hazmat | niemeyer, kapilt.com/files/topology.data | 00:06 |
---|---|---|
SpamapS | hazmat: I hear there's some kind of smack-off going on with 1000+ nodes | 00:40 |
SpamapS | hazmat: still going on? | 00:41 |
hazmat | SpamapS, yeah | 01:09 |
hazmat | SpamapS, f*cking yaml | 01:09 |
hazmat | ;-) | 01:09 |
SpamapS | hazmat: eh? | 02:02 |
SpamapS | hazmat: I thought the problem was watching the topology in zk | 02:02 |
hazmat | SpamapS, it was yaml parsing that was problematic | 02:04 |
SpamapS | oh! | 02:04 |
hazmat | SpamapS, it was really slow with larger yaml docs | 02:04 |
SpamapS | that makes perfect sense | 02:04 |
SpamapS | what yaml do we have that is large? | 02:04 |
hazmat | like 6s for 100k, takes like 0.01s with json | 02:04 |
SpamapS | yeah yaml is not for big docs | 02:04 |
hazmat | SpamapS, topology | 02:04 |
SpamapS | so we keep a yaml.. of the whole topo ?! | 02:04 |
hazmat | SpamapS, yup | 02:05 |
SpamapS | *doh* | 02:05 |
SpamapS | so its not so much that we're sending it to everybody.. its that yaml is complex to parse | 02:18 |
SpamapS | hazmat: so are you guys trying a json based version? | 02:18 |
hazmat | SpamapS, not yet, its done | 02:22 |
hazmat | SpamapS, i will add it though | 02:23 |
hazmat | SpamapS, we should talk tomorrow though about it | 02:23 |
hazmat | its a hoser for existing envs | 02:23 |
Aram | evening. | 16:22 |
Aram | where's gustavo, I wonder? | 16:23 |
fwereade | Aram, gustavo's heading off home today; not ure if he's actually *gone* but he's certainly going | 16:24 |
Aram | thanks fwereade | 16:24 |
Aram | hi rogpeppe | 16:24 |
rogpeppe | Aram: yo! | 16:25 |
Aram | rogpeppe: I might try to do a better os than os in Inferno today. | 16:25 |
Aram | any tips? | 16:25 |
Aram | I never actually written any limbo before :). | 16:25 |
rogpeppe | Aram: "a better os than os" ? | 16:25 |
Aram | the one that sets the current directory. | 16:25 |
rogpeppe | Aram: still in the dark | 16:26 |
rogpeppe | Aram: ah, you mean os(1) ? | 16:26 |
Aram | yes. | 16:26 |
rogpeppe | Aram: was it you i had the discussion with about reverse-mapping paths? | 16:26 |
Aram | yes. | 16:27 |
rogpeppe | Aram: and i didn't ever find that package, right? | 16:27 |
Aram | right. | 16:27 |
Aram | I wonder if there's a better way than trying to make sense of /prog/<pid>/ns | 16:33 |
Aram | I'm not sure if parsing /prog/<pid>/ns is enough, in all cases... | 16:33 |
rogpeppe | Aram: it definitely isn't | 16:35 |
Aram | yeah... | 16:36 |
Aram | so what do you suggest? | 16:36 |
rogpeppe | Aram: but it's the best we can do | 16:36 |
Aram | I see. | 16:36 |
rogpeppe | Aram: -> #inferno | 16:37 |
=== samkottl_ is now known as samkottler | ||
hazmat | m_3, ping | 22:28 |
hazmat | m_3, seems to work for me | 22:32 |
hazmat | m_3, the src wasn't up to date on the controler node | 22:40 |
m_3 | pong | 22:44 |
m_3 | hazmat: ok, I'll pull and try again | 22:44 |
hazmat | m_3, sorry messed with your vim accidentally, i quit it to avoid any random changes | 22:48 |
m_3 | np | 22:50 |
m_3 | kicked it off, we'll dig through the prov agent once it's up | 22:51 |
m_3 | hazmat: sweeet! | 22:53 |
hazmat | m_3, cool | 22:53 |
m_3 | hazmat: (that was supposed to be "Duuuude!") | 22:54 |
m_3 | we'll try again.... | 22:54 |
hazmat | ;-) | 22:54 |
m_3 | hazmat: sweeeet! | 22:54 |
hazmat | m_3, where you at? | 22:54 |
m_3 | in session... up in a few | 22:54 |
m_3 | damn, that's fast | 22:57 |
hazmat | m_3, yeah? | 23:00 |
m_3 | *\o/* | 23:01 |
hazmat | m_3, nice! | 23:02 |
hazmat | m_3, clock it.. i'd like to know if we need to make it faster still | 23:03 |
m_3 | I'll buy the beer! | 23:03 |
m_3 | we're going up to 11! | 23:03 |
m_3 | it's clocked in the deploy.log (head it and tail it and compare) | 23:03 |
m_3 | (when done) | 23:03 |
m_3 | gotta detach... you in 205? | 23:03 |
hazmat | m_3, yup | 23:09 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!