| kota_ | eranrom: it works for me. | 07:46 |
|---|---|---|
| kota_ | eranrom: since today? | 07:46 |
| *** eranrom has joined #openstack-storlets | 08:00 | |
| eranrom | kota_,: I was thinking only for today, but if it is better for you and takashi, we can make it permanent | 08:12 |
| kota_ | eranrom: either is fine, no idea :-) | 08:15 |
| kota_ | anyways, i will be available 12:00 UTC today | 08:30 |
| *** eranrom has quit IRC | 08:34 | |
| openstackgerrit | Takashi Kajinami proposed openstack/storlets: [WIP] Fix the fd leak in SDaemon https://review.openstack.org/272379 | 09:07 |
| *** eranrom has joined #openstack-storlets | 09:18 | |
| *** takashi_ has joined #openstack-storlets | 09:30 | |
| eranrom | takashi,: Hi is it ok for you to meet an hour earlier toady? that is 12:00 UTC | 09:35 |
| takashi_ | eranrom: It's ok for me. :-) | 09:38 |
| *** takashi_ has quit IRC | 09:44 | |
| eranrom | takashi,: Thanks! | 11:23 |
| eranrom | All, we will have today's meeting an hout earlier. That is 12:00 UTC | 11:24 |
| eranrom | That is an hour earlier no an hout earlier, what is "hout" anyway? | 11:33 |
| *** takashi_ has joined #openstack-storlets | 11:44 | |
| openstackgerrit | Merged openstack/storlets: Use Request.is_success instead of checking Response.status_int https://review.openstack.org/272355 | 11:54 |
| kota_ | hello | 12:02 |
| takashi_ | kota_: hi | 12:02 |
| eranrom | Hi | 12:02 |
| eranrom | Sorry for being late, got caried away reviewing the fd leak fix :-) | 12:02 |
| eranrom | Lets get started | 12:02 |
| eranrom | so the agenda is here: https://wiki.openstack.org/wiki/Meetings/Storlets | 12:03 |
| eranrom | First topic Austin talk. | 12:03 |
| eranrom | I will probably post my bio either today or tomorrow. and I assume you are still waiting for a green light. | 12:04 |
| kota_ | right | 12:05 |
| kota_ | i need 2 different boss approval and i confiemed to get one. | 12:06 |
| eranrom | would you like to post it as soon as you get the green light? | 12:06 |
| kota_ | maybe i will get the other tommorow-ish. | 12:06 |
| kota_ | yup, it's getting to close to deadline. | 12:07 |
| eranrom | Alright. | 12:07 |
| eranrom | Next topic? | 12:07 |
| kota_ | :-) | 12:07 |
| takashi_ | yes | 12:07 |
| eranrom | Currently we have a lot of stuff regarding the Hackathon in the meeting wiki. I would like to split it into two: | 12:08 |
| eranrom | 1. Things we want to talk about with the community | 12:08 |
| eranrom | 2. Things we want to discuss between ourselves (call it a Storlet mini hackathon :-) | 12:09 |
| takashi_ | eranrom: sounds good | 12:09 |
| kota_ | cool | 12:10 |
| eranrom | We can do this now, or I can edit the page and let you review - any preferences? | 12:10 |
| takashi_ | eranrom: I think it's not difficult, so we can do this now | 12:11 |
| eranrom | alright. So let's go over the list: | 12:12 |
| eranrom | Unit tests? | 12:12 |
| takashi_ | I think updates are belongs to the 1st group, | 12:13 |
| takashi_ | if we aim to tell the swift community the latest status of storlets | 12:14 |
| eranrom | +1 that is at the level of update. not sure we neet dive into each item, but rather just update that we have it - agree? | 12:14 |
| takashi_ | eranrom: +1 | 12:15 |
| eranrom | anything else we need to add to the updates? | 12:15 |
| eranrom | perhaps the fact that we submitted a talk... | 12:16 |
| takashi_ | eranrom: how about some works about refactoring | 12:16 |
| takashi_ | and I have no other ideas than that | 12:17 |
| eranrom | alright, so I will add refactoring. | 12:17 |
| eranrom | next is adding a non voting gate job - which is clearly for group 1... | 12:18 |
| takashi_ | eranrom: +100 | 12:18 |
| eranrom | :-) | 12:18 |
| eranrom | use footers, also number 1, I actually want to make sure it will work... | 12:18 |
| takashi_ | "footers" means "http footers", right? | 12:20 |
| kota_ | i think MIME used in EC | 12:21 |
| eranrom | I think. The problem is that I want a storlet to be able to write the metadata at the end, that is after processing the body. and to the best of my knowledge it wasn't possible a year ago, but the EC work introduced the ability to use (http) footers in the protocol between the procy and object server | 12:21 |
| eranrom | s/proct/proxy | 12:22 |
| eranrom | s/procy/proxy/ | 12:22 |
| takashi_ | eranrom: I understand | 12:22 |
| eranrom | kota_,: I see. so perhaps it should be mime then. | 12:22 |
| eranrom | need to look. | 12:22 |
| kota_ | yup, but it needs some extra headers to make a conversation with object-server... | 12:23 |
| kota_ | not sure for now it is available from middleware. | 12:23 |
| openstackgerrit | Merged openstack/storlets: Move StorletGatewayBase under storlet_gateway module https://review.openstack.org/272359 | 12:24 |
| kota_ | if not availble right now, it would be due to requirements from encryption stuff though. | 12:24 |
| eranrom | kota_,: I see. So it looks like its a good candidate for a discussion with the community... | 12:24 |
| eranrom | kota_,: Unless you think that with the encryption work we can just use it... | 12:26 |
| kota_ | :-) | 12:26 |
| eranrom | :-) | 12:27 |
| eranrom | Resource limitation - group 2? | 12:27 |
| takashi_ | I noticed that I misplaced the next item, it should not be under the footter stuff. | 12:27 |
| takashi_ | about system header, sorry | 12:27 |
| takashi_ | Resource limitation should belong to group 2 | 12:27 |
| eranrom | takashi,: you mean the bugs section? | 12:28 |
| eranrom | takashi,: or the chunked transfer? | 12:28 |
| takashi_ | no, "Can we also add system header like Content-Type?" this one | 12:28 |
| takashi_ | I'm wondering if we can pass system header into storlets or get it out of storlets | 12:29 |
| eranrom | alright. missed that. np. group 1? | 12:29 |
| takashi_ | I think it is for group 2 | 12:29 |
| takashi_ | It's about interface in storlets | 12:30 |
| takashi_ | s/in/inside/ | 12:30 |
| eranrom | ok, gotcha, so placing this in group 2. | 12:30 |
| takashi_ | eranrom: :-) | 12:31 |
| eranrom | it is a good point | 12:31 |
| eranrom | bugs to be fixed in swift is clearly 1. | 12:31 |
| eranrom | :-) | 12:31 |
| takashi_ | eranrom: yes! | 12:31 |
| eranrom | and I want to add another one to 2, which is documentation. | 12:31 |
| takashi_ | eranrom: +1. very important item | 12:32 |
| eranrom | yup. | 12:32 |
| eranrom | once in Bristol I am sure we can fill up all the spare time we will have with more subjects :-) | 12:33 |
| eranrom | bugs? | 12:33 |
| takashi_ | yes | 12:33 |
| takashi_ | we have two important bugs now. one of them is about range handling, which we discussed in the last meeting | 12:34 |
| takashi_ | and the other is about fd leak, which I added yesterday. | 12:35 |
| eranrom | I started to review the second one. Really good catch. | 12:35 |
| takashi_ | As I described in the bug report, I found that two file descriptors are leaked for each storlet execution | 12:36 |
| takashi_ | log fd and pipe fd | 12:36 |
| eranrom | right | 12:37 |
| takashi_ | I already fixed the first thing, but not yet about the second thing. now I have to find which pipe is not closed | 12:37 |
| eranrom | I thought that you found it.. isn't it taskIdOut_.close(); | 12:38 |
| takashi_ | eranrom: I hope so, but have not tested after I found it today. | 12:39 |
| eranrom | takashi,: gotcha | 12:39 |
| takashi_ | eranrom: I'll test it tommorow morinig. It causes some failure in our testing, and we should solve this. | 12:39 |
| eranrom | Do you want me to send comments or do you prefer to work on it some more? | 12:39 |
| takashi_ | s/this/it/ | 12:39 |
| takashi_ | I'm going to work on it some more, but also open for comments | 12:40 |
| eranrom | alright. | 12:41 |
| takashi_ | I'm now working about this fd leak, so have not so big progress about range handling. | 12:42 |
| eranrom | ok. | 12:42 |
| kota_ | i'm starting proxy chunk read handling again :-) | 12:43 |
| takashi_ | kota_: :-) | 12:43 |
| eranrom | kota_,: you refer to Takashi's Swift fix? | 12:43 |
| kota_ | yup | 12:44 |
| eranrom | Great! tested it successfully on my env. | 12:44 |
| kota_ | I've been working with Takashi how we can fix that | 12:44 |
| kota_ | and i guess it is getting good shape to land. | 12:45 |
| kota_ | nice work. | 12:45 |
| eranrom | that would be great. Once it is in, we can bump our testing to run over 2.6.0 | 12:45 |
| eranrom | that is to the latest Swift that will have the fix... | 12:46 |
| eranrom | clearly it will not be 2.6.0 :-( | 12:46 |
| kota_ | 2.6.0 already released :/ | 12:46 |
| eranrom | right... | 12:46 |
| eranrom | Anything else for today? | 12:47 |
| kota_ | non from me. | 12:47 |
| takashi_ | non | 12:47 |
| eranrom | Alright, so I will update the Swift hackathon list according to what we have discussed. | 12:47 |
| eranrom | Thank you very much! | 12:48 |
| kota_ | thanks eranrom :) | 12:48 |
| takashi_ | eranrom: Thanks | 12:48 |
| *** takashi_ has quit IRC | 12:51 | |
| openstackgerrit | Doron Chen proposed openstack/storlets: Documentation: how to send parameters to a storlet https://review.openstack.org/273124 | 15:59 |
| openstackgerrit | Doron Chen proposed openstack/storlets: Documentation: how to send parameters to a storlet https://review.openstack.org/273124 | 17:07 |
| *** eranrom has quit IRC | 17:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!