00:00:27  * chorrellquit (Client Quit)
00:03:51  * chorrelljoined
00:23:02  * fredkquit (Quit: Leaving.)
00:24:32  * AvianFluquit (Remote host closed the connection)
00:25:22  * chorrellquit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
00:36:25  * CarlosCquit (Quit: Leaving.)
00:37:04  * bcantrillquit (Quit: Leaving.)
00:48:18  * chorrelljoined
01:16:02  * bcantrilljoined
01:30:32  * mjnjoined
01:35:51  * abraxasjoined
02:20:07  * chorrellquit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
02:33:33  * papertigersjoined
02:42:55  * papertigersquit (Quit: papertigers)
03:03:07  * _Tenchi_quit (Quit: Leaving)
03:28:04  * insaneirishjoined
03:39:35  * papertigersjoined
03:42:44  * bcantrillquit (Quit: Leaving.)
03:58:36  * papertigers_joined
03:59:46  * papertigersquit (Ping timeout: 240 seconds)
03:59:46  * papertigers_changed nick to papertigers
04:33:31  * yunongjoined
04:36:01  * yunongquit (Client Quit)
04:41:31  * papertigers_joined
04:42:55  * papertigersquit (Ping timeout: 246 seconds)
04:42:56  * papertigers_changed nick to papertigers
04:53:58  * yunongjoined
04:56:06  * ssandersjoined
04:56:06  * stonecobraquit (Read error: Connection reset by peer)
05:26:45  * foomanjoined
05:28:14  <tjfontaine>fooman: so what exactly are you doing, and what's causing problems?
05:28:54  <fooman>i am trying to do is use manta to do mass compiling of LESS and Stylus files, basically serving as a CSS preprocessor platform, i want to compile the files and store them in manta
05:29:35  <tjfontaine>ok
05:30:11  <fooman>i tar balled my node_modules directory that i had for local development, which included stylus and less processors, i uploaded the tar ball, and extracted in, so now in manta i have /stor/node_modules
05:30:14  * ssandersquit (Ping timeout: 240 seconds)
05:31:47  <fooman>so just as a test, not even processing anything, i created a script that just does var stylus = require("stylus"), that alone is enough to throw an error for that job
05:31:50  * yunongquit (Quit: Leaving.)
05:32:26  <fooman>would nom projects and scripts that make use of them need to be modified to use relative directories for manta?
05:32:29  <fooman>npm*
05:33:21  <tjfontaine>well I guess that depends on what the error is, and how you're getting your modules into the zone
05:34:50  <fooman>i'll edit the error i just got and paste it
05:36:12  <fooman>{"phase":"0","what":"phase 0: map input \"/USER/stor/manta/test.js\"","code":"UserTaskError","message":"user command exited with code 8","stderr":"/USER/jobs/JOB_ID/stor/USER/stor/manta/test.js.0.err.ad96f37b-1769-4933-9422-328d958bc415","input":"/USER/stor/manta/test.js","p0input":"/USER/stor/manta/test.js"}
05:36:26  * stonecobrajoined
05:36:35  <tjfontaine>fooman: did you actually inspect the stderr output?
05:37:17  <fooman>no, how can i do that with manta? i thought mjob errors # was as far as i could go
05:37:40  <tjfontaine>that field for "stderr" is an object you can mget and see the output from your job
05:37:52  <fooman>ah nice
05:38:02  <fooman>checking it out now
05:38:10  <tjfontaine>so on a single job you can do something like: mget $(mjob errors <uuid> | json stderr)
05:38:21  <tjfontaine>(presuming locally you have `npm install -g jsontool`)
05:38:54  <tjfontaine>[which you should] :)
05:39:10  * bixuquit (Remote host closed the connection)
05:39:48  <fooman>i do, trying that now
05:39:58  <fooman>thanks for helping out with all this :)
05:40:05  <tjfontaine>no problemo :)
05:40:30  <tjfontaine>also, you should checkout mlogin, it's a good way to prototype your jobs as you're working on them
05:40:51  <rmustacc>fooman: Are you using that bundle of code as an asset?
05:41:05  <tjfontaine>you can specify your object and assets and then drop into an interactive shell
05:41:47  <rmustacc>It's a bit hard to tell from what you've shared but you want to be using that bundle of node modules as an asset for your job.
05:42:28  <tjfontaine>right, part of why I said: "depends on how you're getting your modules into the zone"
05:43:11  <fooman>ah ha, that may be my problem. there is a difference between uploading just into /stor and registering an asset i'm guessing?
05:43:54  <tjfontaine>fooman: ya, by default you only get the object/key you're working on, if you want additional things in your zone you need to specify them as an asset, or mget them during your task
05:44:34  <tjfontaine>fooman: so, the tarball of node_modules is actually a good thing to use, then first part of your job can be to unpack that into a path you expect
05:44:54  * papertigers_joined
05:45:09  <fooman>so Stil upload it to /stor as normal, then unpack it?
05:45:49  * papertigersquit (Ping timeout: 245 seconds)
05:45:49  * papertigers_changed nick to papertigers
05:46:30  <fooman>mlogin looks interesting, going to try to get that working too
05:47:24  <rmustacc>Yeah, you'd want to upload as normal.
05:47:24  <tjfontaine>mput your file into your space, and specify the path to it with -s before your -m, and then in your job you can do things with it, it will be located in /assets/<path/to/object>
05:47:41  <rmustacc>But I'd basically prototype it in mlogin and use that to write a script.
05:47:44  <tjfontaine>http://apidocs.joyent.com/manta/#running-jobs-using-assets
05:48:06  <tjfontaine>though that's not exactly the example I wanted
05:48:52  <fooman>ah, i think i get it now
05:49:09  <fooman>thanks guys, going to give this a shot and report back in a few mins
05:49:39  <fooman>very much appreciate the help, tjfontaine and rmustacc
05:50:00  <tjfontaine>np
05:50:11  <rmustacc>tjfontaine: Yeah, probalby an example of using a tarball would be more illustrative.
05:50:50  <tjfontaine>rmustacc: there's not really a good one, in my looking of the docs, just the muntar example
05:51:32  <fooman>yeah the docs are a little underwhelming at the moment. i read about the service and my brain wants to explode because of how awesome it sounds, and then it takes a long time to jump in and get the hang of it because there are not enough examples to show off what you can do
05:52:30  <rmustacc>Makes sense. If you have other feedback, do let us know.
05:52:48  <tjfontaine>indeed, exactly what I was going to say, any other place where we can make it more clear please let us know
05:54:09  <fooman>definitely real world examples, showing a simple node project that utilizes something in the node_modules directory would help a lot of people i bet. probably blow a few minds to be honest.
05:55:01  <tjfontaine>I have a blog that will be posted tomorrow that uses assets, but didn't need a tarball
05:57:07  <fooman>looking forward to checking it out, will be on the joyent blog?
05:57:33  <tjfontaine>nodejs.org
05:57:59  * papertigersquit (Quit: papertigers)
05:58:33  <fooman>ok cool, i'll def check it out
06:02:07  * yunongjoined
06:03:37  <fooman>so if i give this a shot, it still says it can not find my module, which is located in /stor/include/ and contains nothing but an index.js file -- i've tested it locally and it works. this is the manta command i am running.
06:03:43  <fooman>echo /$MANTA_USER/stor/manta/test.js | mjob create -o -m -s /$MANTA_USER/stor/node_modules -s /$MANTA_USER/stor/test.js "node"
06:04:12  <fooman>wrong one, actually see an issue with that one, just a sec
06:05:10  <fooman>ok yeah this is the code i use: echo /$MANTA_USER/stor/test.js | mjob create -o -m -s /$MANTA_USER/stor/node_modules -s /$MANTA_USER/stor/test.js "node"
06:05:29  <tjfontaine>ok, so is test.js the script you want to run with node, or the script you want to compress?
06:05:43  <fooman>just the node script i want to run
06:06:07  <rmustacc>And it needs the node modules in that tgz right?
06:06:14  * yunongquit (Ping timeout: 240 seconds)
06:06:52  <fooman>yeah, node_modules has already been extracted -- should i not do it that way? do i leave it as a tar ball and have it init?
06:07:35  <tjfontaine>well, -s doesn't work with a directory, only for an object, so you'll want to use a .tar.gz, something like ...
06:08:50  <tjfontaine>eecho /$MANTA_USER/stor/test.js | mjob create -o -s /$MANTA_USER/stor/node_modules.tar.gz -m 'tar xzf /assets/$MANTA_USER/stor/node_modules.tar.gz && node $MANTA_INPUT_FILE'
06:10:01  <fooman>ah that is slick. so i can just have the script to be executed in /stor and then have node_modules tar'd ?
06:10:40  <rmustacc>So you basically want to specify both as assets and then your input file is always going to be the script to transform/process.
06:11:10  <tjfontaine>right, if something else is going to be what you want to operate on
06:11:26  <tjfontaine>or
06:11:47  <rmustacc>I'm going to disappear, so happy mantaing.
06:12:14  <tjfontaine>echo /$MANTA_USER/stor/myfile | mjob create -o 'npm install -g mypackage && mypackage $MANTA_INPUT_FILE'
06:12:17  <fooman>thanks for the help, rmustacc! goodnight
06:12:43  <fooman>* i'm not leaving, just saying bye to rmustacc *
06:12:43  <tjfontaine>so that way you're always getting the latest version from the registry
06:13:11  <fooman>oh snap, i can run nom install, my whole world just opened up
06:13:25  <fooman>damn autocorrect not liking NPM
06:13:31  <fooman>it's not nom
06:13:42  <tjfontaine>you can do whatever you want, it's a zone/vm that you have (roughly) complete control over
06:15:13  <fooman>please definitely put out some npm install with manta examples for people. i'm telling you this is blowing my mind. i'm a pretty advanced node.js user of about 2 years
06:15:46  <tjfontaine>well, I guess some of the messaging around this could be a bit better :)
06:16:59  <fooman>yeah, it's sometimes hard to explain something that does so much.
06:17:14  <fooman>it's honestly nothing short of amazing
06:17:33  <tjfontaine>in an effort to not pigeon hole, it's easy for people to get lost :)
06:19:30  <fooman>i'm giving the nom install a shot now, very excited
06:19:55  <fooman>damn it, how often do you guys see NPM get autocorrected to nom? i can't be the only one
06:20:20  <tjfontaine>it is very nomnomnom though
06:20:36  <fooman>i give it that indeed
06:29:37  <fooman>so partially good news, i see that npm was working (the progress shows in my error log), but my script still says it can't find the module. here is the line i am using:
06:29:37  <fooman>echo /$MANTA_USER/stor/stylus_test.js | mjob create -o 'npm install -g stylus && node $MANTA_INPUT_FILE'
06:30:31  <fooman>and stylus_test.js is just: var stylus=require('stylus'); console.log(typeof stylus)
06:31:08  <tjfontaine>well, don't forget that -g for npm doesn't actually mean it will show up in your require path, just that scripts/bins will be in your $PATH
06:31:18  <tjfontaine>so remove the -g and it should work
06:31:31  <fooman>ah ok, trying now
06:32:37  <fooman>that worked!
06:32:44  <fooman>yay!
06:33:00  <tjfontaine>heh
06:34:19  <fooman>ok so now i have a best practices type of question about how i should use manta
06:37:22  <fooman>so the web app i'm using this for -- we have a platform that gives customers the ability to theme our product, essentially skinning it to make it match the look of their site, etc... what are trying to do is take those variables that get set from their configurations and then automate the building of their asset files.... the thing is, anytime we push out a change to the core asset files of the product, we need to also recompile the as
06:37:55  <tjfontaine>you got cut off at "recompile the as"
06:38:12  <fooman>right now, there are only about 50 customers, and we're expecting to peak at around 5,000 or so eventually
06:38:19  * ghostbarquit (Remote host closed the connection)
06:38:33  <fooman>ah ok, fixing where i got cutoff
06:38:35  <tjfontaine>but this does sound basically like a good use case for manta :)
06:38:47  * ghostbarjoined
06:39:11  <fooman>* the thing is, anytime we push out a change to the core asset files of the product, we need to also recompile the assets for all of our customers.
06:39:19  <tjfontaine>right
06:40:19  <fooman>my plan is to always keep the latest copies of the client's files and the core project files in manta, so it's just a matter of figuring how i should queue it
06:42:27  <fooman>we have a db table for the assets, to mark which ones are still in need of being updated for each customer, so that's in place, but how would you do this the most effective way? this would eventually be like 10,000 files that suddenly need to be recompiled
06:43:02  * ghostbarquit (Ping timeout: 240 seconds)
06:43:46  <fooman>it's no problem if it takes a few hours, since our plan is to always build and cache new asset versions (with new file names) first, then push HTML changes that utilize those CSS/JS changes after all the CSS/JS is updated
06:45:05  <fooman>is there a way to batch the request? send say 1000 object requests at a time? etc.. ?
06:45:49  <tjfontaine>sure, either figure out the key in manta from your db, or use mfind, and then just give it all the keys
06:46:01  <tjfontaine>or
06:46:58  <tjfontaine>you could do it in multiple phases, where the first part of hte job maps across customers, which generates tarballs for the customer, and then queues other jobs on the tarball for the customer, which gives you a little better object framing
06:47:54  <fooman>that's a good idea
06:48:22  <tjfontaine>it's easy to change to that model later, while you're just starting it would be fine to just queue per each file
06:49:28  <fooman>yeah definitely. how does manta work as far as sending the multiple requests? should i wait for each task to complete or is that unnecessary as long as my queue is working properly and not duplicating jobs
06:49:42  * mamashjoined
06:50:13  <tjfontaine>it's all pretty independent, you can queue jobs as you need, and you can send as many keys to a given job as you need
06:50:48  <fooman>very cool, so i could use the api and just request 500 jobs back to back over and over until everything was processed
06:51:48  <tjfontaine>sure, it's quite flexible, so basically it can adapt to whatever model works for you
06:52:27  <fooman>and that would be individual api requests, right? there's no batching in place that i'm missing, correct?
06:52:37  <tjfontaine>you can create jobs all externally, or you can make jobs from inside jobs
06:52:43  <tjfontaine>I'm not sure what you mean
06:53:53  <fooman>oh, that's too -- i meant say using the node.js SDK, did i need to run the mjob command 500 times, but no i don't, i can do it promatically
06:54:14  <tjfontaine>right
06:55:53  <fooman>i meant like how some APIs (typically query-like reads and writes) allow you to send multiple requests in an array and you get an array back, with each item in the array containing the result for each item you sent -- they do it to prevent so many round trips between your app and their api
06:56:14  <fooman>i don't think any such thing exists for manta, just asking to make sure
06:57:45  <fooman>and another quick question about the assets, i can have multiple -s params in my request? like i could do this? ....
06:58:10  <fooman>echo /$MANTA_USER/stor/test.js | mjob create -o -s /$MANTA_USER/stor/node_modules.tar.gz -s /$MANTA_USER/stor/assets.tar.gz -m 'tar xzf
06:58:19  <tjfontaine>yup
06:58:35  <fooman>awesome
06:59:02  <tjfontaine>it's a bit more clear if you look at the json for createjob http://apidocs.joyent.com/manta/api.html#jobs
06:59:58  <fooman>looking now
07:01:07  <fooman>ok i get it
07:01:20  <fooman>this is all so very impressive
07:01:34  <tjfontaine>spread the good news :)
07:01:40  <fooman>seriously made my day figuring out i can do all this complicated stuff on manta.
07:02:05  <tjfontaine>we're glad you like it, you're one of the prime use cases for it
07:02:12  <fooman>oh i intend to, once i get this working, i'm writing a blog entry about how amazing it is
07:02:42  * yunongjoined
07:02:51  <tjfontaine>are you planning on also using it as the cdn for the assets?
07:04:08  <fooman>i'm going to store the compiled version of the assets in manta, but then use our cdn to serve them, so that manta only gets hit on cache miss
07:04:12  <fooman>sound right?
07:04:39  <tjfontaine>sure, whatever works best for you
07:04:45  <tjfontaine>I was just curious
07:04:56  <fooman>is manta meant to be used as a cdn as well?
07:05:42  <fooman>or should you utilize a cdn and think of manta as your resource origin?
07:05:50  <tjfontaine>well it's meant to be a object-store and compute mechanism, but you get a strongly consistent cdn as a byproduct of that
07:06:42  <tjfontaine>it's up to you to decide if it makes sense for your workflow/usecase
07:06:46  * yunongquit (Ping timeout: 240 seconds)
07:07:06  <fooman>so does that mean that the files are served from servers in various parts of the world, or only from the 2 datacenter (by default) that your files are stored in?
07:07:49  <tjfontaine>right, it's only in the region you stored them in, so if you want localized cdn it's not quite that yet
07:08:33  <fooman>gotcha, well if/when it becomes available, i'll certainly test it, it's all config values really, so easy to switch the CDN around
07:10:33  <fooman>thank you for everything tonight, tjfontaine. it's been a pleasure getting assistance from you, i'll be able to get this whole asset builder stuff wrapped up tomorrow because of how awesome manta is. thank you again for all of your help
07:10:38  <fooman>goodnight!
07:10:52  <tjfontaine>yup, don't hesitate to pop up if you have more questions :)
07:11:11  <fooman>oh i'm sure i'll be back :)
07:11:14  <fooman>thanks!
07:11:18  <tjfontaine>indeed, gnight
07:11:20  * foomanquit (Quit: Page closed)
07:18:01  * ghostbarjoined
07:25:09  * bcantrilljoined
07:28:39  * bcantrill1joined
07:31:37  * bcantrillquit (Ping timeout: 246 seconds)
07:36:24  * bcantrilljoined
07:38:55  * bcantrill2joined
07:39:22  * bcantrill3joined
07:39:36  * bcantrill1quit (Ping timeout: 268 seconds)
07:42:06  * bcantrillquit (Ping timeout: 246 seconds)
07:42:27  * bcantrill2quit (Read error: No route to host)
07:43:20  * bcantrilljoined
07:45:49  * bcantrill3quit (Ping timeout: 256 seconds)
07:47:09  * bcantrill1joined
07:49:16  * bcantrillquit (Ping timeout: 240 seconds)
07:49:28  * bcantrilljoined
07:50:19  * bcantrill2joined
07:53:10  * bcantrill1quit (Ping timeout: 268 seconds)
07:53:41  * bcantrillquit (Ping timeout: 248 seconds)
08:03:18  * yunongjoined
08:07:39  * yunongquit (Ping timeout: 246 seconds)
08:10:57  * ghostbarquit (Remote host closed the connection)
08:11:28  * ghostbarjoined
08:12:45  * bcantrilljoined
08:14:45  * bcantrill2quit (Ping timeout: 268 seconds)
08:15:59  * ghostbarquit (Ping timeout: 268 seconds)
08:18:02  * ghostbarjoined
08:21:32  * bcantrill1joined
08:25:23  * bcantrillquit (Ping timeout: 264 seconds)
08:42:26  * bcantrilljoined
08:45:13  * almostobsoletejoined
08:45:40  <almostobsolete>Hi guys, is there an rsync equivalent for getting data up to manta?
08:45:48  * bcantrill1quit (Ping timeout: 246 seconds)
08:45:51  <almostobsolete>Got a directory of files, some of which are new
08:46:08  <jperkin>I think there was but it didn't work properly so was pulled
08:46:18  <jperkin>I'd really like a working version too, for now I'm just using muntar
08:46:23  <almostobsolete>In this specific case there are no changed files, just new ones and ones that are already uploaded
08:47:50  * ghostbarquit (Remote host closed the connection)
08:47:57  <almostobsolete>Cool, reckon it's something that will make it back sometime?
08:48:19  * ghostbarjoined
08:50:07  <jperkin>almost certainly, the infrastructure is there (e.g. server-side md5)
08:51:13  <mjn>if there aren't any changed files you could fairly easily hack something up with diffing the output of a local find and a remote mfind (after stripping pathnames with 'basename' and sorting)
08:52:03  * bcantrillquit (Quit: Leaving.)
08:52:53  * ghostbarquit (Ping timeout: 248 seconds)
08:56:20  <almostobsolete>Just realised how small these files get when I gzip them, gonna just reload all :p
09:03:46  * yunongjoined
09:08:33  * yunongquit (Ping timeout: 256 seconds)
09:09:11  * abraxasquit (Remote host closed the connection)
09:09:44  <almostobsolete>What's a good way to get stats on my jobs? I want to know what's slow and what's fast
09:36:45  <almostobsolete>Another question, how do I find out my current usage of Manta? Want to see how much this is all costing (I assume very little)
10:04:22  * yunongjoined
10:09:27  * yunongquit (Ping timeout: 268 seconds)
10:30:49  <almostobsolete>Another question for whenever someone is around: what's the best way to have a job output to a named object in the store?
10:33:53  * mamashpart
11:04:51  * yunongjoined
11:09:17  * yunongquit (Ping timeout: 246 seconds)
11:18:51  * mamashjoined
11:23:17  * almostobsoletequit (Ping timeout: 248 seconds)
11:47:06  * AvianFlujoined
12:05:21  * yunongjoined
12:09:41  * yunongquit (Ping timeout: 248 seconds)
12:33:58  * marsellquit (Quit: marsell)
12:34:16  * marselljoined
12:39:03  * insaneirishquit (Quit: Computer has gone to sleep.)
12:39:19  * insaneirishjoined
12:39:37  * insaneirishquit (Remote host closed the connection)
13:00:26  * chorrelljoined
13:00:34  * chorrell_joined
13:05:57  * yunongjoined
13:10:45  * yunongquit (Ping timeout: 268 seconds)
13:22:47  * dapjoined
13:37:45  * _Tenchi_joined
13:56:13  * chorrellquit (Quit: My iMac has gone to sleep. ZZZzzz…)
13:57:34  * almostobsoletejoined
13:58:53  <almostobsolete>Apologies if someone already answered but I'm I lost connection to IRC for a bit so I'm going to ask again...
13:59:06  <almostobsolete>What's the best way to have a job output to a named object in the store?
13:59:32  <tjfontaine>echo "foo" | mpipe /$MANTA_USER/stor/obj_in_job
14:00:09  <tjfontaine>mput will also work
14:00:24  <almostobsolete>Ah, of course, just add that to my reduce step
14:00:26  <almostobsolete>thanks
14:00:42  <tjfontaine>np
14:02:17  <almostobsolete>Are there any ways of getting some insight on the performance of the stages of a job?
14:02:53  <almostobsolete>My log processing job is taking about 20 minutes to run (when I use msplit, a lot longer with a single reduce task) and I'd like to figure out exactly what's taking the time
14:04:53  * dapquit (Quit: Leaving.)
14:04:58  <tjfontaine>almostobsolete: there's some information in /$MANTA_USER/jobs/<uuid>/job.json
14:05:57  <tjfontaine>but if I had to guess, maybe you're running out of memory in your reducer?
14:06:45  * yunongjoined
14:07:37  <almostobsolete>That could slow it down without stopping it from working?
14:08:20  <tjfontaine>ya, if you start swapping
14:08:40  <almostobsolete>I'm getting "mget: ResourceNotFoundError: /andymitchell/jobs/7bd9181d-0e3c-42c8-b6cb-3894c51ac75f/job.json"
14:08:52  <almostobsolete>7bd9181d-0e3c-42c8-b6cb-3894c51ac75f is a job I'm running right now
14:09:39  <tjfontaine>oh, probably not available until job is completed, if it's running you can `mjob get uuid`
14:10:02  <tjfontaine>it will show you where you are in progress, how many tasks or retries are going on
14:10:24  <tjfontaine>oh I guess that's all what mjob get is, heh
14:11:05  * yunongquit (Ping timeout: 246 seconds)
14:11:06  <tjfontaine>anyway, I have to run, best of luck
14:11:21  <almostobsolete>Thanks for your hekp
14:11:24  <almostobsolete>*help
14:11:31  * |woody|joined
14:17:01  <almostobsolete>If anyone else is around could you tell me what the disadvantages of having too many reducers are (if any)
14:25:30  * jcspjoined
14:29:59  * dapjoined
14:31:07  <dap>almostobselete: Unfortunately, there's not a lot you can see (yet).
14:31:47  <dap>You can log whatever you want, of course, and save that to a debug object at the end
14:34:04  * Azbruhjoined
14:48:03  <almostobsolete>cool
14:48:23  <almostobsolete>How much memory does a single mapper/reducer get to play with?
14:49:03  * chorrell_quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
14:50:39  * yunongjoined
14:57:56  <dap>almostobsolete: 1GB by default, but customizable: http://apidocs.joyent.com/manta/jobs-reference.html#resource-usage-memory-and-disk-properties
14:58:10  <almostobsolete>cool
15:25:19  * mcavagejoined
15:27:24  * chorrelljoined
15:36:17  * jcspquit (Ping timeout: 268 seconds)
15:36:39  * mamashpart
15:41:40  * jcspjoined
15:42:17  * jcspquit (Read error: Connection reset by peer)
15:43:10  * jcspjoined
15:43:27  * jcspquit (Read error: Connection reset by peer)
15:44:26  * jcspjoined
15:47:25  * almostob`joined
15:50:16  * almostobsoletequit (Ping timeout: 240 seconds)
15:50:47  * jcspquit (Read error: Connection reset by peer)
15:52:25  * jcspjoined
15:54:04  * fredkjoined
15:54:45  * bcantrilljoined
15:55:48  * jcspquit (Read error: Connection reset by peer)
15:56:43  * jcspjoined
15:59:48  * jcspquit (Read error: Connection reset by peer)
16:00:46  * jcspjoined
16:01:03  * jcspquit (Read error: Connection reset by peer)
16:01:44  * nfitchjoined
16:01:57  * jcspjoined
16:05:16  * yunongquit (Quit: Leaving.)
16:07:29  * dapquit (Quit: Leaving.)
16:14:55  * papertigersjoined
16:17:50  * jcspquit (Read error: Connection reset by peer)
16:18:34  * jcspjoined
16:18:47  * jcspquit (Read error: Connection reset by peer)
16:20:43  * jcspjoined
16:22:18  * cburroughsjoined
16:25:31  * yunongjoined
16:26:16  * dapjoined
16:27:38  * mamashjoined
16:29:51  * jcspquit (Read error: Connection reset by peer)
16:30:17  * jcspjoined
16:30:34  * jcspquit (Read error: Connection reset by peer)
16:31:39  * jcspjoined
16:32:51  * jcspquit (Read error: Connection reset by peer)
16:33:49  * jcspjoined
16:35:19  * ryancnelsonjoined
16:36:10  * jcspquit (Read error: Connection reset by peer)
16:40:44  * ghostbarjoined
16:42:26  * papertigersquit (Quit: papertigers)
16:44:06  * dapquit (Quit: Leaving.)
16:52:57  * jcspjoined
16:54:33  * yunongquit (Quit: Leaving.)
17:00:31  <almostob`>What does "mfind: BadGatewayError" indicate?
17:01:03  <mcavage>almostobsolete: is that reproducible?
17:01:18  <mcavage>typically it means you got unlucky, and something disconnected, for *some reason* behind the scenes.
17:01:31  * dapjoined
17:01:50  <mcavage>i think there are a few straggler cases we have bugs for where sending in bad URLs etc can make that happen, so if it's reproducible, then it's probably the latter, and I would want to see what you're doing.
17:02:10  <almostob`>Turns out I'd dropped a | character and was giving an mjob command as a second argument to mfind
17:03:01  <mcavage>ok - so that one is one we knew about, but if you could give me an example i'll append it to the internal ticket we already have for it.
17:04:41  <almostob`>This does it: mfind /$MANTA_USER/stor/apath/ mjob
17:05:03  <mcavage>yeah that would :)
17:05:05  <mcavage>ok - thanks.
17:05:19  <mcavage>i'll look at a short term fix in node-manta.
17:05:39  <mcavage>and like i said we already have a ticket for the server side erroring on that
17:06:31  <almostob`>cool, now I know what to look for it's not so much of a problem for me
17:06:50  <almostob`>I just didn't connect "Gateway error" with "check for typos in your script" :p
17:06:54  <mcavage>fwiw, b/c the tool is "xargs friendly" it's running two mfinds:
17:06:59  <mcavage>one with .../apath
17:07:02  <mcavage>and one with 'mjob'
17:07:05  <almostob`>ah cool, makes sense
17:07:07  <mcavage>which is not a valid resource (obviously)
17:07:11  <mcavage>yeah, i know - sorry :\
17:07:19  <mcavage>we'll get it fixed up -- we hit that too.
17:07:28  <dap>Is this #51?
17:07:34  <dap>(https://github.com/joyent/node-manta/issues/51)
17:07:53  <mcavage>no that's different i think.
17:08:06  <mcavage>oh - it's related :)
17:08:07  <almostob`>Sounds like that one is fixed now?
17:08:12  <mcavage>it's b/c it's taking the last one.
17:08:21  <mcavage>no - dave is right, there are two here. all the other tools are xargs friendly.
17:08:26  <mcavage>mfind takes the _last_ argument
17:08:39  <mcavage>so in your case you got double screwed and effectively just ran 'mfind mjob'
17:10:12  <almostob`>haha :p
17:11:53  * mcavagequit (Remote host closed the connection)
17:12:20  * mcavagejoined
17:16:51  * mcavagequit (Ping timeout: 260 seconds)
17:24:16  * bcantrill1joined
17:25:32  * bcantrill2joined
17:26:28  * mamashpart
17:26:43  <nahamu>so many bynars!
17:26:46  * bcantrillquit (Ping timeout: 240 seconds)
17:28:33  * bcantrill1quit (Ping timeout: 240 seconds)
17:29:09  * bcantrill1joined
17:29:41  * bcantrill2quit (Ping timeout: 248 seconds)
17:29:42  * mamashjoined
17:37:04  * yunongjoined
17:48:25  * mamashpart
17:51:48  * natefoojoined
17:51:55  <natefoo>bam.
17:51:58  <nahamu>:)
17:53:59  * bcantrilljoined
17:54:25  * bcantrill1quit (Read error: Connection reset by peer)
17:57:07  * bcantrill1joined
17:57:20  * bcantrillquit (Read error: Connection reset by peer)
18:02:57  * bcantrilljoined
18:03:11  * mamashjoined
18:04:23  * bcantrill1quit (Ping timeout: 240 seconds)
18:04:57  * bcantrill1joined
18:07:14  * bcantrillquit (Read error: Connection reset by peer)
18:16:30  * bcantrilljoined
18:16:30  * dapquit (Quit: Leaving.)
18:19:52  * bcantrill1quit (Ping timeout: 276 seconds)
18:22:06  * bcantrillquit (Quit: Leaving.)
18:47:07  * bcantrilljoined
18:49:13  * almostob`quit (Ping timeout: 246 seconds)
18:56:16  * cburroughsquit (Ping timeout: 240 seconds)
19:27:48  * cburroughsjoined
19:31:37  * chorrellquit (Quit: Textual IRC Client: www.textualapp.com)
19:46:41  * dapjoined
19:52:47  * dapquit (Quit: Leaving.)
19:53:08  * dapjoined
20:10:59  * mcavagejoined
20:12:34  * bixujoined
20:14:37  * AvianFluquit (Remote host closed the connection)
20:44:49  * cburroughsquit (Ping timeout: 248 seconds)
21:13:53  * cburroughsjoined
21:36:39  * AvianFlujoined
21:37:52  * mamashpart
21:46:34  * ghostbarquit (Remote host closed the connection)
21:47:03  * ghostbarjoined
21:51:50  * ghostbarquit (Ping timeout: 268 seconds)
21:59:09  * ryancnelsonquit (Quit: Leaving.)
22:21:16  * bcantrillquit (Quit: Leaving.)
22:27:21  * ghostbarjoined
22:44:40  * mjn_joined
22:47:13  * cburroughsquit (Ping timeout: 246 seconds)
22:47:31  * mjnquit (Ping timeout: 240 seconds)
23:04:56  * notmattjoined
23:07:36  * ryancnelsonjoined
23:07:45  * AvianFluquit (Remote host closed the connection)
23:09:27  * AvianFlujoined
23:15:05  * ryancnelsonquit (Quit: Leaving.)
23:17:06  * ryancnelsonjoined
23:17:07  * ryancnelsonquit (Remote host closed the connection)
23:38:26  * bcantrilljoined
23:56:01  * nfitchquit (Quit: Leaving.)
23:59:42  * mcavagequit (Remote host closed the connection)