| *** oshani has quit IRC | 00:00 | |
| *** oshani has joined #cc | 00:20 | |
| *** parker-fcnyu has quit IRC | 00:22 | |
| *** parker-fcnyu has joined #cc | 04:15 | |
| *** Kaetemi has quit IRC | 06:17 | |
| *** parker-fcnyu has quit IRC | 06:27 | |
| *** oshani has quit IRC | 06:32 | |
| *** wormsxulla has quit IRC | 08:34 | |
| *** MarkDude has quit IRC | 08:46 | |
| *** MarkDude has joined #cc | 08:49 | |
| *** wormsxulla has joined #cc | 08:54 | |
| *** igorlukanin has joined #cc | 09:36 | |
| *** igorlukanin has quit IRC | 09:42 | |
| *** igorlukanin has joined #cc | 09:49 | |
| *** igorlukanin has joined #cc | 10:13 | |
| *** akila87 has joined #cc | 10:39 | |
| *** akila87 has quit IRC | 11:16 | |
| *** parker-fcnyu has joined #cc | 11:19 | |
| *** JoiIto has joined #cc | 11:33 | |
| *** oshani has joined #cc | 12:24 | |
| *** everton137 has quit IRC | 12:37 | |
| *** oshani has quit IRC | 12:51 | |
| *** midoubleko has joined #cc | 12:56 | |
| *** akila87 has joined #cc | 13:01 | |
| *** parker-fcnyu has quit IRC | 13:03 | |
| *** nkinkade has joined #cc | 13:22 | |
| *** tvol has joined #cc | 13:23 | |
| *** nkinkade has quit IRC | 13:42 | |
| paroneayea | ;o | 14:03 |
|---|---|---|
| *** akila87 has quit IRC | 14:10 | |
| *** oshani has joined #cc | 14:18 | |
| *** akila87 has joined #cc | 14:18 | |
| *** JoiIto has quit IRC | 14:58 | |
| *** JoiIto has joined #cc | 14:59 | |
| *** JoiIto1 has joined #cc | 15:12 | |
| *** JoiIto has quit IRC | 15:12 | |
| *** nyergler has joined #cc | 15:41 | |
| akila87 | Hello nyergler, | 15:44 |
| *** nkinkade has joined #cc | 15:46 | |
| *** oshani has quit IRC | 15:47 | |
| *** Kaetemi has joined #cc | 15:47 | |
| *** jgay has joined #cc | 16:01 | |
| nyergler | morning, akila87 | 16:04 |
| akila87 | nyergler, I have check that RDF problem. It actually a valid RDF but the only problem is, we cannot add Compact URI like "dc:title" as we want. | 16:06 |
| nyergler | akila87, ok | 16:09 |
| akila87 | and I tried localization support also. Netbeans localization wizards works in OpenOffice too. But still its better to integrate .po files. | 16:14 |
| nyergler | akila87, yes, if we can support PO files that'd be great -- then we can list them for translation with our other tools | 16:14 |
| akila87 | nyergler, And the method used to show the license notification when loading is not successful. | 16:19 |
| akila87 | Because to add the listener to the onload event first we have to call the initialize() method. It won't work until click the creative commons menu. It should be done using Jobs I'm trying to get it done. | 16:19 |
| akila87 | This is just to inform you what I'm doing these day. | 16:19 |
| nyergler | akila87, sounds great, thanks for the update | 16:20 |
| akila87 | nyergler, any time :) | 16:23 |
| *** igorlukanin has quit IRC | 16:36 | |
| *** nyergler has quit IRC | 16:40 | |
| *** oshani has joined #cc | 16:46 | |
| *** akozak has joined #cc | 16:47 | |
| *** nyergler has joined #cc | 16:47 | |
| *** igorlukanin has joined #cc | 16:48 | |
| *** JED3 has quit IRC | 17:02 | |
| nkinkade | paroneayea: Are you pushing the new cc.engine live today still? | 17:04 |
| *** mralex has joined #cc | 17:04 | |
| *** luisv has joined #cc | 17:05 | |
| *** JoiIto has joined #cc | 17:06 | |
| *** JoiIto has left #cc | 17:06 | |
| *** JoiIto1 has quit IRC | 17:09 | |
| paroneayea | nkinkade: yeah | 17:13 |
| *** mralex has quit IRC | 17:15 | |
| *** mralex has joined #cc | 17:15 | |
| paroneayea | nkinkade: in theory, it's serving on 9088 on creativecommons.org, though I found out that the package is missing a file, so working on resolving that | 17:25 |
| paroneayea | actually | 17:25 |
| paroneayea | nkinkade, nyergler: Would we prefer to run cc.engine out of a git checkout of some tag, or | 17:25 |
| paroneayea | would we prefer to have cc.engine pull an egg release | 17:25 |
| paroneayea | I'm doing the latter for now, as you can see on a5:/var/www/creativecommons.org/cc.engine_sanity/ | 17:26 |
| nyergler | paroneayea, not exactly sure what you mean, looking now | 17:26 |
| nyergler | paroneayea, my instinct is that using a git clone is fine, not sure that the egg release actually gives us anything (although I guess it lets us separate deployment and development configuration?) | 17:28 |
| nyergler | (and am i correct that the setup.py is completely unnecessary in that directory?) | 17:29 |
| nyergler | paroneayea, actually i guess having the run-time configuration versioned isn't a bad idea | 17:29 |
| paroneayea | nyergler: yeah, I guess the versioning of the config is the only real advantage rather than checking out a tag | 17:30 |
| nyergler | paroneayea, right... | 17:30 |
| paroneayea | ok, easy enough to switch it over :) | 17:30 |
| nyergler | why don't you stick with this, but make sure it always unzips eggs? | 17:30 |
| nyergler | http://pypi.python.org/pypi/zc.buildout#always-unzipping-eggs | 17:30 |
| paroneayea | ok | 17:31 |
| nyergler | paroneayea, i suppose it also lets us version the combination of wsgi_cache + cc.engine, etc | 17:31 |
| paroneayea | yeah | 17:31 |
| paroneayea | so that leads to a second question: | 17:32 |
| paroneayea | currently when you do a buildout inside of cc.engine it does an svn checkout of all the "resources" it needs (javascript, images and etc) | 17:33 |
| paroneayea | I could package that with the egg but I don't think it's really necessary | 17:33 |
| paroneayea | since theoretically that would only be used in development anyway | 17:33 |
| nkinkade | I'll you and nyergler decide these things since it's all beyond what I know about. | 17:33 |
| nyergler | paroneayea, makes sense | 17:34 |
| paroneayea | ok, cool | 17:34 |
| nyergler | i suppose that's another reason to version the deployment configuration independently | 17:34 |
| nyergler | so that's let us remove the direct_remote_paths from deploy.ini, right? | 17:34 |
| *** igorlukanin has quit IRC | 17:34 | |
| paroneayea | those are used for url generation, not serving | 17:35 |
| nyergler | paroneayea, ah | 17:36 |
| nyergler | (nevermind :) ) | 17:36 |
| paroneayea | brb | 17:36 |
| nyergler | ok | 17:36 |
| nkinkade | nyergler: Does the idea of making our transition to Transifex.net also our transition to git for i18n sound reasonable to you? | 17:45 |
| nyergler | nkinkade, yes, i think so | 17:48 |
| *** JED3 has joined #cc | 17:53 | |
| *** luisv has quit IRC | 17:53 | |
| *** luisv has joined #cc | 17:55 | |
| JED3 | nkinkade: ping | 18:02 |
| nkinkade | JED3: Hi. | 18:02 |
| *** akila87 has left #cc | 18:12 | |
| *** mick_laptop has joined #cc | 18:32 | |
| mick_laptop | anyone know if there is a license that says basically CC: By (and if you improve on it, let me know) type of license? | 18:33 |
| greg-g | mick_laptop: what do you mean by "(and if you improve on it, let me know)" | 18:35 |
| paroneayea | mick_laptop: you mean... CC BY? | 18:35 |
| mick_laptop | yes | 18:35 |
| mick_laptop | CC BY | 18:35 |
| paroneayea | I mean, if you want a license that is like CC BY, why not use CC BY | 18:36 |
| nyergler | mick_laptop, i don't know of any license that requires notification (which is what I think you're getting at) | 18:36 |
| paroneayea | oh | 18:36 |
| mick_laptop | yes | 18:36 |
| mick_laptop | ok, thanks | 18:36 |
| luisv | no free licenses that require notification, no | 18:36 |
| luisv | ISTR it is not strictly non-free, but debian and others strongly dislike notification requirements? but it may also be non-free/non-OSI to require notification | 18:36 |
| mick_laptop | is that still considered "free" if it requires (well i'd like to make it where it would be nice) | 18:37 |
| greg-g | mick_laptop: as luisv said, many people would say it is not a "Free" license in that respect. | 18:38 |
| mick_laptop | i'm interested in creating CC content for CPR training (I found out that the American Red Cross charges quite a bit for that type of stuff). I'm going to get an instructors license (well that is my plan) and then release something under a CC license - but if there is something that is factually incorrect - i'd like to be notified | 18:39 |
| mick_laptop | i'd see that it gets peer review first | 18:39 |
| mick_laptop | but still, it would be nice to have something like that | 18:39 |
| luisv | mick_laptop: you can always say 'I release under CC, so you're not technically required to notify me of changes, but it would be super-awesome if you notify me of changes' | 18:39 |
| mick_laptop | i want to stick in the realm of "free" as much as possible though | 18:39 |
| mick_laptop | i guess i'll do that | 18:40 |
| mick_laptop | along w/ a huge disclaimer :) | 18:40 |
| paroneayea | mick_laptop: I think most people *want* to help and inform of changes and etc | 18:40 |
| luisv | and if you explain *why* you want notifications, that would also be awesome | 18:40 |
| mick_laptop | due to the nature of the content | 18:40 |
| paroneayea | huge disclaimer is sensible :) | 18:40 |
| paroneayea | CPR non-warranty | 18:40 |
| mick_laptop | right | 18:40 |
| luisv | right, what paroneayea said- most people want to help, so if you explain to them 'helping is super-important because of the nature of this information- please help me build something super-useful!' then they'll be even more inclined to do so | 18:41 |
| mick_laptop | "please don't practice this on your kid sister... actually get certified.... etc. | 18:41 |
| mick_laptop | luisv: will do, i like it | 18:41 |
| mick_laptop | i think this type of information is like asperin - it should be free | 18:42 |
| greg-g | mick_laptop: to save you some time, check out the disclaimer on these resources from a medical school. You can use them (the disclaimer) under the terms of the CC:BY license, so just attribute it to the "Open.Michigan initiative" https://open.umich.edu/education/med/m1cardioresp-fall2007/lectures.html | 18:46 |
| greg-g | full disclosure: I work with that initiative | 18:47 |
| greg-g | mick_laptop: also, there might be some good images/other resources available here that you can use under a CC license (it varies) that has effectively been vetted by the University of Michigan Medical School: https://open.umich.edu/education/med | 18:48 |
| greg-g | welcome to the OER community :) | 18:48 |
| mick_laptop | cool, thanks :) | 18:49 |
| *** jgay has quit IRC | 19:06 | |
| nyergler | paroneayea, how are things going? | 19:07 |
| paroneayea | nyergler: things run. I tested against the local server with links and that worked | 19:08 |
| paroneayea | backing up the apache and varnish configs and about to work on that | 19:08 |
| nyergler | great | 19:08 |
| JED3 | oh yeah today's the day eh? | 19:14 |
| paroneayea | yup | 19:24 |
| *** christian_h has joined #cc | 19:30 | |
| *** christian_h has quit IRC | 19:32 | |
| *** luisv has quit IRC | 19:52 | |
| *** luisv has joined #cc | 19:54 | |
| nkinkade | paroneayea: One the live chooser I noticed that U.S. is my default option. | 19:55 |
| nkinkade | for jurisdiction. | 19:55 |
| paroneayea | nkinkade: well, it's not up yet :) | 19:55 |
| nkinkade | Oh. | 19:55 |
| nkinkade | Sorry. :-) | 19:56 |
| paroneayea | I'm still editing the conf files and stuff | 19:56 |
| paroneayea | np | 19:56 |
| *** pktck has joined #cc | 20:43 | |
| *** erlehmann has joined #cc | 21:18 | |
| *** oshani has quit IRC | 21:26 | |
| paroneayea | okayyyy | 21:27 |
| paroneayea | looks like everything's in order to move things over to sanity | 21:28 |
| * paroneayea wipes brow nervously | 21:28 | |
| *** oshani has joined #cc | 21:29 | |
| nyergler | paroneayea, great | 21:29 |
| nyergler | paroneayea, let me know if you want review of anything | 21:29 |
| paroneayea | nyergler: sure. Why not review these config files before I move them over and restart varnish & apache | 21:29 |
| paroneayea | you and/or nkinkade :) | 21:30 |
| nyergler | paroneayea, where are they now? | 21:30 |
| nyergler | if nkinkade is around and can review, that'd be great, too | 21:30 |
| nkinkade | I'm here. | 21:30 |
| erlehmann | hai again. | 21:30 |
| nkinkade | paroneayea: Where are the temporary config files? | 21:30 |
| paroneayea | /etc/varnish/creativecommons.vcl.new /var/www/creativecommons.org/conf/creativecommons.conf.new | 21:30 |
| paroneayea | 21:30 | |
| paroneayea | and backed up to the respective .old files | 21:31 |
| nyergler | paroneayea, looking | 21:31 |
| paroneayea | there's also a new init script at /etc/init.d/cc.engine which works | 21:31 |
| nkinkade | paroneayea: Nothing looks obviously wrong to me, but it's hard for me to know since I don't know how the underlying engine works. | 21:34 |
| nyergler | paroneayea, the only thing i see that i'm not certain about is that I don't see /publicdomain/ being handled int he deeds | 21:34 |
| nyergler | er, in the apache configuration | 21:34 |
| paroneayea | I don't think it is, it's handled by varnish | 21:34 |
| paroneayea | that's how it is on staging | 21:34 |
| paroneayea | if ( req.url ~ "^/(license|choose|characteristic|publicdomain)/" ) { | 21:35 |
| paroneayea | set req.backend = cc_engine; | 21:35 |
| paroneayea | } | 21:35 |
| nyergler | paroneayea, i assumed the CC0 deeds, for example, would be handled by the wsgi_cache, as well | 21:35 |
| paroneayea | oh | 21:35 |
| paroneayea | :) | 21:35 |
| paroneayea | I never thought about that. | 21:35 |
| paroneayea | for some reason. | 21:35 |
| nyergler | paroneayea, it's probably fine to send it direct -- i think that's what we do now, anyway | 21:35 |
| nyergler | but in the fullness of time, we should use the same approach for all the deeds | 21:36 |
| paroneayea | I'll set up a ticket | 21:36 |
| nyergler | thanks | 21:36 |
| nyergler | paroneayea, i also wonder if the Alias line in the apache configuration needs to come before the rewritecond lines that rely on it | 21:36 |
| paroneayea | I can put it before that. Though it *is* working that way on staging :) | 21:38 |
| nyergler | paroneayea, oh, ok :)( | 21:38 |
| nyergler | ok :) | 21:38 |
| paroneayea | regardless, looks cleaner and can't hurt | 21:38 |
| nyergler | paroneayea, it was idle speculation on my part :) | 21:38 |
| paroneayea | 1sec, gonna grab fresh tea before I throw this switch | 21:40 |
| nyergler | ok | 21:40 |
| JED3 | nyergler: good call on POST'ing to retire users in roundup | 21:47 |
| nyergler | JED3, awesome | 21:50 |
| JED3 | nyergler: you wouldn't happen to have a micro usb cable here in the office i could use for an hour or so? | 21:56 |
| nyergler | no, sorry; mralex may | 21:56 |
| JED3 | akozak: might you have one? | 21:57 |
| mralex | he already asked me :) | 21:58 |
| mralex | i left mine at home | 21:58 |
| *** tvol has quit IRC | 22:01 | |
| paroneayea | JED3: nyergler: nkinkade & etc: it's up! | 22:07 |
| nyergler | paroneayea, sweet! | 22:07 |
| paroneayea | take a look to make sure all is in working order plz plz plz | 22:07 |
| nyergler | yup, doing so | 22:07 |
| nyergler | paroneayea, where does the deed cache live? (just curious) | 22:09 |
| nyergler | oh, in /var/www/creativecommons.org/cc.engine_sanity/cache? | 22:09 |
| paroneayea | /var/www/creativecommons.org/cc.engine_sanity/cache | 22:09 |
| nyergler | :) | 22:09 |
| paroneayea | yes :) | 22:09 |
| nyergler | paroneayea, did it just die? | 22:10 |
| nyergler | getting a 503 on choose and publicdomain | 22:10 |
| paroneayea | it's working for me ? | 22:10 |
| paroneayea | oh | 22:11 |
| paroneayea | not working on publicdomain | 22:11 |
| paroneayea | shoot | 22:11 |
| nkinkade | paroneayea: I'm getting a 503 too. | 22:11 |
| paroneayea | ok, switching back | 22:11 |
| nyergler | or choose | 22:11 |
| nkinkade | on /choose | 22:11 |
| nyergler | well did it fall over? | 22:11 |
| paroneayea | not sure yet | 22:11 |
| paroneayea | I set it back to zope, looking | 22:11 |
| paroneayea | cwebber@a5:/var/www/creative | 22:12 |
| paroneayea | Server running in PID 25197 | 22:12 |
| paroneayea | 22:12 | |
| paroneayea | oops | 22:12 |
| paroneayea | cwebber@a5:/var/www/creativecommons.org/conf$ sudo /etc/init.d/cc.engine statusServer running in PID 25197 | 22:12 |
| paroneayea | looks like it's running... | 22:12 |
| nyergler | paroneayea, if it's running, shouldn't http://creativecommons.org:9088/choose/ work? | 22:13 |
| paroneayea | it hasn't worked for some reason | 22:13 |
| paroneayea | must be firewalled | 22:13 |
| paroneayea | I've been using lynx | 22:13 |
| *** wormsxulla has quit IRC | 22:13 | |
| paroneayea | er, links | 22:13 |
| paroneayea | oh :\ | 22:14 |
| paroneayea | I think I might know what killed it | 22:14 |
| nyergler | oh? | 22:14 |
| paroneayea | I set --pid-file and --log-file to the same file >_< | 22:14 |
| nyergler | :) | 22:14 |
| paroneayea | which is also why I couldn't kill it | 22:14 |
| nyergler | heh | 22:14 |
| *** MarkDude has quit IRC | 22:15 | |
| paroneayea | pl | 22:16 |
| paroneayea | er | 22:16 |
| paroneayea | ok | 22:16 |
| paroneayea | should be back up | 22:16 |
| nkinkade | paroneayea: One tiny thing is that the link in the info box for jurisdictions opens in the same window. To me, I feel it should open in a new window, but others may have other feels on that. | 22:16 |
| nyergler | paroneayea, and, it seems to be 503 again :/ | 22:17 |
| paroneayea | it went down again | 22:17 |
| paroneayea | yeah | 22:17 |
| paroneayea | :< | 22:17 |
| nyergler | nkinkade, you're probably right (by info box you mean the "i" link?) | 22:17 |
| nkinkade | nyergler: yeah. | 22:17 |
| nyergler | nkinkade, great, can you just open a ticket for that? | 22:18 |
| nyergler | that way paroneayea can focus on getting it rolled out, then sweep up minor points | 22:18 |
| nkinkade | Yeah. I think it's actually in an existing ticket. | 22:18 |
| nyergler | ah | 22:18 |
| nkinkade | I'll break it out though if necessary. Let me look. | 22:18 |
| paroneayea | ugh. I didn't kill the old process may have been the problem the second time. Trying to bring it up once more... | 22:19 |
| nkinkade | nyergler: It's already here: http://code.creativecommons.org/issues/issue572 | 22:20 |
| nkinkade | ... so I think we're good on that trivial point. | 22:20 |
| nyergler | nkinkade, thanks | 22:21 |
| *** MarkDude has joined #cc | 22:21 | |
| *** akozak has quit IRC | 22:22 | |
| paroneayea | okay | 22:22 |
| paroneayea | it is up one more time | 22:22 |
| paroneayea | I think I've cleared out my stupidity this time | 22:22 |
| paroneayea | nope :\ | 22:24 |
| paroneayea | the paste logs aren't showing any relevant error | 22:24 |
| nyergler | paroneayea, i suggest we try to reproduce the problem elsewhere, possibly using apache bench (ab) | 22:25 |
| paroneayea | I think it's happening somewhere around the varnish layer? | 22:25 |
| nyergler | paroneayea, what makes you think that? | 22:25 |
| paroneayea | well, maybe foolishness, but | 22:26 |
| nyergler | :) | 22:26 |
| paroneayea | the 503 error is coming from varnish | 22:26 |
| paroneayea | and varnish talks directly to cc.engine | 22:26 |
| paroneayea | I'm not getting errors in links | 22:26 |
| nyergler | paroneayea, so when it falls over you're able to still use links to get a page? | 22:27 |
| nkinkade | From a quick run through the basic functionality, all seems to be working as expected. | 22:27 |
| paroneayea | yeah | 22:27 |
| paroneayea | it's happening on stuff like /choose/ | 22:27 |
| paroneayea | so | 22:27 |
| paroneayea | that shouldn't be part of the apache config | 22:27 |
| nyergler | nkinkade, do i remember that we had trouble at one point with varnish connection handling? like we had to "pipe" something from zope? | 22:28 |
| nyergler | (i could be making that up) | 22:28 |
| nkinkade | nyergler: There have been some issues in the past, but we haven't seen any of them in a long time, I believe. | 22:29 |
| nyergler | ah | 22:29 |
| nkinkade | Since we upgraded Varnish. | 22:29 |
| paroneayea | one difference between staging and live is | 22:29 |
| paroneayea | if ( req.http.host == "labs.creativecommons.org" ) { | 22:29 |
| paroneayea | pass; | 22:29 |
| paroneayea | I didn't add that on live | 22:29 |
| nkinkade | For example bzr wasn't working right through Varnish back a couple years ago. That may be fixed now. | 22:29 |
| paroneayea | I don't know what that's for or how I came across it | 22:29 |
| nyergler | paroneayea, that's just for the labs.cc.org vhost | 22:30 |
| nkinkade | paroneayea: That's a rule to simply not cache anything on Labs. | 22:30 |
| *** wormsxulla has joined #cc | 22:30 | |
| nyergler | nkinkade, is there any way to find out what varnish thinks the problem is when it 403s? | 22:30 |
| nkinkade | nyergler: If it emitted a 503, then likely it couldn't connect to the requested backend. | 22:31 |
| nkinkade | or at least Varnish felt that it couldn't for whatever reason. | 22:31 |
| nyergler | right, but we have circumstantial evidence to the contrary | 22:31 |
| paroneayea | one thing I could do | 22:31 |
| nkinkade | I'm not sure how much debugging info can be pulled from that. | 22:31 |
| paroneayea | is run cc.engine in daemon mode | 22:31 |
| nyergler | you mean non-daemon mode? | 22:32 |
| nkinkade | We could inspect the Varnish log. | 22:32 |
| paroneayea | x_x | 22:32 |
| paroneayea | yes | 22:32 |
| paroneayea | non-daemon mode | 22:32 |
| paroneayea | ah yes, varnish log is a good idea also :) | 22:32 |
| nyergler | paroneayea, great, let's try that; when it falls down again (assuming it does), we'll look at what the varnish log reports; reasonable, nkinkade? | 22:32 |
| nkinkade | nyergler: It's possible, but the file is massive. | 22:33 |
| nyergler | sigh | 22:33 |
| nkinkade | 8.7G at the moment, and entries flying by at a mile a minute. I could try to pass the output through grep or pass some option to only pull certain things. | 22:33 |
| nyergler | ok, paroneayea, let's see if running non-detached gives us anything interesting | 22:34 |
| nkinkade | Like, "only show up things from host XX" | 22:34 |
| nkinkade | Let me see if there is any quick way to get what we want from the Varnish log. | 22:34 |
| nyergler | thanks | 22:34 |
| paroneayea | tail? | 22:35 |
| paroneayea | tail and enough lines could give us something recent at least | 22:36 |
| paroneayea | it's up | 22:39 |
| paroneayea | hasn't fallen over yet... | 22:39 |
| nyergler | paroneayea, bonk | 22:39 |
| nyergler | 403 | 22:39 |
| nyergler | 504 | 22:39 |
| nyergler | ack | 22:39 |
| nyergler | 503 | 22:39 |
| paroneayea | ope | 22:39 |
| paroneayea | zope back up | 22:39 |
| paroneayea | paste showed nothing | 22:40 |
| nyergler | nkinkade, can we set up pipe or pass on a per backend basis in varnish? | 22:42 |
| nkinkade | nyergler: I'm not sure. I can find out. I suspect though that it won't be at that level. | 22:42 |
| nyergler | nkinkade, i guess i'm thinking it might be interesting to try and remove varnish from the equation as much as possible | 22:43 |
| nyergler | by either piping or passing the request | 22:43 |
| nkinkade | nyergler: Is the 503 error still present? | 22:43 |
| nyergler | yes, still having trouble, back on the zope back end right now | 22:43 |
| nyergler | paroneayea, i assume you're just copying config files to switch between the two? | 22:44 |
| nyergler | nkinkade, i suppose we can just say "pipe" after setting the backend, huh? | 22:45 |
| nkinkade | nyergler: I don't think so. I think that can only be done from the vcl_recv function. | 22:45 |
| nyergler | nkinkade, sorry, wasn't clear; i meant in the vcl_recv function, after the line that assigns req.backend | 22:46 |
| nyergler | (not after the backed is defined) | 22:46 |
| nkinkade | Oh, yeah, In that case, for sure. | 22:46 |
| nyergler | paroneayea, i think we should try to pipe the request... i honestly don't know exactly how it's different, but i do know that it's resolved issues in the past | 22:47 |
| nyergler | paroneayea, i can make the change to the VCL, just want to make sure i make it in the right copy; cc.vcl.new, right? | 22:47 |
| nkinkade | nyergler: We might be able to add a rule like "if ( req.backend.port == "9080" ) { pipe; } | 22:48 |
| nyergler | nkinkade, i don't think we even need to do that | 22:48 |
| nyergler | we can just add the pipe line after the two req.backend = ccengine lines, i think | 22:49 |
| paroneayea | nyergler: yeah | 22:50 |
| nyergler | paroneayea, ok, i'm going to make that change, can you switch over to paste after i do that? | 22:50 |
| paroneayea | yeah | 22:50 |
| nyergler | paroneayea, great, done | 22:51 |
| nyergler | go :) | 22:51 |
| paroneayea | it's up | 22:52 |
| nyergler | paroneayea, i think it fell over again :/ | 22:54 |
| paroneayea | set it back | 22:54 |
| nyergler | :) | 22:54 |
| nyergler | ok, we can try passing the request (varnish's lowest level of interaction) | 22:55 |
| nyergler | actually, let's defer to tomorrow morning. | 22:56 |
| nyergler | nkinkade, can you look into debugging the varnish log (either today or first thing tomorrow) so we can see if it's spitting up? | 22:56 |
| nkinkade | nyergler: I can right now. | 22:56 |
| nkinkade | I was just looking over the output. | 22:57 |
| nkinkade | The data is somewhat limited. | 22:57 |
| nyergler | nkinkade, thanks; maybe email chris and i with what to do, and we can take another stab tomorrow morning | 22:57 |
| nyergler | i think we should try "pass" instead of "pipe" first -- just to try and get it out of the way as much as possible. | 22:57 |
| nkinkade | I think pipe is more useful here. | 22:58 |
| nkinkade | paroneayea: nyergler: The output from varnish log is like this: | 22:58 |
| nkinkade | http://pastebin.com/8YUku4xa | 22:58 |
| nyergler | nkinkade, is the only different that pipe caches, pass does not? | 22:59 |
| nkinkade | Basically we get response code from the backend and any headers. | 22:59 |
| nkinkade | nyergler: pipe does not cachen. | 22:59 |
| nyergler | so what's the difference? | 22:59 |
| nkinkade | pipe will pipe an entire connection right to the backend, whereas pass, I believe, goes request by requests, even on the same connection. | 22:59 |
| nyergler | ah | 22:59 |
| nkinkade | paroneayea: So are we just bailing on this for today and will look at fresh in the morning? | 23:00 |
| nyergler | nkinkade, yes | 23:01 |
| nkinkade | paroneayea: Why can we not setup this thing to run side by side with the live version, but on a different port, for example? Maybe we could run it off a URL like /chooze | 23:01 |
| *** MarkDude has quit IRC | 23:01 | |
| nkinkade | That way we can debug leisurely without having to flip-flop back and forth. | 23:02 |
| nyergler | nkinkade, sounds like a good idea | 23:05 |
| *** oshani has quit IRC | 23:20 | |
| *** oshani has joined #cc | 23:32 | |
| *** MarkDude has joined #cc | 23:34 | |
| *** luisv has quit IRC | 23:55 | |
Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!