*** tonyb has quit IRC | 00:16 | |
openstackgerrit | Bradley Jones proposed openstack/cinder master: [DNM] Testing Cisco Third party CI https://review.openstack.org/588411 | 00:22 |
---|---|---|
*** edmondsw has joined #openstack-cinder | 00:37 | |
*** edmondsw has quit IRC | 00:42 | |
openstackgerrit | Merged openstack/cinder master: NetApp ONTAP: Fix driver force detach operations https://review.openstack.org/588082 | 00:58 |
*** mriedem_afk has quit IRC | 01:34 | |
*** tonyb has joined #openstack-cinder | 02:22 | |
*** edmondsw has joined #openstack-cinder | 02:25 | |
*** edmondsw has quit IRC | 02:30 | |
openstackgerrit | zhufl proposed openstack/cinder master: Fix none-ascii char in doc https://review.openstack.org/588424 | 02:33 |
*** psachin has joined #openstack-cinder | 02:34 | |
openstackgerrit | yenai proposed openstack/cinder master: Add multiattach attribute when do backup restore https://review.openstack.org/587667 | 02:35 |
*** dave-mccowan has quit IRC | 02:35 | |
*** itlinux_ has joined #openstack-cinder | 03:31 | |
openstackgerrit | caixiaoyu proposed openstack/cinder master: Add test case: create volume from source volume https://review.openstack.org/588431 | 03:32 |
openstackgerrit | Boxiang Zhu proposed openstack/cinder master: swap out easily readable strings for constants https://review.openstack.org/588189 | 03:55 |
*** dklyle has quit IRC | 04:39 | |
*** gnufied has quit IRC | 05:27 | |
openstackgerrit | Boxiang Zhu proposed openstack/cinder master: swap out easily readable strings for constants https://review.openstack.org/588189 | 06:10 |
*** Luzi has joined #openstack-cinder | 06:13 | |
openstackgerrit | Kumar Prashant proposed openstack/cinder master: VMAX Driver - Fix for multiple clones failure https://review.openstack.org/578486 | 06:19 |
openstackgerrit | cheng li proposed openstack/cinder master: Using "availability_zone" property when restoring https://review.openstack.org/588447 | 06:24 |
*** whoami-rajat__ has quit IRC | 06:34 | |
*** rcernin has quit IRC | 06:54 | |
*** kaisers has quit IRC | 07:03 | |
openstackgerrit | Yu Zhang proposed openstack/cinder master: IBM XIV: fix issues for replication https://review.openstack.org/588453 | 07:09 |
openstackgerrit | caixiaoyu proposed openstack/cinder master: Add test case: create volume from source volume https://review.openstack.org/588431 | 07:25 |
openstackgerrit | caixiaoyu proposed openstack/cinder master: Add test case: create volume from source volume https://review.openstack.org/588431 | 07:33 |
openstackgerrit | yenai proposed openstack/cinder master: Add Flase check when do _retype_error https://review.openstack.org/588144 | 07:49 |
*** tommylikehu is now known as tommylikehu2 | 08:02 | |
*** tommylikehu2 is now known as tommylikehu | 08:03 | |
*** tommylikehu is now known as tommylikehu_afk | 08:04 | |
*** tommylikehu_afk is now known as tommylikehu | 08:09 | |
*** vivsoni_ has quit IRC | 08:29 | |
*** deepak_mourya has quit IRC | 09:25 | |
*** vivsoni has joined #openstack-cinder | 09:31 | |
openstackgerrit | cheng li proposed openstack/cinder master: Using "availability_zone" property in restoration https://review.openstack.org/588447 | 09:48 |
*** obre is now known as obre_ | 09:58 | |
*** obre_ is now known as obre | 09:58 | |
*** obre has quit IRC | 10:04 | |
*** obre has joined #openstack-cinder | 10:06 | |
openstackgerrit | caixiaoyu proposed openstack/cinder master: Add test case: create volume from source volume https://review.openstack.org/588431 | 10:09 |
*** luizbag has joined #openstack-cinder | 10:47 | |
*** e0ne has joined #openstack-cinder | 11:01 | |
*** e0ne has quit IRC | 11:01 | |
*** e0ne has joined #openstack-cinder | 11:02 | |
*** e0ne has quit IRC | 11:02 | |
*** e0ne has joined #openstack-cinder | 11:17 | |
*** e0ne has quit IRC | 11:17 | |
*** vivsoni has quit IRC | 11:17 | |
*** walshh_ has joined #openstack-cinder | 11:38 | |
*** dave-mccowan has joined #openstack-cinder | 11:43 | |
*** tpsilva has joined #openstack-cinder | 11:48 | |
*** abishop has joined #openstack-cinder | 11:52 | |
*** walshh_ has quit IRC | 11:56 | |
*** zul has joined #openstack-cinder | 12:02 | |
*** hemna_ has quit IRC | 12:04 | |
*** hemna_ has joined #openstack-cinder | 12:15 | |
*** mriedem has joined #openstack-cinder | 12:39 | |
*** nicolasbock has joined #openstack-cinder | 13:00 | |
*** edmondsw_ has joined #openstack-cinder | 13:00 | |
*** eharney has joined #openstack-cinder | 13:07 | |
*** dustins has joined #openstack-cinder | 13:07 | |
*** dave-mccowan has quit IRC | 13:25 | |
*** mriedem is now known as mriedem_afk | 13:25 | |
*** lbragstad has quit IRC | 13:26 | |
*** stephenfin is now known as finucannot | 13:28 | |
*** psachin has quit IRC | 13:33 | |
*** ganso has joined #openstack-cinder | 13:37 | |
*** _ix has joined #openstack-cinder | 13:46 | |
*** gnufied has joined #openstack-cinder | 13:49 | |
*** gnufied has quit IRC | 13:55 | |
kukacz_ | hi, what is the status of capacity based qos implementation? is there a release where this works with front-end policy? | 14:03 |
*** edmondsw_ is now known as edmondsw | 14:08 | |
kukacz_ | I would like to implement qos with specs like `total_iops_sec_per_gb=30`, as in the example at https://docs.openstack.org/cinder/latest/admin/blockstorage-capacity-based-qos.html | 14:09 |
*** Luzi has quit IRC | 14:15 | |
jungleboyj | kukacz_: By front end do you mean at the attachment point? | 14:20 |
kukacz_ | hmm, I mean by Nova with KVM - so libvirt/qemu perhaps | 14:22 |
kukacz_ | ie. defined the same way as in the example | 14:22 |
*** _ix has quit IRC | 14:22 | |
jungleboyj | kukacz_: Ok. So that has been implemented for a while. The documentation should be correct. | 14:22 |
kukacz_ | I did a test and it did not work - applied as in the example, created new volume from the appropriate volume type (the one where this qos was associated), attached to existing instance and tested using fio from within an instance | 14:24 |
smcginnis | luizbag: Morning. I didn't see a backport proposed for that os-brick patch. Did you need some help with that? | 14:24 |
kukacz_ | jungleboyj: is there perhaps something which needs to be configured in nova or cinder controllers or compute nodes, to make it work? | 14:25 |
luizbag | No, thanks. I need to pass it through our internal code review on NetApp | 14:25 |
smcginnis | luizbag: Should I just do it then? It's just a backport. I thought NetApp got better about those things. | 14:26 |
jungleboyj | kukacz_: I didn't think there was anything extra that needs to be done. | 14:26 |
jungleboyj | What storage backend do you have? | 14:26 |
luizbag | smcginnis: no need, I'll get it through this afternoon | 14:26 |
smcginnis | Hmm, OK. We could have it merged by now. :) | 14:27 |
kukacz_ | jungleboyj: solidfire and ceph | 14:29 |
kukacz_ | jungleboyj: does it make a difference? | 14:31 |
*** cmart has quit IRC | 14:31 | |
jungleboyj | kukacz_: Interesting. | 14:32 |
jungleboyj | kukacz_: So, I am not as familiar with how this all works. According to the support matrix RBD doesn't support QoS but it should work on Solidfire. | 14:32 |
kukacz_ | jungleboyj: I thought that "front-end" consumer means that the qos is applied at hypervisor, disregard the type of backend | 14:35 |
jungleboyj | I think that it still requires some support from the backend? | 14:36 |
jungleboyj | smcginnis: ^^^ Am I off the rails here? | 14:36 |
kukacz_ | jungleboyj: should mention that common "total_iops_sec" applied by "front-end" consumer works. the solidfire specific qos specs (all 3 of them) applied by "back-end" consumer works as well | 14:38 |
kukacz_ | jungleboyj: what I'm looking for though, is the capacity based model. I could not find it in solidfire driver description, so I thought I would implement it using front-end | 14:39 |
jungleboyj | Oh, so it is just the iops/gb ? | 14:40 |
jungleboyj | That isn't working? | 14:40 |
smcginnis | jungleboyj: I would look to eharney or maybe mnaser on that. I have not touched QoS much. | 14:41 |
jungleboyj | smcginnis: That was what I was thinking. | 14:41 |
jungleboyj | kukacz_: What release are you on? | 14:41 |
kukacz_ | jungleboyj: yes, the iops/gb is what I'm seeking | 14:41 |
kukacz_ | jungleboyj: ocata | 14:42 |
jungleboyj | kukacz_: Hmmm, let me look here. The iops/gb was added more recently. I am not totally sure when. Give me a minute to look here. | 14:42 |
smcginnis | Capacity based merged in pike or queens I think. | 14:42 |
jungleboyj | smcginnis: Ah, there would be the rub. :-) | 14:42 |
kukacz_ | thanks, that would explain it | 14:43 |
jungleboyj | kukacz_: That was my first thought but I missed the capacity based part of your first note. :-) | 14:43 |
smcginnis | Yep, spec was merged in pike - https://specs.openstack.org/openstack/cinder-specs/specs/pike/capacity_based_qos.html | 14:43 |
*** mriedem_afk is now known as mriedem | 14:43 | |
jungleboyj | smcginnis: Thank you for confirming sir. | 14:44 |
kukacz_ | pike, great! that's the easier upgrade path for me :-) | 14:44 |
kukacz_ | smcginnis, jungleboyj: thanks a lot guys! | 14:44 |
jungleboyj | kukacz_: No problem. Glad to help. | 14:45 |
jungleboyj | kukacz_: Out of curiosity what are you going to do instead? | 14:45 |
kukacz_ | hmm, sorry, what do you mean? are you asking about the usecase? | 14:46 |
*** cmart has joined #openstack-cinder | 14:47 | |
jungleboyj | Since it isn't implemented in ocata what will you do instead? | 14:48 |
kukacz_ | I'll consider proceeding with upgrade or will have to build a pricing model specifying both the $/IOPS and $/GiB values to stay on the safe side with storage utilization | 14:52 |
kukacz_ | my motivation for the IOPS/GB is to keep a static ratio on a performance tier and to avoid bearing costs for the extreme situations where customers consume large IOPS with just small GiBs and vice versa | 14:55 |
kukacz_ | ... means I prefer the GCE storage pricing model over the AWS one :-) | 14:57 |
* mnaser reads buffer | 14:58 | |
jungleboyj | kukacz_: Sounds good. Hope you are able to do the upgrade to get the functionality in place. | 14:58 |
mnaser | ah yeah, looks like jungleboyj figured it out | 14:59 |
mnaser | it was added in pike so yeah | 14:59 |
jungleboyj | mnaser: The credit goes to smcginnis :-) | 14:59 |
mnaser | gah i fixed the pep8 issue yesterday but forgot to psuh up change | 15:00 |
mnaser | darn | 15:00 |
smcginnis | Likely story. :) | 15:00 |
openstackgerrit | Mohammed Naser proposed openstack/cinder master: Add ability to specify minimum value when using capacity based QoS https://review.openstack.org/588302 | 15:00 |
mnaser | voila | 15:00 |
*** gnufied has joined #openstack-cinder | 15:16 | |
*** cmart has quit IRC | 15:23 | |
*** gnufied has quit IRC | 15:25 | |
*** pooja_jadhav has quit IRC | 15:25 | |
*** cmart has joined #openstack-cinder | 15:41 | |
*** Shilpa has quit IRC | 15:43 | |
*** itlinux_ has quit IRC | 15:43 | |
*** itlinux has joined #openstack-cinder | 15:49 | |
*** mriedem is now known as hansmoleman | 15:57 | |
-openstackstatus- NOTICE: The infra team is renaming projects in Gerrit. There will be a short ~10 minute Gerrit downtime in a few minutes as a result. | 16:01 | |
*** melwitt is now known as jgwentworth | 16:14 | |
*** cmart has quit IRC | 16:27 | |
*** pocketprotector- has quit IRC | 16:27 | |
*** pocketprotector- has joined #openstack-cinder | 16:27 | |
*** pocketprotector- is now known as pocketprotector | 16:27 | |
*** psachin has joined #openstack-cinder | 16:29 | |
*** imacdonn has quit IRC | 16:37 | |
*** imacdonn has joined #openstack-cinder | 16:37 | |
*** cmart has joined #openstack-cinder | 16:41 | |
*** openstackgerrit has quit IRC | 16:49 | |
*** zhaochao has quit IRC | 16:52 | |
hansmoleman | smcginnis: anyone ever talk about upstreaming something like this? https://github.com/starlingx-staging/stx-cinder/commit/ac9c475f6f598be8312b7480918525ac87eddf21#diff-ce3ab5d94902facf73ddd54d10739347R116 | 17:13 |
hansmoleman | this is mriedem btw | 17:13 |
*** gouthamr has quit IRC | 17:23 | |
*** erlon has quit IRC | 17:26 | |
*** gouthamr has joined #openstack-cinder | 17:26 | |
-openstackstatus- NOTICE: Project renames and review.openstack.org downtime are complete without any major issue. | 17:27 | |
smcginnis | hansmoleman: We have the messaging API for a more general way of getting details on failures with async operations. | 17:28 |
smcginnis | hansmoleman: Does that track any faults with the volume, even outside of async operations. | 17:28 |
smcginnis | (I know you are not involved enough to expect you to know that) | 17:28 |
*** psachin has quit IRC | 17:29 | |
smcginnis | Quick skim it looks like it's just on volume create, so we do have that now with the messages API. | 17:31 |
*** gouthamr has quit IRC | 17:33 | |
hansmoleman | ok i sort of knew about the messages API so i figured that was the answer | 17:33 |
smcginnis | hansmoleman: Looks like we could add some better description here - https://developer.openstack.org/api-ref/block-storage/v3/index.html#messages | 17:34 |
smcginnis | Trying to track down the spec from that too. | 17:34 |
hansmoleman | i'd think people using the messages API would also want to filter messages based on resource_uuid | 17:35 |
hansmoleman | so gimme all the messages for my volume | 17:35 |
smcginnis | Hmm, I thought we did that, but it wouldn't appear so from the api-ref. | 17:36 |
smcginnis | Original spec - https://specs.openstack.org/openstack/cinder-specs/specs/newton/summarymessage.html | 17:37 |
*** hemna_ has quit IRC | 17:41 | |
jungleboyj | hansmoleman: That would be nice. | 17:47 |
hansmoleman | similar for listing attachments for a given volume | 17:48 |
hansmoleman | we need that on the nova side for some stuff, | 17:48 |
hansmoleman | but i'm too busy (lazy) to write a spec / change to cinder for that | 17:48 |
hansmoleman | the volume also has the attachments list in it though | 17:48 |
*** e0ne has joined #openstack-cinder | 18:15 | |
*** gnufied has joined #openstack-cinder | 18:42 | |
*** luizbag has quit IRC | 18:47 | |
*** hemna_ has joined #openstack-cinder | 18:48 | |
*** gouthamr has joined #openstack-cinder | 19:04 | |
*** openstackgerrit has joined #openstack-cinder | 19:17 | |
openstackgerrit | Merged openstack/cinder master: Add ability to specify minimum value when using capacity based QoS https://review.openstack.org/588302 | 19:17 |
*** woojay has left #openstack-cinder | 19:29 | |
*** gouthamr has quit IRC | 20:00 | |
openstackgerrit | Sean McGinnis proposed openstack/cinder master: Improve messages api-ref description https://review.openstack.org/588653 | 20:21 |
*** harlowja has joined #openstack-cinder | 20:21 | |
*** tswanson has quit IRC | 20:27 | |
smcginnis | jungleboyj: We still have a bug with python-cinderclient for v3 volume transfers. Filing a bug now. | 20:35 |
jungleboyj | Dang was going to try and test that. What is wrong? | 20:37 |
smcginnis | Not sure yet, but I can repro the failure using the cinder CLI. | 20:37 |
smcginnis | Works with v2 API, not with v3. | 20:37 |
smcginnis | Doesn't work with the 3.55 microversion either. | 20:38 |
*** e0ne has quit IRC | 20:40 | |
*** e0ne has joined #openstack-cinder | 20:41 | |
*** abishop has quit IRC | 20:45 | |
*** e0ne has quit IRC | 20:45 | |
smcginnis | Pffffft | 21:05 |
openstackgerrit | Sean McGinnis proposed openstack/cinder master: Correct volume-transfers API endpoint https://review.openstack.org/588664 | 21:17 |
*** dustins has quit IRC | 21:27 | |
*** pcaruana has quit IRC | 21:38 | |
openstackgerrit | Sean McGinnis proposed openstack/cinder master: Add api-ref for mv 3.55 volume transfer https://review.openstack.org/588666 | 21:40 |
*** gouthamr has joined #openstack-cinder | 21:47 | |
*** KeithMnemonic has quit IRC | 21:53 | |
openstackgerrit | Sean McGinnis proposed openstack/python-cinderclient master: Fix backwards compat for volume transfer < 3.55 https://review.openstack.org/588673 | 22:05 |
*** Guest43093 is now known as amrith | 22:12 | |
*** ganso has quit IRC | 22:18 | |
openstackgerrit | Sean McGinnis proposed openstack/cinder master: Add api-ref for mv 3.55 volume transfer https://review.openstack.org/588666 | 22:36 |
*** nicolasbock has quit IRC | 22:43 | |
*** tpsilva has quit IRC | 22:54 | |
openstackgerrit | Sean McGinnis proposed openstack/cinder master: Correct volume-transfers API endpoint https://review.openstack.org/588664 | 22:56 |
*** harlowja has quit IRC | 23:48 | |
*** tswanson has joined #openstack-cinder | 23:48 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!