that goes01:27
rm_worki don't think they ship to my area, i kinda checked briefly a while ago01:27
*** ducttape_ has quit IRC01:39
*** ducttape_ has joined #openstack-lbaas01:54
*** ducttape_ has quit IRC02:13
*** ducttape_ has joined #openstack-lbaas02:20
*** ducttape_ has quit IRC02:20
*** ducttape_ has joined #openstack-lbaas02:20
*** yuanying has quit IRC02:21
*** armax has quit IRC03:05
*** itsuugo has joined #openstack-lbaas03:23
*** itsuugo has quit IRC03:28
*** sticker_ has joined #openstack-lbaas03:29
*** sticker has quit IRC03:32
*** sanfern has joined #openstack-lbaas03:55
*** ducttape_ has quit IRC04:05
*** csomerville has joined #openstack-lbaas04:20
*** armax has joined #openstack-lbaas04:27
*** ducttape_ has joined #openstack-lbaas04:28
*** beardedeagle has joined #openstack-lbaas04:28
*** beardedeagle has quit IRC04:37
*** beardedeagle has joined #openstack-lbaas04:37
*** ducttape_ has quit IRC04:38
*** dileepr has joined #openstack-lbaas04:43
*** ducttape_ has joined #openstack-lbaas04:47
*** jamespage has quit IRC05:00
*** jamespage has joined #openstack-lbaas05:00
*** ducttape_ has quit IRC05:13
*** Alex_Staf has joined #openstack-lbaas05:31
*** beardedeagle has quit IRC06:16
*** beardedeagle has joined #openstack-lbaas06:17
*** beardedeagle has quit IRC06:21
*** gcheresh_ has joined #openstack-lbaas06:26
*** ducttape_ has joined #openstack-lbaas06:27
*** rcernin has joined #openstack-lbaas06:29
*** ducttape_ has quit IRC06:31
*** armax has quit IRC06:35
*** itsuugo has joined #openstack-lbaas06:37
*** itsuugo has quit IRC06:47
*** itsuugo has joined #openstack-lbaas06:47
*** itsuugo has quit IRC06:52
*** tesseract has joined #openstack-lbaas07:18
*** eezhova has joined #openstack-lbaas07:18
*** pcaruana has joined #openstack-lbaas07:19
*** yamamoto has quit IRC07:25
*** ducttape_ has joined #openstack-lbaas07:57
*** ducttape_ has quit IRC08:02
*** itsuugo has joined #openstack-lbaas08:14
*** eezhova has quit IRC08:16
*** yuanying has joined #openstack-lbaas08:30
*** cody-somerville has joined #openstack-lbaas08:43
*** cody-somerville has quit IRC08:43
*** cody-somerville has joined #openstack-lbaas08:43
*** eezhova has joined #openstack-lbaas08:46
*** amotoki has quit IRC08:49
*** sanfern has quit IRC08:49
*** sanfern has joined #openstack-lbaas08:55
*** ducttape_ has joined #openstack-lbaas09:28
*** amotoki has joined #openstack-lbaas09:30
*** ducttape_ has quit IRC09:32
*** amotoki has quit IRC09:38
*** amotoki has joined #openstack-lbaas09:45
*** amotoki has quit IRC09:47
*** amotoki has joined #openstack-lbaas09:50
*** sanfern has quit IRC10:04
*** sanfern has joined #openstack-lbaas10:05
*** sanfern has quit IRC10:21
*** sanfern has joined #openstack-lbaas10:27
*** ducttape_ has joined #openstack-lbaas10:58
*** ducttape_ has quit IRC11:03
*** amotoki has quit IRC11:18
*** sanfern has quit IRC11:25
*** dileepr has quit IRC11:59
*** catintheroof has joined #openstack-lbaas12:28
*** ducttape_ has joined #openstack-lbaas12:29
*** ducttape_ has quit IRC12:33
*** ducttape_ has joined #openstack-lbaas12:57
*** afranc has quit IRC13:09
*** afranc has joined #openstack-lbaas13:09
*** Alex_Staf has quit IRC13:13
*** adam_g has quit IRC13:17
*** sanfern has joined #openstack-lbaas13:23
*** adam_g has joined #openstack-lbaas13:25
*** Alex_Staf has joined #openstack-lbaas13:27
*** ducttape_ has quit IRC13:42
openstackgerritNir Magnezi proposed openstack/octavia: Add placement services to devstack example files  https://review.openstack.org/42771013:57
nmagnezijohnsom, rm_work ^^13:57
*** amotoki has joined #openstack-lbaas13:59
*** amotoki has quit IRC14:08
*** ducttape_ has joined #openstack-lbaas14:09
*** matt-borland has joined #openstack-lbaas14:14
*** beardedeagle has joined #openstack-lbaas14:16
*** ducttape_ has quit IRC14:22
*** ducttape_ has joined #openstack-lbaas14:22
*** ducttape_ has quit IRC14:23
*** ducttape_ has joined #openstack-lbaas14:28
*** ducttape_ has quit IRC14:31
*** csomerville has quit IRC15:12
*** csomerville has joined #openstack-lbaas15:13
*** mugsie has joined #openstack-lbaas15:13
*** kobis has joined #openstack-lbaas15:16
*** gcheresh_ has quit IRC15:24
*** ducttape_ has joined #openstack-lbaas15:34
*** ducttape_ has quit IRC15:34
*** ducttape_ has joined #openstack-lbaas15:35
*** armax has joined #openstack-lbaas15:35
*** rcernin has quit IRC15:51
*** kevo has joined #openstack-lbaas15:54
*** fnaval has joined #openstack-lbaas15:55
*** eezhova has quit IRC16:07
*** _ducttape_ has joined #openstack-lbaas16:09
*** ducttape_ has quit IRC16:13
openstackgerritAnkur proposed openstack/octavia: Add v2 load balancer type and controllers  https://review.openstack.org/42354316:30
*** kobis has quit IRC16:38
openstackgerritMerged openstack/octavia: Add placement services to devstack example files  https://review.openstack.org/42771016:48
*** kobis has joined #openstack-lbaas16:56
*** tesseract has quit IRC16:56
*** beardedeagle has quit IRC16:57
*** itsuugo has quit IRC16:57
*** beardedeagle has joined #openstack-lbaas16:57
*** Alex_Staf has quit IRC17:00
*** kevo has quit IRC17:06
*** Alex_Staf has joined #openstack-lbaas17:53
*** _ducttape_ has quit IRC17:54
*** ducttape_ has joined #openstack-lbaas17:54
*** ducttape_ has quit IRC17:54
*** ducttape_ has joined #openstack-lbaas17:55
*** ducttape_ has quit IRC17:55
*** ducttape_ has joined #openstack-lbaas17:56
johnsomStrangest thing, this placement thing is breaking our py3x gates, but not the py27 gates....17:58
johnsomhttp://logs.openstack.org/90/424890/3/check/gate-neutron-lbaasv2-dsvm-py3x-scenario-ubuntu-xenial-nv/aac3428/logs/screen-n-cpu.txt.gz#_2017-02-01_01_31_16_76617:59
johnsomCan't boot the cirros hosts17:59
*** kobis has quit IRC18:03
*** harlowja_ has joined #openstack-lbaas18:05
*** harlowja has quit IRC18:06
johnsomHmm, I'm getting that on both working and not, but it fixed it local.  Argh.18:12
*** pcaruana has quit IRC18:28
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - Initial Cluster Manager  https://review.openstack.org/40523818:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology: Initial Distributor Noop Driver  https://review.openstack.org/31300618:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology : create distributor network flow  https://review.openstack.org/40976318:37
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - register/uregister amphorae tasks  https://review.openstack.org/40976518:37
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - Cluster DB Tasks  https://review.openstack.org/40976418:37
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - LB create cluster impl  https://review.openstack.org/42656018:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology OVS-based Distributor Driver  https://review.openstack.org/31762918:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology OVS-based Distributor Backend  https://review.openstack.org/32042218:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - Distributor related tasks  https://review.openstack.org/40695118:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE - controller network tasks  https://review.openstack.org/32348118:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - Distributor image creation  https://review.openstack.org/40359418:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE - network driver related changes  https://review.openstack.org/32249418:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE - distributor certificate tasks  https://review.openstack.org/40695218:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - distributor creation flow  https://review.openstack.org/40695318:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - create shared distributor  https://review.openstack.org/40695418:37
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - Distributor open_flow back-end  https://review.openstack.org/42785818:37
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - Initial Cluster Manager  https://review.openstack.org/40523818:43
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - Distributor related tasks  https://review.openstack.org/40695118:43
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - Distributor open_flow back-end  https://review.openstack.org/42785818:43
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - Distributor image creation  https://review.openstack.org/40359418:43
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology : create distributor network flow  https://review.openstack.org/40976318:43
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - register/uregister amphorae tasks  https://review.openstack.org/40976518:43
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE - distributor certificate tasks  https://review.openstack.org/40695218:43
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - Cluster DB Tasks  https://review.openstack.org/40976418:43
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - distributor creation flow  https://review.openstack.org/40695318:43
openstackgerritValleriya Perelman proposed openstack/octavia: Active-Active Topology - LB create cluster impl  https://review.openstack.org/42656018:43
openstackgerritValleriya Perelman proposed openstack/octavia: ACTIVE-ACTIVE Topology - create shared distributor  https://review.openstack.org/40695418:43
*** Alex_Staf has quit IRC18:45
*** kevo has joined #openstack-lbaas18:53
*** ducttape_ has quit IRC19:02
*** ducttape_ has joined #openstack-lbaas19:03
*** yuanying has quit IRC19:03
*** dileepr has joined #openstack-lbaas19:14
johnsomJust a heads up, I did a full re-clone devstack this morning and it's totally broken19:16
*** harlowja_ has quit IRC19:19
rm_workffff19:28
rm_workok19:28
rm_worki'll umm19:29
rm_worktry not to do that?19:29
rm_work<_<19:29
rm_workthat's difficult for my workflow T_T19:29
*** sanfern has quit IRC19:40
rm_workjohnsom: finishing up reviewing the RH Amps19:50
rm_workjohnsom: I can merge when I'm done?19:50
johnsomYes, please do if you are happy with it19:51
rm_workOK :)19:51
openstackgerritShashank Kumar Shankar proposed openstack/octavia: Introduce Octavia v2 API for pools  https://review.openstack.org/40592219:53
nmagnezijohnsom, hey, re: broken devstack. do we know what was broken? doe that also break / will break the gates?20:00
nmagneziwell, it's meeting time :)20:01
johnsomIt is nova20:01
*** beardedeagle has quit IRC20:07
*** beardedeagle has joined #openstack-lbaas20:07
rm_workahhh shit20:13
rm_workyeah that'll break the gate >_>20:13
*** csomerville has quit IRC20:13
johnsomWell, we can give it a try20:29
rm_workoh right i forgot to mention that spares-health bug, but we can always just backport it20:36
rm_workOK, just did +A20:37
nmagnezimay the force be with the gates.20:38
*** harlowja has joined #openstack-lbaas20:39
johnsomWe are good, it's going straight to the gate-gate, so should pass20:40
openstackgerritNakul Dahiwade proposed openstack/octavia: Add v2 load balancer type and controllers  https://review.openstack.org/42354320:43
openstackgerritAnkur proposed openstack/octavia: Add v2 load balancer type and controllers  https://review.openstack.org/42354320:49
*** beardedeagle has quit IRC20:53
*** beardedeagle has joined #openstack-lbaas20:53
nmagnezirm_work, johnsom, maybe this will save the day (nova): https://review.openstack.org/#/c/427747/20:55
nmagnezioh.. wrong link? sec20:55
*** _ducttape_ has joined #openstack-lbaas20:58
*** ducttape_ has quit IRC21:02
rm_workbrb21:03
rm_workhopefully that merges fine :)21:03
rm_workI'll try stacking later and see if anything jumps out to me21:03
*** kobis has joined #openstack-lbaas21:08
*** eezhova has joined #openstack-lbaas21:08
*** kevo has quit IRC21:08
*** armax has quit IRC21:11
openstackgerritMerged openstack/octavia: Fix the amphora image support for RH Linux flavors  https://review.openstack.org/33184121:24
nmagneziyay!21:29
nmagnezilooks like we are ready for RC121:29
*** raj_ has joined #openstack-lbaas21:31
raj_A toss-up question - why does Octavia seem to want to have Rather Useful Things (tm) like TCP Selective ACKnowledgements and TCP Timestamps enabled?  A maze of twisty troubleshooting passages have led me to https://github.com/openstack/octavia/blob/master/elements/haproxy-octavia/post-install.d/20-haproxy-tune-kernel21:32
raj_grrr... s/enabled/disabled/21:33
johnsomThose are optimizations recommended to us by the folks at haproxy technologies21:33
raj_How are they optimizations???21:34
raj_They rather fully trash TCP's ability to recover from non-trivial bursts of packet loss.21:34
raj_I guess I should also ask if those same folks provided suggestions as to what the client and server timeouts should be for haproxy itself.21:35
raj_ftp://ftp.netperf.org/candidate.png is an example of what happens without SACK and Timestamps with a non-trivial loss burst.21:36
*** beardedeagle has quit IRC21:37
*** _ducttape_ has quit IRC21:37
*** catintheroof has quit IRC21:37
johnsomSelective acks does seem a bit counter intuitive.21:37
*** ducttape_ has joined #openstack-lbaas21:37
*** beardedeagle has joined #openstack-lbaas21:37
*** catintheroof has joined #openstack-lbaas21:37
johnsomFor client and server timeouts are you talking about the connect/response timeouts?21:38
raj_Most of the other sysctl settings in 2-haproxy-tune-kernel look more or less sane I suppose, but I'm not sure about the clipping of the maximum TCP window/socket buffer size to about half a megabyte.21:38
raj_  timeout  connect 10s21:39
raj_  timeout  client 1m21:39
raj_  timeout  server 1m21:39
raj_  timeout  check 10s21:39
johnsomYeah, ok.21:39
raj_timeout client and timeout server in this case21:39
johnsomThose are really dependent on the application.  We do have an open bug to expose those to users, but it has not been implemented yet.21:39
raj_I've not gotten around to finding-out what those are in an Octavia image.21:39
johnsomraj_ https://github.com/openstack/octavia/blob/master/octavia/common/jinja/haproxy/templates/base.j2#L3221:40
raj_Units there are milliseconds then I suppose?21:41
johnsomCorrect21:41
raj_Without SACK and Timestamps, a non-trivial packet loss burst will result in what you see in that .png  - the RTO will continue to double until one hits the client (or server I suppose) timeout on forward progress and haproxy will nuke the connection with extreme prejudice.21:42
raj_All the quicker/more effective loss recovery heuristics of TCP tend to depend on SACK and/or Timestamps...21:43
johnsomI am pretty sure the timestamp disable is about the latency to generate the timestamps.21:44
raj_???21:44
johnsom(Trying to pull up the way-back machine as this has been in the code for a long time)21:44
johnsomAs such, this could be a hold over from older kernels too21:45
raj_Those settings, and perhaps some il-behaved image building on baremetal build nodes, seem to be at the root of a problem which has been festering close to home for a long time too :)21:46
raj_The only thing I've ever heard as negatives about TCP Timestamps is they tend to shrink the effective TCP Maximum Segment Size (MSS) by 12 bytes.21:47
raj_I don't recall anyone reporting they significantly increased say the service demand reported by a netperf benchmark.21:47
johnsomTroubles with helion?21:47
raj_Well, hypothetically speaking, let's assume the following:21:48
raj_A build pipeline involving some bare metal systems separated from the likes of git repositories by an occasionally congested path...21:48
raj_From time to time, build jobs report connection reset by peer when trying to access the git server(s) which happen to be behind an haproxy (not octavia)...21:49
johnsomSo, some googling to refresh my memory has RedHat giving advice on the timestamps: https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux_for_Real_Time/7/html/Tuning_Guide/Reduce_TCP_performance_spikes.html21:50
raj_And an image building job that seems to end-up with the settings in https://github.com/openstack/octavia/blob/master/elements/haproxy-octavia/post-install.d/20-haproxy-tune-kernel ending-up being set on the baremetal nodes.21:50
*** eezhova has quit IRC21:50
raj_Thus, hypothetically speaking, one might end-up looking at those Octavia-desired sysctl settings and asking Why?!? :)21:50
raj_That redhat link is blissfully free from the ravages of details.21:51
johnsomHaha21:51
johnsomYeah.  We had a person participating from HAProxy tech back when those were created.  They are probably worth revisiting.21:52
raj_Since it is in an area related to "real time" I would guess that in the RHEL7 (?) kernel, there can be situations where going to get time can cause a bit of a short hiccup.21:52
johnsomGlad to know folks thought highly of our customizations and borrowed them.  I'm pretty sure I had nothing to do with that when I was hypothetically involved with that product.21:53
raj_I'd be happy to file a ticket - just need to be pointed at where.21:53
johnsomhttps://bugs.launchpad.net/octavia/+filebug21:53
raj_Well, in this hypothetical situation, those settings were not consciously adopted, they seem to have leaked-out onto the baremetal for reasons I don't yet fully understand.21:54
*** armax has joined #openstack-lbaas21:54
johnsomThe DIB setup in that hypothetical environment needed to be burned to the ground IMHO.  It was a hack job that gave me no end of headaches trying to integrate with...21:55
*** eezhova has joined #openstack-lbaas21:55
johnsomDigging a little deeper, I think those settings were directly related to issues in the 2.4 and 2.6 kernel series21:56
*** catintheroof has quit IRC21:57
*** catintheroof has joined #openstack-lbaas21:58
raj_https://bugs.launchpad.net/octavia/+bug/166110522:01
openstackLaunchpad bug 1661105 in octavia "Octavia should not disable TCP Selective ACKnowledgements or Timestamps" [Undecided,New]22:01
raj_johnsom, hypothetically speaking, is there a later, better image building mechanism out there?22:04
johnsomDIB works reasonably well when it is not "customized" in such a way as I saw22:05
raj_Does the dimm memory have few enough bit errors to have an idea as to what was behind the clipping of tcp_[rw]mem?22:06
*** kobis has quit IRC22:06
raj_(That is how I tend to describe my own wayback :) )22:06
*** beardedeagle has quit IRC22:07
*** beardedeagle has joined #openstack-lbaas22:07
*** eezhova has quit IRC22:08
*** ducttape_ has quit IRC22:09
*** ducttape_ has joined #openstack-lbaas22:10
johnsomI am pretty sure that was to reduce the memory footprint while handling a large number of connections22:11
johnsomconcurrent connection metrics are a "thing" in the load balancing world22:12
*** ducttape_ has quit IRC22:15
raj_Sounds reasonable, though the Tput <= Window/RTT cost might be a bit high.22:15
xgermanhi, like all the hypotheticals22:18
johnsomHaha22:18
xgermanand concur with johnsom22:18
xgermanI did some benchmarks on a now defunct publi cloud and they were bad so people asked me to apply tuning parameters and here we are22:19
raj_I'm not sure how such settings would have increased performance but OK.22:20
johnsomThe guy that helped us out with that was super sharp and on top of this stuff.  My suspicion is that these have not withstood the test of time and said kernel is much better at some of this than it was at the time.22:21
raj_Do Octavia images get built "upstream?"  If so, I wonder if these settings might be leaking-out there too.22:21
xgermanwe build them afresh every time you install devstack22:22
johnsomWe provide a script to build images for Octavia amphora, but we do not ship pre-built images anywhere22:22
* raj_ was just wondering if this hypothetical leakage might be affecting other folks22:23
johnsomTo my knowledge, this config was only in the haproxy element for octavia.  It's not even in the native DIB haproxy element to my knowledge.22:26
johnsomYeah, don't see it there, so they didn't borrow either22:27
johnsomhttps://github.com/openstack/tripleo-image-elements/tree/master/elements/haproxy22:28
johnsomAnyhow, those configs are ~3 years old now, so probably time to update.22:29
xgerman+10022:30
*** catintheroof has quit IRC22:35
openstackgerritMichael Johnson proposed openstack/octavia: Remove outdated kernel tuning for haproxy  https://review.openstack.org/42795722:43
johnsomraj_ ^^^22:43
*** matt-borland has quit IRC22:44
raj_Works for me.  Paint it yellow and ship it :)  Do you want me to weigh-in in review?22:45
*** ducttape_ has joined #openstack-lbaas22:47
*** ducttape_ has quit IRC22:49
*** _ducttape_ has joined #openstack-lbaas22:49
*** _ducttape_ has quit IRC22:53
johnsomMore reviews the merrier....22:56
*** yuanying has joined #openstack-lbaas23:07
*** kobis has joined #openstack-lbaas23:08
*** kobis has quit IRC23:12
xgermanl8ter23:21
johnsomSee ya23:21
raj_Thanks for all the help/feedback/whatnot23:21
johnsomSure, NP23:21
*** yamamoto has joined #openstack-lbaas23:24
*** beardedeagle has quit IRC23:32
*** beardedeagle has joined #openstack-lbaas23:32
*** beardedeagle has quit IRC23:37
*** ducttape_ has joined #openstack-lbaas23:51
*** cody-somerville has quit IRC23:59
« Tuesday, 2017-01-31 Index Thursday, 2017-02-02 »

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