| *** brinzhang_ is now known as brinzhang | 09:09 | |
| tobias-urdin | sean-k-mooney: ack | 09:39 |
|---|---|---|
| sean-k-mooney | o/ | 09:40 |
| sean-k-mooney | tobias-urdin: by the way are you using the nats stream functionality or just nats-core in the poc | 09:44 |
| sean-k-mooney | apparently the stream funcitonality has a key value store and object store | 09:44 |
| sean-k-mooney | depending on the performance that may mean that nats could repace memcache too for our caching usecases | 09:45 |
| sean-k-mooney | if it was plugged into oslo.cache | 09:45 |
| sean-k-mooney | i dont know if that level consolidation is desireable or not but just something i noticed looking at the docs | 09:46 |
| tobias-urdin | sean-k-mooney: just core, interesting didn't know about that | 09:57 |
| sean-k-mooney | it used ot be called nats-stream but got renamed to jetstream | 09:58 |
| sean-k-mooney | not sure when/why | 09:58 |
| sean-k-mooney | but when using jetstream it adds message persitance and then built on top of that they added a key value strore and object storage | 09:59 |
| sean-k-mooney | i have not really looked into either much | 09:59 |
| tobias-urdin | yeah, i want to investigate that, the biggest issue was that the official python library was async and i didn't know how to implement that | 09:59 |
| tobias-urdin | so went with third-party for poc with just core | 09:59 |
| sean-k-mooney | ack ya i have played with pythons async freatures but just in toy scripts | 10:04 |
| sean-k-mooney | i havent really used it for anything real | 10:04 |
| sean-k-mooney | although you proably could hide much of that by just awiting everything | 10:04 |
| sean-k-mooney | im not sure how eventlets and async python works together at the same time | 10:05 |
| sean-k-mooney | oslo may or may not be monkey patch but nova certenly will be for example so one way or another eventlets will show up | 10:06 |
| tobias-urdin | ye that's the not-so-fun part :p | 10:18 |
| damani | sean-k-mooney, tobias-urdin, if you need help to implement new drivers i'm intersted to help | 10:45 |
| sean-k-mooney | damani: good to know i proably wont have much time next cycle to work on openstack. ill still be around but i will be doing some openstack installer work which will reduce my time to actully work on openstack projects | 10:54 |
| sean-k-mooney | i do plan to pick back up my unix socket notification driver in antelope | 10:55 |
| sean-k-mooney | but im not sure ill have much tiem to look at nats or zeromq ectra | 10:55 |
| sean-k-mooney | however i woudl not mind trying out nats at some point possible in that installer work | 10:56 |
| sean-k-mooney | ill be working on deploying nova via and openshift operator on openshift and while we are currently considring using rabbit nats might be a better choice eventually in such an envionment | 10:57 |
| damani | sean-k-mooney, okk | 11:04 |
| damani | ok | 11:04 |
| sean-k-mooney | from a disto perspective maintaining rabbitmq is a bit of a pain since erlang is not a skillset / technology that is widely used in our products | 11:05 |
| tobias-urdin | damani: cool, feel free to shime in on the mailing list discussion | 12:13 |
| tobias-urdin | sean-k-mooney: openstack-k8s-operators? :) | 12:13 |
| sean-k-mooney | yes the name is ironic since it does not work with vanilla k8s | 12:16 |
| sean-k-mooney | i would like it to evenutally | 12:16 |
| sean-k-mooney | once we are done with the zed feature freeze and rc period im going to be shifing some of my attention to the nova operator | 12:17 |
| sean-k-mooney | https://github.com/openstack-k8s-operators/nova-operator | 12:18 |
| sean-k-mooney | currently tring to figure out internally where my time will be best spent as i have other pocs in flight also one of them being the poc that need the unix socket notification driver | 12:22 |
| tobias-urdin | sean-k-mooney: so i'm assuming above is the future of ooo then | 12:23 |
| sean-k-mooney | not nessisarly | 12:25 |
| sean-k-mooney | i think the future of tripleo is the way that its moving with the standalone ansibel roles | 12:25 |
| tobias-urdin | ok i was assuming since why otherwise would redhat spend time on it | 12:25 |
| tobias-urdin | anyway, i guess nats would make even more sense in that case :) | 12:33 |
| sean-k-mooney | it may. we discussed a few options adn for now rabbit is what we know how to use and imporantly openstack knows how to use | 12:39 |
| sean-k-mooney | if there were lighter weight alternitives that could be intersting to explore. | 12:39 |
| sean-k-mooney | once we actully have things working. | 12:39 |
| hberaud | stephenfin, tkajinam: o/ Do you mind to review this patch. Thanks in advance. https://review.opendev.org/c/openstack/oslo.config/+/851836/ | 14:07 |
| stephenfin | hberaud: done | 14:17 |
| hberaud | thx | 14:17 |
| opendevreview | Merged openstack/oslo.config master: fix CLI mismatch in unit tests https://review.opendev.org/c/openstack/oslo.config/+/851890 | 16:01 |
| opendevreview | Merged openstack/oslo.config master: Disable a config test if ran under the root user https://review.opendev.org/c/openstack/oslo.config/+/851836 | 16:02 |
| damani | tobias-urdin, ok, i will check that | 16:14 |
| opendevreview | Merged openstack/oslo.messaging master: update hacking pin to support flake8 3.8.3 https://review.opendev.org/c/openstack/oslo.messaging/+/842984 | 19:45 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!