| *** flaper87 has quit IRC | 06:13 | |
| kota_ | eranrom, takashi: if you're getting here, please ping me, I'd like to ask one thing on the storlets behavior | 07:26 |
|---|---|---|
| eranrom | kota_: Hi | 07:58 |
| kota_ | eranrom: hi | 08:00 |
| kota_ | wait a bit | 08:00 |
| kota_ | eranrom: here | 08:04 |
| kota_ | eranrom: so my question is for copy handling on storlets | 08:04 |
| eranrom | yes. | 08:04 |
| eranrom | kota_: want me to describe? | 08:05 |
| kota_ | when the client issue COPY to the storlets, where the storlet handler launchs the invocation, i mean proxy vs object | 08:05 |
| *** flaper87 has joined #openstack-storlets | 08:06 | |
| *** flaper87 has quit IRC | 08:07 | |
| eranrom | kota_: gotcha, let me have a quick look at the code | 08:07 |
| *** flaper87 has joined #openstack-storlets | 08:07 | |
| kota_ | sure | 08:08 |
| eranrom | we do this in the proxy. | 08:09 |
| kota_ | I thought, we can run the storlets on the object-server to reduce the network but it looks... | 08:09 |
| eranrom | see here: https://github.com/openstack/storlets/blob/master/storlets/swift_middleware/handlers/proxy.py#L420 | 08:09 |
| kota_ | yeah, i saw that | 08:09 |
| eranrom | kota_: I think you are right. | 08:09 |
| kota_ | and now wondering why we are doing for now. | 08:09 |
| eranrom | kota_: I do not see a reason why not to do this in the object node. | 08:11 |
| eranrom | At least not from looking at the code... | 08:11 |
| eranrom | kota_: I guess we should also ask Takashi what he thinks | 08:12 |
| kota_ | eranrom: ok. and actually, I want to do it in the object-serverside since my current PoC testing requires local fd directly | 08:14 |
| eranrom | kota_: ok, I will be happy to review the patch! | 08:15 |
| kota_ | eranrom: ok, will try then | 08:17 |
| eranrom | goodluck | 08:17 |
| kota_ | thx | 08:18 |
| *** openstackgerrit has quit IRC | 09:03 | |
| *** openstackgerrit has joined #openstack-storlets | 11:55 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/storlets master: WIP: run copy source storlets on object-server if possible https://review.openstack.org/449099 | 11:55 |
| kota_ | the amazing thing is that no tests failed with that change :/ | 11:55 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/storlets master: Run copy source storlets on object-server if possible https://review.openstack.org/449099 | 12:38 |
| kota_ | eranrom: ^^ | 12:38 |
| eranrom | kota_: :-) looking. | 12:43 |
| kota_ | eranrom: sorry, I didn't have enough time with my local env on the sort of functional testing (i.e. deploying running s2aio and check where the invocation running) yet but if you could check it while I'm asleep, that will be super helpful | 12:44 |
| eranrom | kota_: np. I will update here. good night | 12:45 |
| kota_ | eranrom: much appreciated | 12:45 |
| eranrom | kota_: my pleasure | 12:45 |
| openstackgerrit | Eran Rom proposed openstack/storlets master: Run copy source storlets on object-server if possible https://review.openstack.org/449099 | 14:26 |
| openstackgerrit | Eran Rom proposed openstack/storlets master: Running .functests directly not through tox https://review.openstack.org/449177 | 14:28 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!