00:06:45  * bixu_joined
00:07:18  <bixu_>Are there plans to have ACLs in Manta (so that I can have a user access a dir or something but not my whole Joyent infrastructure)
00:08:03  <wesolows>it's at the top of the list.
00:08:05  <nahamu>bixu_: you missed discussion earler about just that
00:08:30  <mikeal>is there a way to get the headers for $MANTA_INPUT_OBJECT
00:08:57  <mikeal>i need to copy them over to my mput
00:17:11  <mikeal>actually, better question, how do I copy a header from one manta object to another
00:18:16  <tjfontaine>I feel like the answer is mln
00:19:57  <mikeal>only the headers?
00:20:05  <mikeal>i thought mln was basically the manta version of a sym link
00:20:35  <tjfontaine>I guess I didn't read enough scrollback for this window
00:21:58  <mikeal>i need to get the header from inside a job for $MANTA_INPUT_OBJECT so i can pass it to my mput
00:22:27  <mikeal>or, alternatively, i could set it after the mput but i'd still need to get it inside the job
00:23:05  <mikeal>wish there was a $MANTA_INPUT_HEADER_*
00:23:18  <mikeal>so i could just do $MANTA_INPUT_HEADER_CONTENT_TYPE
00:23:24  <nfitch>mikeal: try HEADER_VAL=$(mls -j $MANTA_INPUT_OBJECT | json <header>)?
00:23:51  <nfitch>For example: [nate@dev node-manta]$ mls -j /nfitch/stor/a.txt | json type
00:23:51  <nfitch>text/plain
00:24:59  <mikeal>the json property is just called "type", strange
00:25:13  <mikeal>cool, i'll try this
00:26:48  <bixu_>nahamu: Is there a log somewhere or can you recap?
00:27:25  <nahamu>bixu_: as wesolows said, it's at the top of their list of features to implement.
00:27:55  <bixu_>Thanks.
00:28:03  <wesolows>the specifics are interesting, so mark's take is useful. But I don't know of any loggers in here.
00:28:07  <wesolows>Maybe echelog has one.
00:28:17  <nahamu>I have my logs, hang on
00:28:27  <wesolows>yep
00:28:34  <wesolows>http://echelog.com/logs/browse/manta/1375394400
00:29:43  <nahamu>oh, cool
00:29:50  <nfitch>mikeal: We actually delete the headers explicitly before printing the object: https://github.com/joyent/node-manta/blob/master/bin/mls#L187
00:30:00  <nfitch>That's weird. We should have an option that leaves them in...
00:30:28  <mikeal>that worked
00:30:29  <mikeal>sweet!
00:30:38  <mikeal>the rest should be easy :)
00:31:19  <nfitch>mikeal: Good deal. I've opened this to track the ability to leave the headers in: https://github.com/joyent/node-manta/issues/103
00:32:10  * nfitchquit (Quit: Leaving.)
00:32:20  * mikealquit (Quit: Leaving.)
00:34:48  * yunongquit (Quit: Leaving.)
00:34:49  <mcavage>nahamu/wesolows/bixu: also, the logs are on libuv as well (I find the interface slightly nicer. YMMV): http://logs.libuv.org/manta/latest
00:35:37  * fredkquit (Quit: Leaving.)
01:14:27  * mikealjoined
01:28:28  * abraxasjoined
01:45:04  * dapquit (Quit: Leaving.)
02:03:21  <mikeal>strange error in two images https://gist.github.com/mikeal/a15ef4a56a78148bead3
02:03:28  <mikeal>doing convert from imagemagick
02:03:57  <mikeal>looks like imagemagick might not be entirely configured properly in manta
02:09:50  <mikeal>is there a way to get rid of old jobs? even after i cancel them i can see them
02:21:14  * mikealquit (Quit: Leaving.)
02:56:24  * mikealjoined
03:21:49  * nfitchjoined
03:48:21  * mikealquit (Quit: Leaving.)
04:16:12  * mikealjoined
04:30:25  * bixu_quit (Ping timeout: 245 seconds)
04:51:06  * mikealquit (Quit: Leaving.)
04:52:38  * mikealjoined
05:13:10  * bixu___quit (Remote host closed the connection)
05:16:26  * mikealquit (Quit: Leaving.)
05:43:23  * mikealjoined
05:56:39  * AvianFluquit (Remote host closed the connection)
06:13:31  * mamashjoined
06:23:03  * ghostbarquit (Remote host closed the connection)
06:39:26  * nfitchquit (Quit: Leaving.)
07:09:59  * ghostbarjoined
10:04:43  * abraxasquit (Remote host closed the connection)
10:20:31  * ghostbarquit (Remote host closed the connection)
11:37:46  * AvianFlujoined
13:00:22  * chorrelljoined
13:42:52  * AvianFluquit (Remote host closed the connection)
14:39:17  * mikealquit (Quit: Leaving.)
15:12:15  * mikealjoined
15:32:40  * nfitchjoined
15:34:22  * nfitchquit (Client Quit)
15:39:50  * mikealquit (Quit: Leaving.)
15:46:58  <mcavage>mikeal: we'll look at the imagemagick thing. To your question of how to remove old jobs:
15:47:08  <mcavage>jobs are "archived" to objects once they're done. Running this:
15:47:17  <mcavage>mls -rt /$MANTA_USER/jobs
15:47:33  <mcavage>will list jobs in ascending order. So to get rid of old ones, I typically do something like this:
15:48:03  <mcavage>mls -rt /$MANTA_USER/jobs | head -50 | sed 's|^|/mark.cavage/jobs/|' | parallel -j 20 mrm -r {}
15:52:21  <nahamu>s/\/mark.cavage/~~\// ;)
15:53:33  <mcavage>yeah given the new CLI.
15:53:46  <mcavage>i just wanted to batch up a few more fixes before I publish to npm
15:57:55  <nahamu>fair
16:03:47  * mikealjoined
16:07:07  * trentmjoined
16:09:13  * trentmquit (Client Quit)
16:11:38  * nfitchjoined
16:13:39  * fredkjoined
16:13:40  * fredkquit (Client Quit)
16:13:57  * fredkjoined
16:21:41  * mikealquit (Quit: Leaving.)
16:25:26  * dapjoined
16:26:35  * yunongjoined
16:46:25  * nfitchquit (Quit: Leaving.)
16:51:43  * mamashpart
16:53:07  * yunongquit (Quit: Leaving.)
16:59:37  * mikealjoined
17:10:34  * bixujoined
17:14:20  * bixuquit (Remote host closed the connection)
17:14:47  * bixujoined
17:25:09  * bixu_joined
17:28:24  * bixu_quit (Remote host closed the connection)
17:29:14  * ghostbarjoined
17:29:53  * bixuquit (Remote host closed the connection)
17:34:00  <mcavage>mikeal: we looked into your transcode problem -- your images aren't images. they're "rate limit" 403 error pages.
17:34:50  * nfitchjoined
17:38:37  <mikeal>yeah
17:38:49  <mikeal>i'm working it out, it's a bug in the generation/scrape code
17:39:05  <mikeal>there's definitely a check in their for statusCode but there must be a bug
17:39:11  <mikeal>there
17:39:20  <mcavage>you might want to sanity check content as well (or at least content-type)
17:39:22  <mikeal>been talking to jperkin on twitter
17:39:32  <mcavage>since you're saving HTML as .png right now.
17:39:33  <mikeal>so many image content types :(
17:39:33  <mcavage>k
17:39:43  <mikeal>well, i don't actually stick a .png on there :)
17:40:01  <mikeal>and i use the content-type given from the http response in the scraper
17:44:51  * yunongjoined
17:52:37  * papertigers_joined
17:53:32  * papertigersquit (Read error: Connection reset by peer)
17:53:32  * papertigers_changed nick to papertigers
18:09:22  * bixujoined
18:09:33  * bixu_joined
18:17:11  * sbraswe1joined
18:41:36  * papertigersquit (Ping timeout: 260 seconds)
19:02:14  * papertigersjoined
19:09:29  <mcavage>I know a few people have asked here as well -- heads up that I just landed "client.createReadStream" and "client.createWriteStream" in master of node-manta.
19:10:27  <mcavage>https://github.com/joyent/node-manta/commit/4e73a6c0e7b5a1de9ad4471e5a3a7d0405f8b495
19:11:47  * yunong1joined
19:20:15  * yunongquit (*.net *.split)
19:46:38  * bixu_quit (Remote host closed the connection)
19:47:05  * bixu_joined
20:24:53  <nahamu>mcavage: might it be possible for one of the tools (mfind strikes me as the obvious choice) to spit out the full http (or https) URL for the results?
20:25:09  <nahamu>clearly one can ad-hoc sed the output, or make an alias or function to do it
20:25:25  <mcavage>you mean instead of /foo/stor/bar https://us-east.manta.joyent.com/foo/stor/bar?
20:25:31  <nahamu>yes
20:25:50  <mcavage>i suppose it could be optional (mfind is the only one that spits out full paths, iirc) -- I guess is it worth it over | sed ?
20:26:19  <nahamu>I feel like I do it often enough that I'm tempted to make a funcion in my .bashrc
20:26:30  <mcavage>hmm
20:26:33  <nahamu>if I feel like doing that, others might as well and then it might as well be a part of the tool.
20:26:40  <mcavage>ooc: why?
20:26:51  <mcavage>unless it's /public you can't pipe that to a browser or curl or anytihng?
20:27:14  <nahamu>I use /public as a dumping ground for uploading stuff to share.
20:27:44  <mcavage>oh ok so it is /public.
20:27:50  <nahamu>other option would be for mput or muntar to spit out a url at the end.
20:28:32  <mcavage>ok - well, file a GH issue, and i can make it optional.
20:28:44  <mcavage>i know at least *I* don't want it doing that all the time ;)
20:29:00  <nahamu>certainly not as default behavior, agreed.
20:29:54  <nahamu>(and probably not worth implmenting unless other people are asking for it)
20:37:36  <nahamu>https://github.com/joyent/node-manta/issues/104
20:37:59  <nahamu>can I not label it myself? sad...
20:40:06  <mcavage>oh, i don't know - i never use GH labels anyway :)
20:40:39  <nahamu>fair enough.
20:40:43  <tjfontaine>you can't label yourself, no.
20:41:47  <nahamu>now I need to file and RFE on github itself. I should be able to mark an issue I submit as some sort of low priority... but too lazy. enough bug filing for one day
20:54:44  * mamashjoined
20:58:38  <jperkin>fwiw I'd definitely +1 that, echo'ing $MANTA_URL and then copy/pasting that + the filename is somewhat annoying
20:59:43  * sbraswe1part
21:00:05  <nahamu>jperkin: +1 it on github. :)
21:05:35  * yunong1quit (Quit: Leaving.)
21:05:48  * yunongjoined
21:09:42  * mamashpart
21:24:19  * mikeal1joined
21:24:59  * mikealquit (Read error: Connection reset by peer)
21:34:43  * bixuquit (Remote host closed the connection)
21:35:10  * bixujoined
22:35:21  * mikeal1quit (Quit: Leaving.)
22:52:22  * mikealjoined
23:14:49  * yunongquit (Quit: Leaving.)