Wednesday, 2019-09-25

*** lseki has quit IRC00:03
*** armax has quit IRC00:03
*** links has joined #openstack-meeting00:06
*** links has quit IRC00:06
*** yangjianfeng has joined #openstack-meeting00:14
*** yangjianfeng has quit IRC00:16
*** markvoelker has joined #openstack-meeting00:24
*** jamesmcarthur has joined #openstack-meeting00:29
*** jamesmcarthur has quit IRC00:35
*** diablo_rojo has joined #openstack-meeting00:37
*** zhangchi has joined #openstack-meeting00:43
*** donghao has joined #openstack-meeting00:48
*** nitinuikey has joined #openstack-meeting00:52
*** nitinuikey has quit IRC00:52
*** gyee has quit IRC00:52
*** donghao has quit IRC00:53
*** mriedem has quit IRC01:00
zhangchi#startmeeting tricircle01:00
openstackMeeting started Wed Sep 25 01:00:26 2019 UTC and is due to finish in 60 minutes.  The chair is zhangchi. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.01:00
*** openstack changes topic to " (Meeting topic: tricircle)"01:00
openstackThe meeting name has been set to 'tricircle'01:00
zhangchiMorning all01:00
zhangchiwelcome back home to tricircle bio-weekly meeting01:01
zhangchiA new version will be published soon in a few days01:03
*** markvoelker has quit IRC01:08
*** ekcs has quit IRC01:10
zhangchiregion_name has been added in kuryr for neutron_client retriving data in different regions01:15
*** Liang__ has joined #openstack-meeting01:21
zhangchiAny other issues01:28
*** XSM_HUST has joined #openstack-meeting01:30
zhangchiHi XSM_HUST01:31
zhangchineutron_client in kuryr can retrieving data from different region now.How about your container guide manual01:32
XSM_HUSTI'm sorry that I'm busy looking for a job recently. The progress of the project may not be pushed forward recently.01:34
XSM_HUSTI will finish the container guide manual as soon as I finish my job-hunting.01:35
*** armax has joined #openstack-meeting01:36
zhangchiYeah , Good luck for your job-hunting. It does not matter much as a new development circle will be started . Hope to see you then01:38
*** diablo_rojo has quit IRC01:38
zhangchiThere has been a few exceptions for uni-test then as neutron master udpated. jobs listed here
zhangchiNext meeting is 9th Oct Hope to see you then01:42
*** armax has quit IRC01:42
zhangchiThanks for you attending XSM_HUST Good luck to you01:43
XSM_HUSTOk, thank you for your understanding.01:43
zhangchiI see.01:44
*** XSM_HUST has quit IRC01:45
*** hongbin has joined #openstack-meeting01:49
zhangchiHi hongbin01:52
zhangchiSee you all next time01:53
*** openstack changes topic to "OpenStack Meetings ||"01:53
openstackMeeting ended Wed Sep 25 01:53:40 2019 UTC.  Information about MeetBot at . (v 0.1.4)01:53
openstackMinutes (text):
*** njohnston has quit IRC02:03
*** armax has joined #openstack-meeting02:05
*** apetrich has quit IRC02:10
*** ekcs has joined #openstack-meeting02:24
*** ekcs has quit IRC02:25
*** ekcs has joined #openstack-meeting02:26
*** markvoelker has joined #openstack-meeting02:35
*** hongbin has quit IRC02:36
*** igordc has quit IRC02:39
*** markvoelker has quit IRC02:40
*** ricolin has joined #openstack-meeting02:47
*** donghao has joined #openstack-meeting02:50
*** donghao has quit IRC02:54
*** rfolco has quit IRC03:09
*** ekcs has quit IRC03:09
*** whoami-rajat has joined #openstack-meeting03:11
*** ekcs has joined #openstack-meeting03:21
*** psachin has joined #openstack-meeting03:31
*** Liang__ has quit IRC03:34
*** boxiang has quit IRC04:16
*** zhubx has joined #openstack-meeting04:17
*** ociuhandu has joined #openstack-meeting04:30
*** ociuhandu has quit IRC04:35
*** enriquetaso has joined #openstack-meeting04:38
*** dviroel has quit IRC04:40
*** artom has quit IRC04:41
*** donghao has joined #openstack-meeting04:44
*** donghao has quit IRC04:49
*** belmoreira has joined #openstack-meeting04:51
*** enriquetaso has quit IRC04:57
*** belmoreira has quit IRC05:05
*** ekcs has joined #openstack-meeting05:07
*** ekcs has quit IRC05:07
*** seongsoocho has joined #openstack-meeting05:34
*** hyunsikyang__ has joined #openstack-meeting05:37
*** hyunsikyang has quit IRC05:40
*** zbr|ruck has quit IRC05:43
*** zbr has joined #openstack-meeting05:43
*** brault has joined #openstack-meeting05:46
*** radeks has joined #openstack-meeting06:06
*** e0ne has joined #openstack-meeting06:16
*** kopecmartin|off is now known as kopecmartin06:20
*** enriquetaso has joined #openstack-meeting06:32
*** belmoreira has joined #openstack-meeting06:35
*** pcaruana has joined #openstack-meeting06:42
*** eharney has joined #openstack-meeting06:46
*** rpittau|afk is now known as rpittau06:47
*** markvoelker has joined #openstack-meeting06:47
*** donghao has joined #openstack-meeting06:47
*** donghao has quit IRC06:52
*** markvoelker has quit IRC06:52
*** slaweq has joined #openstack-meeting06:55
*** e0ne has quit IRC06:56
*** tesseract has joined #openstack-meeting07:01
*** ykatabam has quit IRC07:15
*** brault has quit IRC07:21
*** apetrich has joined #openstack-meeting07:27
*** enriquetaso has quit IRC07:34
*** enriquetaso has joined #openstack-meeting07:35
*** zhubx has quit IRC07:38
*** zhubx has joined #openstack-meeting07:38
*** e0ne has joined #openstack-meeting07:39
*** yaawang has quit IRC07:47
*** yaawang has joined #openstack-meeting07:49
*** ralonsoh has joined #openstack-meeting07:51
*** rpittau is now known as rpittau|bbl08:01
*** ykatabam has joined #openstack-meeting08:03
*** moguimar has joined #openstack-meeting08:16
*** e0ne has quit IRC08:28
*** priteau has joined #openstack-meeting08:34
*** brault has joined #openstack-meeting08:39
*** zbr is now known as zbr|ruck08:40
*** brault has quit IRC08:43
*** markvoelker has joined #openstack-meeting08:48
*** donghao has joined #openstack-meeting08:49
*** markvoelker has quit IRC08:53
*** donghao has quit IRC08:54
*** markvoelker has joined #openstack-meeting08:56
*** markvoelker has quit IRC09:00
*** ttsiouts has joined #openstack-meeting09:12
*** brault has joined #openstack-meeting09:24
*** e0ne has joined #openstack-meeting09:31
*** electrofelix has joined #openstack-meeting09:35
*** ociuhandu has joined #openstack-meeting09:35
*** ociuhandu has quit IRC09:36
*** ociuhandu has joined #openstack-meeting09:38
*** ociuhandu has quit IRC09:39
*** brault has quit IRC09:43
*** bbowen has quit IRC09:46
*** bbowen has joined #openstack-meeting09:47
*** brault has joined #openstack-meeting09:51
*** rpittau|bbl is now known as rpittau09:57
*** ociuhandu has joined #openstack-meeting09:59
*** enriquetaso has quit IRC10:02
*** ociuhandu has quit IRC10:04
*** brault has quit IRC10:17
*** pcaruana has quit IRC10:23
*** rcernin has quit IRC10:23
*** enriquetaso has joined #openstack-meeting10:30
*** pcaruana has joined #openstack-meeting10:36
*** zhangchi has quit IRC10:49
*** donghao has joined #openstack-meeting10:51
*** artom has joined #openstack-meeting10:53
*** oneswig has joined #openstack-meeting10:55
oneswigHi All, I'm proposing to cancel today's meeting.  I'm currently in a design workshop and can't give it due attention.10:56
*** markvoelker has joined #openstack-meeting10:57
*** donghao has quit IRC10:57
*** ociuhandu has joined #openstack-meeting10:58
*** janders has joined #openstack-meeting11:00
oneswigApologies, I'm in a workshop currently, can't chair the meeting11:01
*** ttsiouts has quit IRC11:01
*** markvoelker has quit IRC11:01
jandersno worries11:02
*** b1airo has joined #openstack-meeting11:02
jandershey Blair! :)11:02
b1airoHey hey11:02
b1airoHow goes it?11:02
oneswigHi b1airo janders11:02
janderswe hit 100GB/s on the GPFS cluster I'm working on today11:03
janders6 servers, 6 clients11:03
oneswigjanders: sheesh11:03
b1airoSounds like there's a story behind that claim janders ...11:03
oneswigGoing for IO500?11:03
jandershaha :)  maybe11:03
oneswigI have to head off I'm afraid.11:03
oneswigapologies, can't join you today.11:04
oneswiggotta go :-(11:04
*** oneswig has quit IRC11:04
janderswe'll chat later11:04
b1airoAnyone else along this evening?11:04
jandersBlair, do you have any experience with running IOPSy workloads on GPFS?11:04
b1airoI wouldn't say "much"11:06
b1airoWhatever it is probably works better inside a container though ;-)11:06
jandersI haven't reached that part yet, but it's in the pipeline11:06
b1airoWhat are you seeing with IOPS?11:07
jandersI'm seeing interesting IOPS performance profile - much of which is to do with the underlying hardware, but GPFS itself plays a role in this11:07
jandersthe NVMes this thing is build on can do 700k IOPS each11:07
jandersbut a single client can only see 70k against the entire fs11:07
jandersit's 70k per client, no matter how many clients11:08
jandersso on 6 we can go over 300k IOPS11:08
jandersGPFS block size might have something to do with this (we figured 1M should be reasonable for a general purpose fs)11:09
*** brault has joined #openstack-meeting11:09
jandersbut yeah I find it interesting - it seems that the single client is IOPS constrained, the filesystem - much less so11:09
b1airoDoesn't scale similarly with higher I/O queue depths?11:09
jandersI only tested on 32 queue depth and 16 threads11:10
janders(getting the fio file)11:10
*** njohnston has joined #openstack-meeting11:11
b1airoI wouldn't be surprised if GPFS's client side metadata caching ended up adding some bottlenecks that need to be tuned away to realise full potential of such a backend11:11
jandersvery good insight - will discuss with the team tomorrow11:12
jandersI've seen it causing side effects in the past (not specifically this though)11:12
b1airoI can read that fio config fine btw11:12
jandersanother interesting lesson learned is... HDR200 heat issues11:12
janderswe had some thermal shutdowns of the HCAs with some neighbouring 10GE adapters falling victim of the hotspots11:13
*** brault has quit IRC11:13
jandersCX6 warning temperature is 105C and shutdown temperature is 125C11:13
b1airoWhat wattage are they pulling?11:14
jandersWe ended up taking plastic bits out of the server (these maximise memory cooling at the expense of NICs/HCAs)11:14
jandersand fixing the fan speed to 80%11:14
jandersrunning at around 70C now11:14
jandersdown from 90-12011:14
jandersgood question.. I know how to check temperatures but not wattage - do you have a command handy?11:14
b1airoHa, resourceful!11:14
*** thgcorrea has joined #openstack-meeting11:16
b1airoNot sure if you can ask the HCA itself, would expect to dig in the system management interfaces11:16
janders[ansible@s23 ~]$ ansible spectrum -a 'mget_temp -d mlx5_1's204-ib3 | SUCCESS | rc=0 >>68             s202-ib3 | SUCCESS | rc=0 >>70             s205-ib3 | SUCCESS | rc=0 >>66             s201-ib3 | SUCCESS | rc=0 >>70             s203-ib3 | SUCCESS | rc=0 >>64             s206-ib3 | SUCCESS | rc=0 >>70             [ansible@s23 ~]$11:16
b1airoHmm, just remembered something I was going to investigate earliest today but didn't, so you're getting the question now :-) ...11:17
janderssure will try to help11:18
b1airoAny ideas if the memory cgroup can be configured to allow prioritised ... Hmmmm actually, think I just realised I should be looking at tuning OOM behaviour11:19
b1airoSometimes "verbalisation" of the problem solves it :-D11:19
jandersis OOM taking out things running in cgroups?11:20
b1airoI'll give you some context...11:20
b1airoWe've got some hugemem nodes that are particularly popular among genomics users11:20
b1airoThing is, they are generally pretty bad at estimating and optimising their memory requirements, and anyway, they're often dealing with algorithms and datasets that make it hard to predict correctly11:22
*** ttsiouts has joined #openstack-meeting11:22
b1airoBut our resources of this kind are scarce11:23
*** raildo has joined #openstack-meeting11:24
b1airoI was thinking about whether there'd be a cgroup based way to achieve a sort of memory overcommit (as these users usually over-estimate their memory limits quite a lot to avoid surprises 2 days into an assembly)11:24
b1airoSo we could let other workloads run and just ensure OOM priorities are set appropriately within that memory cgroup11:26
janderscrazy idea - wondering if some clever form of swap could help?11:26
jandersmaybe combined with custom nice value..11:27
b1airoPotentially, though hard to avoid swap activity hurting everyone on the node11:28
janderstrue but the nice value could differentiate that, perhaps11:28
b1aironiceness only impacts CPU time scheduling doesn't it? Wonder what the nice value for "yield on major fault that would cause demand pageout" ... :-)11:32
jandersyoure right (checked the doco)11:33
b1airoAnyway, I'll let you know if I (or more likely someone smarter that has the misfortune to work for me) come up with something11:33
b1airoYou heading to eResearch Australasia this year?11:33
jandersmemory.* bits might be relevant11:34
b1airoCan get you a talk invite for a storage BoF if you're interested..11:35
b1airoARDC Infrastructure Summit is on too11:35
jandersit would be nice, but with the current workload I don't think I will get approval to go anywhere before mid November unfortunately11:36
jandersI need the GPFS stuff to be wrapped up by mid Oct and I need to have OpenStack running against it by end October11:37
b1airo"running against it"..?11:37
b1airoWe talking Manila shares or something else entirely?11:38
jandersthis GPFS will be OpenStack's nova/glance/cinder backend - and parallel fs for the baremetal nodes, possibly11:38
jandersbut not sure if the fs part will be integrated with OpenStack yet - might be similar to Stig's BeeGFS approach11:39
janderswith the way things are going, we'll likely be helping users set their stuff up with Ansible, so if we take root off them in the process that'll probably be okay11:40
jandersgotta check where things are with Kerberos and GPFS11:40
b1airoPretty sure that's a working thing11:41
janderswill ask my GPFS guys about this11:42
janderswould be supernice11:42
*** markvoelker has joined #openstack-meeting11:42
b1airoI guess alternatively you could have on-demand GPFS-NFS protocol node exports into private network overlays11:43
*** brault has joined #openstack-meeting11:43
janderson a slightly different note, I noticed some interesting performance implications of using LSI expander cards11:43
jandersto plug in all the NVMes, I need some of those in passthru11:43
jandersI've got enough PCIe lanes for everything to be non blocking11:44
jandersand with read, this is indeed the case11:44
janderswith write, less so, there's a bit of a hit on bandwidth11:44
b1airoNVMe expanders this is?11:44
janderswith IOPS, it's quite signigicant, in order of 70%11:44
janders5e:00.0 RAID bus controller: LSI Logic / Symbios Logic MegaRAID Tri-Mode SAS3508 (rev 01)11:45
janders86:00.0 RAID bus controller: LSI Logic / Symbios Logic MegaRAID Tri-Mode SAS3408 (rev 01)11:45
b1airoThat's curious11:45
janderswith IOPS, I can sort of understand cause at least it's symmetrical. Capped to 200k read/write, end of story11:46
b1airoThey're not doing any silly caching are they?11:46
jandersbut with bandwidth - how come read is non blocking but write is?11:46
jandersI was assured they are just dumb devices providing extra ports11:46
jandersbut will chat to my hardware guys about this11:46
janders1) why read and write differ via the expander, when they do not differ without11:47
janders2) what's on with the IOPS11:47
jandersI suppose forcing NVMe to speak SAS might force them to communicate in sub-optimal ways11:47
jandersbut I do find this very curious11:48
b1airoI'll be interested to hear11:48
jandersit doesn't hurt performance much as we're capped with HDR200 bandwidth anyway11:48
b1airoI'm fading, almost midnight here. Might have to bail11:48
janders(and similarly, there's a cap on IOPS somewhere and it's not the expander)11:48
jandersI just realised I should benchmark IOPS on two NVMes hanging off the same expander...11:49
jandersno worries - thank you for the chat!11:49
janderstill next time11:49
b1airoCheers and likewise!11:49
*** brault has quit IRC11:50
*** belmoreira has quit IRC11:54
*** dviroel has joined #openstack-meeting11:56
*** belmoreira has joined #openstack-meeting11:58
*** yaawang has quit IRC12:03
*** janders has quit IRC12:06
*** rfolco has joined #openstack-meeting12:12
*** hyunsikyang has joined #openstack-meeting12:29
*** hyunsikyang__ has quit IRC12:32
*** mriedem has joined #openstack-meeting12:47
*** donghao has joined #openstack-meeting12:50
*** donghao has quit IRC12:54
*** lseki has joined #openstack-meeting12:55
*** psachin has quit IRC13:04
*** ociuhandu has quit IRC13:04
*** ociuhandu has joined #openstack-meeting13:07
*** ociuhandu has quit IRC13:11
*** belmoreira has quit IRC13:12
*** jamesmcarthur has joined #openstack-meeting13:19
*** belmoreira has joined #openstack-meeting13:23
*** brault has joined #openstack-meeting13:24
*** rossella_s has joined #openstack-meeting13:24
*** ykatabam has quit IRC13:24
*** yaawang has joined #openstack-meeting13:26
*** jamesmcarthur has quit IRC13:27
*** brault has quit IRC13:28
*** ociuhandu has joined #openstack-meeting13:31
*** liuyulong has joined #openstack-meeting13:31
*** ociuhandu has quit IRC13:34
*** ociuhandu has joined #openstack-meeting13:35
*** yaawang has quit IRC13:37
*** ociuhandu has quit IRC13:37
*** ociuhandu has joined #openstack-meeting13:37
*** ociuhandu has quit IRC13:39
*** whoami-rajat has quit IRC13:39
*** ociuhandu has joined #openstack-meeting13:40
*** zhangchi has joined #openstack-meeting13:46
*** ociuhandu has quit IRC13:49
*** zhangchi has quit IRC13:52
*** slaweq has quit IRC13:54
*** slaweq has joined #openstack-meeting13:57
*** panda is now known as panda|bbl13:59
*** yaawang has joined #openstack-meeting13:59
liuyulong#startmeeting neutron_l314:00
openstackMeeting started Wed Sep 25 14:00:19 2019 UTC and is due to finish in 60 minutes.  The chair is liuyulong. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
*** openstack changes topic to " (Meeting topic: neutron_l3)"14:00
openstackThe meeting name has been set to 'neutron_l3'14:00
liuyulong#chair haleyb14:01
openstackCurrent chairs: haleyb liuyulong14:01
*** b1airo has quit IRC14:01
*** liuyulong__ has joined #openstack-meeting14:02
liuyulongOooooK, let's ssstart.14:02
liuyulong#topic Announcements14:02
*** openstack changes topic to "Announcements (Meeting topic: neutron_l3)"14:02
liuyulongWe have a Projects Release Date: 16 October, 201914:04
liuyulongit will come soon14:05
liuyulongAnd I received many RC release mail.14:05
slaweqyes, we already released rc1 for many stadium projects14:06
*** brault has joined #openstack-meeting14:06
liuyulongThis is our train-rc1 related topics. IMO, they are in good shape now, we could catch up the scheduler.14:08
liuyulongCall for topics, summit is also coming soon, we will have 5 days meeting?14:11
slaweqptg starts on Wednesday14:11
slaweqMonday - Wednesday is summit and Forum14:12
liuyulongGood, so 3 days, another thing is that we will also hold the meeting online in parallel?14:13
slaweqliuyulong: I sent You an email recently about that14:14
slaweqI will probably need Your help to organize some streaming14:14
slaweqbut we can sync later with that14:14
liuyulongAnd mlavalle send us the mail to query if there will have a nice tool for those who can not go to Shanghai.14:15
*** slaweq has quit IRC14:15
*** slaweq has joined #openstack-meeting14:16
slaweqliuyulong: yes, but I will now take care of it instead of mlavalle14:16
*** ociuhandu has joined #openstack-meeting14:16
liuyulongWechat is widely used in China. So I may want to give a try for our L3 team members to test some tools before the PTG.14:18
slaweqliuyulong: will You be available tomorrow morning my time? So afternoon Your time14:18
slaweqwe can sync about it then14:18
slaweqor on Friday if You want14:18
liuyulongAnd a Proxy Server can also be another choice, so I wonder if there are any free server for us to user from the community?14:19
liuyulong__Backup is coming...14:19
liuyulong__Seems that connection is broken...14:20
*** ociuhandu has quit IRC14:20
liuyulong__slaweq, sure, I will ping you tomorrow. Friday can also be fine.14:22
slaweqliuyulong__: great14:22
liuyulong__All right, I have no more announcement. If you have please go ahead.14:23
liuyulong__My irc connection is unstable. Let's speedup.14:25
liuyulong__#topic Bugs14:25
*** ociuhandu has joined #openstack-meeting14:25
*** liuyulong has quit IRC14:26
*** liuyulong has joined #openstack-meeting14:26
liuyulong#topic Bugs14:27
*** openstack changes topic to "Bugs (Meeting topic: neutron_l3)"14:27
liuyulong#chair liuyulong__14:27
openstackCurrent chairs: haleyb liuyulong liuyulong__14:27
openstackLaunchpad bug 1845150 in neutron "[FT] "keepalived" needs network interfaces configured as in its own config" [High,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)14:27
ralonsohthe patch is here14:28
ralonsohas commented in the patch, if keepalived starts and doesn't see the tracked interfaces with the needed ips14:28
liuyulongralonsoh, thanks, it looks good to me.14:28
ralonsohthe process stops14:29
ralonsohok, thanks! thats all14:29
liuyulonghaleyb have +wed it :)14:29
openstackLaunchpad bug 1844688 in neutron ""radvd" daemon does not work by default in some containers" [Medium,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)14:30
ralonsohyes, thanks for the comment14:30
liuyulongThe fix:
ralonsohI need to add a reno14:30
ralonsohand test cases14:30
*** ociuhandu has quit IRC14:31
liuyulongA new config is added for this bug, so IMO, a release note is needed. Users should know such config option in many ways.14:31
*** ociuhandu has joined #openstack-meeting14:31
ralonsohfor sure, today I'll push a new patch14:31
liuyulongralonsoh, np14:32
openstackRemoving item from minutes: #link
openstackLaunchpad bug 1845227 in neutron "new l3 agent router factory breaks neutron-fwaas functional tests" [Critical,Fix committed]14:33
ralonsohfixed here #link
*** rpittau is now known as rpittau|afk14:34
liuyulongA Critical bug, and it is fixed.14:34
*** diablo_rojo has joined #openstack-meeting14:34
liuyulongLast one from me14:35
openstackLaunchpad bug 1845145 in neutron "[L3] add abilitiy for iptables_manager to ensure rule was added only once" [High,In progress] - Assigned to LIU Yulong (dragon889)14:35
liuyulongBasically I did not see that WARNING log from our local queens deployments. But when create new agent extension with parallel processing, I noticed many warning log.14:38
liuyulongIn order to ensure the reentrancy of rule addition, I raised such bug.14:39
slaweqliuyulong: but doesn't it means that Your agent extension is doing something wrong?14:39
liuyulongslaweq, we can add more code to prevent that, but it will need more cache and event more lock to achieve that. A simple reentrancy or idempotent can be a nice approach.14:42
liuyulongIMO, iptables_manager should have that ability.14:42
slaweqI wonder why it wasn't dont like that in the past14:43
liuyulongI did not find any reason that we want the duplicated rules in that managers rule cache.14:44
slaweqmaybe haleyb or mlavalle will remember something14:44
* haleyb sticks his head up14:44
liuyulongThe fix is quite simple.14:45
haleybslaweq: we've never needed that before, i don't know why it's happening now14:45
haleybthat's why i was wondering if something outside neutron was adding a rule, or if we were calling something twice14:47
slaweqhaleyb: IIUC Neutron U/S don't need it still but liuyulong is doing some d/s agent extension which needs it to make his life easier14:48
liuyulonghaleyb, yes, that upstream log search can approve that.14:48
liuyulongLOG a warning does not make sense to me, we are neutron developers, so we can make it better.14:48
haleybright, i'm just wondering how that ironic devstack setup is causing issues, since that seems to be where that port rule is added, not via the neutron api14:49
liuyulongThis fix should still use list() instead of set(), because the rule needs the proper order.14:50
haleybbasically i'm saying it would be good to better understand how it's happening before adding a duplicate check14:50
liuyulongAnd that's why the patch set 2 has functional test failures.14:50
liuyulongI will run test locally and update it soon.14:51
*** enriquetaso has quit IRC14:51
liuyulongSorry, I have to run now. Let's stop here.14:51
liuyulongThank you for your participation.14:52
liuyulongand sorry for my bad network connection...14:52
*** openstack changes topic to "OpenStack Meetings ||"14:52
openstackMeeting ended Wed Sep 25 14:52:41 2019 UTC.  Information about MeetBot at . (v 0.1.4)14:52
openstackMinutes (text):
*** liuyulong__ has quit IRC14:53
*** liuyulong has quit IRC14:53
*** brault has quit IRC14:57
*** e0ne has quit IRC14:59
*** brault has joined #openstack-meeting14:59
*** e0ne has joined #openstack-meeting14:59
*** brault has quit IRC15:00
*** belmoreira has quit IRC15:05
*** brault has joined #openstack-meeting15:10
*** zhubx has quit IRC15:13
*** ttsiouts has quit IRC15:13
*** boxiang has joined #openstack-meeting15:14
*** eharney has quit IRC15:14
*** artom has quit IRC15:15
*** artom has joined #openstack-meeting15:15
*** ttsiouts has joined #openstack-meeting15:15
*** ttsiouts has quit IRC15:15
*** ttsiouts has joined #openstack-meeting15:20
*** sridharg has joined #openstack-meeting15:23
*** ttsiouts has quit IRC15:24
*** ekcs has joined #openstack-meeting15:28
*** ekcs has quit IRC15:29
*** pcaruana has quit IRC15:30
*** gyee has joined #openstack-meeting15:35
*** whoami-rajat has joined #openstack-meeting15:37
*** macz has joined #openstack-meeting15:40
*** markvoelker has quit IRC15:45
*** macz has quit IRC15:47
*** rosmaita has joined #openstack-meeting15:51
*** davidsha has joined #openstack-meeting15:52
*** _pewp_ has joined #openstack-meeting15:53
*** woojay has joined #openstack-meeting15:53
*** lbragstad_ has joined #openstack-meeting15:54
*** lbragstad has quit IRC15:55
*** lbragstad has joined #openstack-meeting15:57
*** lbragstad_ has quit IRC15:58
jungleboyj#startmeeting Cinder16:00
openstackMeeting started Wed Sep 25 16:00:23 2019 UTC and is due to finish in 60 minutes.  The chair is jungleboyj. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
*** openstack changes topic to " (Meeting topic: Cinder)"16:00
openstackThe meeting name has been set to 'cinder'16:00
jungleboyjcourtesy ping:  jungleboyj whoami-rajat rajinir lseki carloss pots woojay erlon geguileo eharney rosmaita enriquetaso e0ne smcginnis davidsha walshh_ xyang hemna _hemna tosky sfernand16:00
_pewp_jungleboyj ヾ(-_-;)16:01
*** carloss has joined #openstack-meeting16:01
jungleboyjOh pewp bot.  I guess you know how my week is going.16:01
rosmaitae0ne: you have been a reviewing maniac, thank you16:02
jungleboyje0ne:  ++16:02
jungleboyjThough the mental image of e0ne being a maniac is a little scary.16:03
jungleboyjOk.  It looks like we have the usual crowd here.16:04
jungleboyj#topic announcements16:04
*** openstack changes topic to "announcements (Meeting topic: Cinder)"16:04
jungleboyjNot a lot to say other than the fact that we are at RC week for Train.16:04
rosmaitathere's nothing unreviewed on there, just things that need another review16:05
jungleboyjOooh, look at the PTL to be being organized!16:05
* jungleboyj is happy16:05
rosmaitahopefully the stuff in the gate will merge OK16:05
rosmaitathere's a small problem with the default types patch, i think16:05
jungleboyjI will take a look through all of these after the meeting.16:05
rosmaitathe release note needs a revision, but that patch is so big, i don't want to request a change16:06
jungleboyjrosmaita:  I see you comments.16:06
rosmaitaso i think it would be good to merge it and then do a followup16:06
jungleboyj@whoami-rajat  Any concerns with the plan there?16:06
*** rfolco is now known as rfolco|bbl16:06
rosmaitabut it will have to be quick because once the stable branch is created, the release notes behave weirdly16:06
rosmaitamy proposal would be to propose the RC-1 release tomorrow afternoon?16:07
whoami-rajatrosmaita: jungleboyj that sounds good, thanks, i've been trying to get zuul+1 since past few days and if the update is releasenotes related, i would be glad to do it in a followup16:07
rosmaitawhich is a day later than usual for cinder, but i think it is still within legal time16:07
e0nerosmaita, jungleboyj: thanks. I'm doing my best to make cinder better16:07
jungleboyjwhoami-rajat:  Ok.  Cool.  Lets do that instead of complicating things.16:08
rosmaitae0ne: ++16:08
rosmaitawhoami-rajat: that would be great16:09
rosmaitai will put more notes on your current patch about what i think is needed16:09
jungleboyjAre there any bug fixes that are not on Brian's list?16:09
whoami-rajatjungleboyj: rosmaita thanks, also i would like to ask when is cinderlib release planned?16:09
e0nejungleboyj, rosmaita, smcginnis: fyi,
rosmaitai didn't look at the unmerged changes, i just pulled the FFEs and bugs i knew about16:10
*** yonglihe has quit IRC16:10
rosmaitawhoami-rajat: cinderlib is a cycle-trailing release, so probably within a month after actual Train release16:10
rosmaitait can wait longer, but i think gorka will want to release as soon as it makes sense16:11
jungleboyje0ne:  Thank you.  I just put a hold on that.16:11
rosmaitayes, thanks!16:11
whoami-rajatrosmaita: great, i've a dependent patch of the same feature in cinderlib so i think it will be in on time.16:11
walshh_I have a few driver rev16:11
walshh_I have a few driver reviews not on the list above16:12
jungleboyjwalshh_:  I looked at the one this morning you sent.  Can you add yours to the list in the etherpad?16:12
jungleboyjThat way everyone can take a look.16:12
jungleboyjNot just me.  :-)16:12
walshh_thank you.  Just a few more, all have clean CI now, minor fixes16:12
jungleboyjOkie Dokie.16:12
jungleboyjBug fixes are good.16:12
jungleboyjMust mean people are actually using your driver?  :-)16:13
walshh_one or two :-)16:14
jungleboyjThat is a good thing.16:14
jungleboyjOk, if you can add your reviews in the etherpad that would be great and we will review.16:14
walshh_I will, thank you16:15
jungleboyjOtherwise testing of the RC would be good if anyone is able to do that.16:15
rosmaitaok, i am in prague time zone this week, so let's say to be in RC-1, stuff must be merged by 22:00 UTC tomorrow (26 Sept)16:15
rosmaitai will also commit to looking at walshh_'s patches16:15
* jungleboyj would like to go to Prague some time.16:15
hemnaany pressing reviews?16:16
e0nejungleboyj: +116:16
jungleboyjWhere you been man?  ;-)16:16
hemnathis job thing....16:16
jungleboyjhemna: ^^16:16
jungleboyjhemna: Yeah, I feel your pain.16:16
hemnakinda gets in the way of cinder sometimes16:16
jungleboyjrosmaita:  Anything else on RC discussion?16:17
rosmaitai don't mean to be a jerk, but i think we need someone to commit to reviewing the default types patch16:17
rosmaita(it's a bit of a monster)16:17
rosmaitacode looks pretty clean, just a lot of files are touched16:17
hemnayah that's a big one the last time I looked16:17
jungleboyjrosmaita:  I have been through it before.16:17
jungleboyjSo, not sure that I am going to have additional input.16:18
rosmaitai don't think it has changed too much since your last comments16:18
hemnaso this is a new feature16:18
hemnaare we ok with merging this at this stage of the game?16:18
rosmaitayes, it got a FFE last week16:18
hemnawhat could go wrong?16:18
davee__It could provide an excuse to drink lots of caffeine16:19
rosmaitawell, if people have set CONF.default_type already, i don' think there will be any problems16:19
whoami-rajathemna: it's adding a volume type with db migration, migrating all existing untyped volumes/snapshots to the default type, that's it :)16:19
jungleboyj:-)  What could possibly go wrong.  :-)16:20
hemnawell if we are going to merge it, it's best to merge early16:20
hemnaand test it16:20
hemnawe can always back it out...16:20
whoami-rajatrosmaita: that's the aim :)16:20
jungleboyjhemna:  Agreed.16:20
*** brault has quit IRC16:20
rosmaitaor, more positively, we can backport a minor bugfix :)16:21
jungleboyjrosmaita:  Agreed.16:21
hemnaok I'll take a look today16:21
jungleboyjhemna:  Thank you.16:22
rosmaitahemna: ty!16:22
jungleboyjIs there anyone planning to test it?16:22
whoami-rajathemna: thank you!16:22
rosmaitai won't have time until after RC to really test it out16:22
*** ociuhandu has quit IRC16:22
hemnaso we need to create a bunch of vols/snaps w/o a type16:22
*** priteau has quit IRC16:23
hemnaI presume there is a migration script that has to be run to migrate those ?16:23
*** ociuhandu has joined #openstack-meeting16:23
whoami-rajathemna: the type migration of existing vols/snaps is done with online migration16:24
*** mattw4 has joined #openstack-meeting16:25
jungleboyjwhoami-rajat:  Is there a way to do it without having to install Stein and try to upgrade?16:25
jungleboyjTo test it?16:26
whoami-rajatjungleboyj: we can run the db migrations (cinder-manage db sync) and online migrations (cinder-manage db online_data_migrations)16:27
jungleboyjOk.  Cool and then it should deal with the untyped volumes?16:27
*** electrofelix has quit IRC16:27
*** ociuhandu has quit IRC16:28
whoami-rajatjungleboyj: untyped volumes and snapshots and if any new volume is created without any provided type will be assigned `__DEFAULT__`16:28
jungleboyjwhoami-rajat:  Cool.16:28
jungleboyjWill try to test but not promising I will get to it right away.16:29
rosmaitawhoami-rajat: you aren't going on vacation during the next 2 weeks, are you?16:30
jungleboyjrosmaita:  :-)  Good question.16:30
hemnaok so most of the filed changed in that patchset are unit tests16:30
hemnate rest looks ok I guess16:30
whoami-rajatjungleboyj:  thanks for the help :)16:30
whoami-rajatrosmaita: no16:31
rosmaitaok, great16:31
jungleboyjOk.  That is good.16:31
rosmaitahemna: before you arrived, we agreed to do the release note changes in a follow up patch16:31
rosmaitathe gate has been really slow, don't want to hold up the main patch16:31
hemnaugh ok16:31
hemnaok, it's on it's way16:32
whoami-rajatjungleboyj: rosmaita hemna thanks for the reviews!16:33
jungleboyjhemna:  Thanks.16:33
jungleboyjwhoami-rajat:  Let me know when you have the updates for the release note done.16:33
rosmaitawhoami-rajat: what jungleboyj said16:34
whoami-rajatjungleboyj: rosmaita sure, will do it quickly after the meeting16:34
rosmaitathanks, please add it to the etherpad when you post it16:35
jungleboyjSounds good!16:35
whoami-rajatrosmaita: ok16:35
whoami-rajatwill do16:35
jungleboyjAdded notes to the patch on what we have discussed today for documentation purposes.16:35
jungleboyjSo, anything else we need to discuss?16:36
rosmaitanothing from me16:36
jungleboyjA friendly reminder to add topics to the etherpad for the Stein PTG:
whoami-rajatjungleboyj: i think there is one thing to discuss16:36
jungleboyjwhoami-rajat:  Ok.  Go ahead.16:37
whoami-rajatit's regarding the tempest schema validation16:37
jungleboyjOh yeah.16:37
whoami-rajatit seems to be breaking cinder in stable branches16:37
rosmaitawell, that is because they aren't testing first16:37
rosmaitait happened in O and now it's happening in P.16:37
*** markvoelker has joined #openstack-meeting16:38
rosmaitaguess it will happen in Q next week16:38
whoami-rajati checked the cinder stable/pike code and seems fine16:38
whoami-rajatrosmaita: hah :)16:38
rosmaitayes, my vote is we ignore it, matt has a patch up to pin tempest16:38
rosmaitahopefully they will pin Q also before anything breaks16:38
rosmaitajust realized that people may not know what i am talking about16:40
whoami-rajatthe question is, is the schema validation really necessary for stable branches ? a lot of code is added removed until the introduction of MV's so IMO it will be valid to do the testing after v316:41
whoami-rajatrosmaita: i should've started with the thread link, thanks16:42
jungleboyjrosmaita:  We had asked why they were doing this right?16:42
jungleboyjI don't remember the answer though.16:42
rosmaitawell, i thought eharney had a well-reasoned discussion going on this last time16:42
rosmaitabut it wasn't really followed up on16:42
rosmaitaso, in my capacity of being a general PITA, i am inclined to not worry about this and let the qa team keep breaking the stable branches and then having to pin tempest to fix it16:43
rosmaitabut it is distressing that they are merging stuff without checking to see what they will break16:44
whoami-rajatthere are different reasons for failures, last one was because of not doing an MV bump/ testing a dynamic response in tempest16:44
rosmaitaand with that, i will shut up16:44
whoami-rajatrosmaita: seems valid16:44
jungleboyjrosmaita:  ++16:45
*** igordc has joined #openstack-meeting16:46
jungleboyjSo, any further discussion required there?16:47
whoami-rajatnot from my side16:47
jungleboyjOk cool.16:48
*** markvoelker has quit IRC16:48
jungleboyjFor those who didn't see, they have announced that the next event after Shanghai will be a PTG like event in Vancouver.16:48
* jungleboyj is happy!16:48
*** rosmaita has left #openstack-meeting16:49
jungleboyjThen a summit like event later next year.  Location still not determined.16:49
e0nejungleboyj: sounds good16:49
*** zhangchi has joined #openstack-meeting16:49
jungleboyjAlso, if you know anyone that might be new to OpenStack and in Shanghai please let them know about OpenStack Upstream Institute.16:50
jungleboyjWould like to see more interest there.16:50
whoami-rajatjungleboyj:  ++16:51
jungleboyjOk.  That is all I had for this week then.  Please go review the outstanding train patches.16:52
*** markvoelker has joined #openstack-meeting16:53
whoami-rajatjungleboyj: Thanks!16:54
jungleboyjYep.  Thanks everyone.16:54
*** openstack changes topic to "OpenStack Meetings ||"16:55
openstackMeeting ended Wed Sep 25 16:55:03 2019 UTC.  Information about MeetBot at . (v 0.1.4)16:55
openstackMinutes (text):
*** ekcs has joined #openstack-meeting16:55
*** davidsha has quit IRC16:55
*** zhangchi has quit IRC16:55
jungleboyjSorry for being distracted this week.  :-)16:55
*** woojay has left #openstack-meeting16:56
*** brault has joined #openstack-meeting16:56
*** e0ne has quit IRC16:59
*** mattw4 has quit IRC17:04
*** ociuhandu has joined #openstack-meeting17:09
*** sridharg has quit IRC17:14
*** ociuhandu has quit IRC17:16
*** panda|bbl is now known as panda17:19
*** mattw4 has joined #openstack-meeting17:44
*** njohnston has quit IRC17:51
*** igordc has quit IRC17:52
*** njohnston has joined #openstack-meeting17:56
*** njohnston has quit IRC17:59
*** tesseract has quit IRC18:08
*** armax has quit IRC18:18
*** e0ne has joined #openstack-meeting18:25
*** ekcs has quit IRC18:43
*** armax has joined #openstack-meeting18:50
*** ricolin has quit IRC18:55
*** belmoreira has joined #openstack-meeting18:55
*** gary_perkins has quit IRC18:57
*** gary_perkins has joined #openstack-meeting18:57
*** belmoreira has quit IRC18:58
*** ociuhandu has joined #openstack-meeting19:17
*** e0ne has quit IRC19:23
*** ociuhandu has quit IRC19:25
*** zhangchi has joined #openstack-meeting19:52
*** zhangchi has quit IRC19:57
*** radeks has quit IRC20:12
*** diablo_rojo has quit IRC20:24
*** ykatabam has joined #openstack-meeting20:29
*** ralonsoh has quit IRC20:30
*** slaweq has quit IRC20:55
*** ekcs has joined #openstack-meeting20:59
*** alecuyer has joined #openstack-meeting21:01
timburke#startmeeting swift21:02
openstackMeeting started Wed Sep 25 21:02:15 2019 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.21:02
*** openstack changes topic to " (Meeting topic: swift)"21:02
openstackThe meeting name has been set to 'swift'21:02
timburkewho's here for the swift meeting?21:02
*** zaitcev has joined #openstack-meeting21:03
*** igordc has joined #openstack-meeting21:03
timburkeagenda's at
timburkei've got more stuff ahead of updates than usual, so i'll try to hustle :-)21:04
timburke#topic User survey results21:04
*** openstack changes topic to "User survey results (Meeting topic: swift)"21:04
*** raildo has quit IRC21:04
timburkethis was actually available last week, but i finally got around to looking at them with some amount of depth21:04
timburkewe only had a couple swift-specific questions: what features are you using, and a free-form "Other feedback" field21:05
timburkewe didn't get any free-form comments, but i still saw some interesting things21:06
timburke62 of 342 respondents included feedback about which Swift features they use21:06
*** diablo_rojo has joined #openstack-meeting21:06
mattoliverauand... what did they use?21:07
timburkebut there were an additional 6 who "contribute maintenance resources, such as patches for bug fixes and code reviews on master or stable branches" to Swift without mentioning which Swift features they use21:07
timburketop 5 features:21:07
timburkeS3 compatibility (54)21:07
timburkeQuotas (35)21:07
timburkeObject versioning (26)21:07
timburkeMulti-region clusters (26)21:07
timburkeTemporary URLs (25)21:08
timburkeso i feel rather vindicated in the recent focus on S3 versioning ;-)21:08
mattoliverauyeah, nice one21:08
tdasilvamulti-region == 26!!21:09
timburketdasilva, yeah, definitely interesting21:09
timburkethere were other questions posed (by the TC, i think?) asking about stable branch usage21:09
timburke54 of 62 Swift respondents use stable branches to some degree, with 28 "using only official point releases"21:09 i should maybe tag more stable releases21:10
rlediseztdasilva: 26 means 26 clusters in 26 regions or 1 cluster/1 ring with 26 "zones/regions" ?21:10
timburke26 of the 62 respondents said they "use" multi-region clusters21:11
tdasilvarledisez: i have no idea, i'm hoping it's 26 different people, meaning 26 different clusters21:11
rledisezah, ok :D21:11
rledisezi was wondering who would do a 26-regions ring ;)21:11
timburkeanyway, thought i would share. i'll make sure i get the full tally of feature usage up somewhere for the curious21:12
timburke(or you can grab the dataset yourself and take a look ;-)21:12
tdasilvatimburke: i'd argue we should almost do a stable branch tag almost every time we do a stable backport (or a string of backports)21:12
timburkeyeah... i kinda agree21:12
tdasilvalike if there's nothing else left in the pipeline, tag21:12
mattoliverauor maybe at least every time we do a release we should see if there has been any backports and if so.. tag21:13
timburkei think the biggest hurdle has been a sense that "oh, if i'm proposing a tag, i should probably see what other bug fixes maybe ought to get backported..."21:13
timburkemattoliverau, definitely. one of the nice things about hand-curating the release notes is that i get a whole bunch of patches and their impacts loaded into my head anyway, which can then be used on my previous concern21:14
timburke#topic Vancouver21:15
*** openstack changes topic to "Vancouver (Meeting topic: swift)"21:15
mattoliverau\o/ back to YVR21:15
timburkei know we haven't even gotten to Shanghai yet, but apparently we're going back to Vancouver for the next OpenStack event in June21:15
zaitcevWhat's in Vancouver? Also, which Vancouver?21:15
zaitcevreading now21:16
* rledisez loves Vancouver21:16
timburkerledisez, and your flight got so much shorter!21:16
rledisezyeah, makes it even better :)21:16
timburkeyou all now know as much as me about all this :-) i'm sure we'll find out more in the coming months21:17
timburke#topic Shanghai21:17
*** openstack changes topic to "Shanghai (Meeting topic: swift)"21:17
timburkejust a quick reminder again to put topic on the etherpad21:18
timburke(i need to do it, too)21:18
timburke#topic train release21:18
*** openstack changes topic to "train release (Meeting topic: swift)"21:18
timburkeso within the next week or so, i'd like to tag a 2.23.021:18
timburkei updated the priority reviews wiki to highlight a few patches i'd like to get in if we can21:19
*** ekcs has quit IRC21:19
timburkethe top two in particular would be good -- and one of them is actually a bug that still needs a patch21:20
timburkespecifically, those have to to with our new py3 support -- and one of them is a (now public) security bug21:21
timburkeif you know of any other patches that ought to get in, feel free to update the wiki. i'll try to get a patch up for in the next couple days21:22
openstackLaunchpad bug 1844368 in OpenStack Object Storage (swift) "fallocate_reserve cannot be specified as a percentage running under python3" [High,Confirmed]21:22
zaitcevIf anyone knows anything about eventlet monkey-patching, there21:22
timburkemattoliverau or zaitcev, you've done a good bit with py3 -- would you be able to review ?21:23
zaitcev's also this:
zaitcevtimburke: will do21:23
timburkethanks. and i can try to take a look at the triple-o patch -- sounds kinda like the issue i hit in where the module that i wanted patched was imported before i realized that it would be21:25
timburkeok, updates!21:26
timburke#topic versioning21:26
*** openstack changes topic to "versioning (Meeting topic: swift)"21:26
zaitcevAs you can see, it is not a review in Swift. The story is, Swift requires eventlet 0.25 for py3. So, I made RDO update the required eventlet. Result is, TripleO fell apart, because the module git has issues with eventlet. It's a mad word, but if anyone can understand what happend, they could add +1.21:26
*** ekcs has joined #openstack-meeting21:26
timburketdasilva, what's the state of the world?21:26
*** whoami-rajat has quit IRC21:27
tdasilvawith the new object versioning patch per se, we've been making progress adding some unit and functional tests21:27
tdasilvathere are still some corner cases to discuss the design about. e.g., how to handle error scnearios of creating versions container, how to handle versions and overwrites when versioning is suspended21:28
tdasilvabut it's coming along. For now we are also punting on the naming scheme, as we have started discussion on the null character piece21:29
claygdid i miss the whole thing?21:30
tdasilvaclayg: nope21:30
timburkeclayg, i think you're right on time :-)21:30
tdasilvatimburke: so we've made progress, but there's still quite a bit of work to do :)21:31
timburkei know we've got some etherpads for both versioning and the null-byte ideas21:31
tdasilvaif folks can keep an eye and contribute to both the discussion on the etherpad and the patches listed here, it would be super helpful:
mattoliverauoh i didn't realise there was a null namespace one. I'll read it today21:32
timburkemattoliverau, i think clayg added it today ;-)21:33
claygyeah that's new fresh off the presses21:33
mattoliverauoh, well then I was alseep anyway :P21:33
kota_interesting, the null namespace idea21:33
tdasilvakota_: it's revolutionary! :D21:34
kota_it seems like similar idea s3api (swift3) uses `+` for versioning namespace that character can not be used by users.21:35
timburkeclayg, tdasilva: anything in particular you'd like input on this week?21:35
timburkekota_, i think that may be where we got the idea :D21:35
claygkota_: yes, lots of s3 features and s3api implementation rely on resources the user can't access except through approved apis!21:35
claygtdasilva: do we have the versioning api draft anywhere?21:36
claygsomething that might eventually become the docs?21:37
tdasilvatimburke: yeah, i think in terms of order, the null character would come first?? it would probably be good to get a good design on that21:37
tdasilvaclayg: no yet, /me needs to get started on that :(21:37
claygtdasilva: right, everyone please read and comment on the null-namespace etherpad - I'm going to be implementing what's in there so expect to be doing some early review stuff soon21:37
claygtdasilva: i was thinking more fast/dirty wiki style for folks to get an idea - but... I guess we have the s3 api for that 🙄21:38
tdasilvaclayg: yeah, I was thinking about adding to the etherpad to begin with, it makes for editing very easy21:39
tdasilvaby anyone21:39
claygtdasilva: I had an idea about versioned_writes - that maybe we call the new mode "object versioning" and the old one was *truly* "just" versioned_writes (i.e. no versioned reads, or restore, or listing)21:39
claygmight help us avoid calling the "object versioning" feature "new versioned writes"21:39
claygor maybe.. there's the "object versions api" and "legacy versioned writes" or something - idk - i get hung up on names21:40
claygtdasilva: hell yeah, etherpad it up!21:40
timburkeFWIW, i feel like the S3 docs aren't always sufficient -- exploratory things like tdasilva's are certainly useful, too21:40
tdasilvaclayg: i've been really scratching my head on naming, open to suggestions21:41
tdasilvatimburke: yeah, also found out from testing that s3 doesn't allow POSTing to older versions, I haven't seen that documented21:42
timburkeclayg, would it be worth breaking out to its own patch as a lace for us to document what AWS does?21:42
timburketdasilva, GTK21:42
claygtdasilva: yeah i breifly let myself think that we might back off on allowing POST to static links via versioning and instead stick with our re-direct behavior (i.e. allow POST key?version=)21:43
timburkeall right, i still want to check in on other endeavors -- maybe we can continue this in -swift21:43
timburke#topic lots of small files21:44
*** openstack changes topic to "lots of small files (Meeting topic: swift)"21:44
claygtimburke: I guess that'd be ok just for organization - I don't think we need/want s3api tests that fail against swift merged to master?21:44
timburkeclayg, fair enough21:44
timburkealecuyer, i saw another patch this morning!21:44
timburkean updated version of
alecuyeryes ! I'm getting to the patches that were pending,21:45
timburkehave you had a chance to sort out the bugs found when trying to deploy to prod?21:45
alecuyerI've updated it to work with HTTP21:45
alecuyerso about the bug, the good news, it's unrelated to the "rpc over http" patch21:45
timburke👍 (on both messages)21:45
alecuyerwe quickly tacked on a patch to bring back the hashes.pkl - and hum the "hang" was actually that the algorithm cost would go up with the square of the suffixes below the partition21:46
alecuyerso, hm, that was quickly reverted21:46
alecuyerand I'll look at it later. But, no issues found with the HTTP patch itself21:46
timburkeis there anything we can be helping with? i'll look at the py3 failures21:47
alecuyerthanks ! haven't looked at it yet but I will. So yes that would be great if you could take a look at the patch21:48
alecuyerthen I'll go back to the other pending patches (they are smaller than this one)21:49
timburkesounds good, thanks!21:49
timburke#topic sharding21:49
*** openstack changes topic to "sharding (Meeting topic: swift)"21:49
timburkelots of cleave-context cleanup!21:50
mattoliveraulol, yeah, I felt guilty that we weren't using an existing last-modified timestamp and adding a new one which would be annoyting for NM and needless storage.. So fixed that.21:50
mattoliverauThen clayg went and cleaned it up even further!21:51
timburkethanks for that, i think it's a bunch better :-)21:51
timburkemattoliverau, is there anything else there that you want to bring up, or do i just need to get to reviewing more patches? ;-)21:51
claygminor stuff, paper cuts21:51
claygis there more to review!?  😬21:51
timburkealways :D21:51
timburke for the empty handoff problem21:52
timburke to latch stats reporting21:52
mattoliveraunope, nothing to bring up. but reviews welcome. I'll review timburke's sharding patches21:52
timburkeand of course, the autosharder chain:
timburke#topic open discussion21:53
*** openstack changes topic to "open discussion (Meeting topic: swift)"21:53
timburkeanything else people would like to bring up?21:53
zaitcevNot me21:53
timburkeone thing i thought of -- the last time we thought much about our current meeting time, there were more of us in SF -- does this time slot still work well for people?21:55
mattoliveraudaylight savings changes in a few weeks for me then it'lll be 8am.. a much more civilised time :P21:56
timburkesomething to think about, anyway. it's hard with the spacing of US, Europe, and Japan/Australia21:56
alecuyerI'm fine with the time - similar to mattoliverau, in winter time it'll be 10pm vs 11 :)21:57
timburkeon that note, let's let alecuyer go to bed ;-)21:57
timburkethank you all for coming, and thank you for working on Swift!21:57
*** openstack changes topic to "OpenStack Meetings ||"21:57
openstackMeeting ended Wed Sep 25 21:57:42 2019 UTC.  Information about MeetBot at . (v 0.1.4)21:57
openstackMinutes (text):
mattoliverauthanks timburke21:57
*** zaitcev has left #openstack-meeting21:57
*** rfolco|bbl is now known as rfolco22:09
*** rcernin has joined #openstack-meeting22:15
*** diablo_rojo has quit IRC22:16
*** diablo_rojo has joined #openstack-meeting22:16
*** thgcorrea has quit IRC22:27
*** ekcs has quit IRC22:31
*** ekcs has joined #openstack-meeting22:49
*** mriedem has quit IRC22:53
*** zhangchi has joined #openstack-meeting22:55
*** zhangchi has quit IRC23:09
*** igordc has quit IRC23:10
*** gyee has quit IRC23:44
*** carloss has quit IRC23:47
*** mattw4 has quit IRC23:50

Generated by 2.15.3 by Marius Gedminas - find it at!