| *** vinsh has quit IRC | 00:19 | |
| *** panbalag has joined #openstack-swift | 00:20 | |
| *** vinsh has joined #openstack-swift | 00:20 | |
| *** ho has joined #openstack-swift | 00:36 | |
| *** haomaiwa_ has quit IRC | 00:39 | |
| *** vinsh has quit IRC | 00:43 | |
| *** vinsh has joined #openstack-swift | 00:44 | |
| *** vinsh has quit IRC | 00:48 | |
| mattoliverau | Sharding update: Debugging container sharding over the weekend and this morning. it now compresses the trie data and only needs to pass back a trie of distributed nodes when dealing with anything but a HEAD or GET. So now seems to be working pretty well on this SAIO.. Might need to upgrade testing to a bigger cluster soon. NOTE: it currently only shards up, doesn't shrink down yet.. that's a future Matt progamming task | 01:44 |
|---|---|---|
| mattoliverau | ;P | 01:44 |
| mattoliverau | Now I should probably be a good boy and continue writing the SPEC and go do some reviews. | 01:45 |
| mattoliverau | which I think may involve taking my laptop outside to enjoy the sun and a coffee :) | 01:49 |
| * mattoliverau realises how quiet the channel is when it's a Monday in Oz and the Japanese peeps are on holidays... I've become that guy who talks to himself | 01:50 | |
| *** pberis has quit IRC | 02:42 | |
| ho | mattoliverau: :-) we are on holidays but I just re-started about keystoneauth rebasing | 02:52 |
| *** vinsh has joined #openstack-swift | 02:59 | |
| *** silor has joined #openstack-swift | 03:05 | |
| *** vinsh has quit IRC | 03:33 | |
| *** vinsh has joined #openstack-swift | 03:33 | |
| *** vinsh has quit IRC | 03:38 | |
| *** haomaiwang has joined #openstack-swift | 03:38 | |
| *** vinsh has joined #openstack-swift | 04:33 | |
| *** vinsh has quit IRC | 04:33 | |
| *** vinsh has joined #openstack-swift | 04:34 | |
| *** vinsh has quit IRC | 04:38 | |
| *** bkopilov has joined #openstack-swift | 04:46 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Imported Translations from Transifex https://review.openstack.org/179635 | 06:07 |
| *** wer has quit IRC | 06:10 | |
| *** wer has joined #openstack-swift | 06:11 | |
| *** rameshg87 has joined #openstack-swift | 06:39 | |
| *** rameshg87 has left #openstack-swift | 06:39 | |
| *** Anticimex has quit IRC | 06:48 | |
| *** tamizh_geek has joined #openstack-swift | 07:04 | |
| *** ppai has joined #openstack-swift | 07:18 | |
| *** geaaru has joined #openstack-swift | 07:32 | |
| *** chlong has quit IRC | 07:37 | |
| *** tamizh_geek has quit IRC | 07:37 | |
| *** swat30 has quit IRC | 07:49 | |
| *** swat30 has joined #openstack-swift | 07:51 | |
| *** krykowski has joined #openstack-swift | 07:52 | |
| *** jistr has joined #openstack-swift | 07:59 | |
| *** proteusguy has joined #openstack-swift | 08:01 | |
| *** slo_ has joined #openstack-swift | 08:08 | |
| slo_ | Hi there. One quick question. After creating *.ring.gz files and distribution of those to each storage node, I have made an change to object-server.conf file. Therefore, as I found similar problem (https://answers.launchpad.net/swift/+question/238269), i get 503 internal error. Due i need to make just this call swift-ring-builder object.builder rebalance and resdistribute new object ring files? | 08:14 |
| ho | slo_: what did you change about object-server.conf file? | 08:29 |
| *** tamizh_geek has joined #openstack-swift | 08:38 | |
| *** tamizh_geek has quit IRC | 08:42 | |
| slo_ | ho: i have commented bind_ip, since the solution setting bind_ip within object-server.conf was not good | 08:43 |
| ho | slo_: are you in initial-setup phase? | 08:44 |
| slo_ | yes | 08:44 |
| ho | slo_: how about the processes of object-server on each storage node? | 08:44 |
| ahale | did you change bind ip from 127.0.0.1, what ip/ports did ring build use ? | 08:46 |
| slo_ | As said in guide (http://docs.openstack.org/juno/install-guide/install/yum/content/swift-install-storage-node.html), i have set for object-server.conf -> bind_ip = MANAGEMENT_INTERFACE_IP_ADDRESS, where management IP was set to my internal replication IP | 08:47 |
| slo_ | but i was having some errors due to that | 08:47 |
| slo_ | I thing it was the problem with calling the swift stat command already, so i commented bind_ip and than it worked | 08:49 |
| slo_ | i also commet bind_ip in account and contianer config files | 08:50 |
| slo_ | ho: so I just rebalance? | 08:55 |
| ahale | i dunno really, would make sure the ips in the ring are what the daemons are liststening on, shouldnt need to rebalance as thats for changing distribution of data around | 08:55 |
| slo_ | i have fresh install | 08:56 |
| *** krykowski has quit IRC | 08:57 | |
| *** krykowski_ has joined #openstack-swift | 08:57 | |
| slo_ | I have set up two VM with fresh CentOS 7 minimal install. Setting Keystone+Proxy/Object storage on first node and just object storage on second node. I have internal storage network 172.100.10.51, 172.100.10.52 and internal replication network 172.100.9.51, 172.100.9.51. Setting bind_ip to replication network ip for each node within object/container/account configuration files did not give me swift stat command results. So | 09:01 |
| slo_ | Than it was OK. Why's that? | 09:01 |
| slo_ | just with swift stat command and creating a map. Still exist trouble of uploading a file | 09:02 |
| slo_ | similar to this: https://answers.launchpad.net/swift/+question/238269 | 09:03 |
| *** krykowski_ has quit IRC | 09:03 | |
| *** krykowski has joined #openstack-swift | 09:03 | |
| ho | slo_: if you use the ips which listed above for rings, you don't need to rebalance again. | 09:04 |
| ho | slo_: is it possible paste your log file (/var/log/message)? you can find 503 in there. | 09:05 |
| ho | slo_: in por | 09:07 |
| ho | slo_: in proxy-server | 09:07 |
| *** jordanP has joined #openstack-swift | 09:14 | |
| *** silor has quit IRC | 09:43 | |
| *** tamizh_geek has joined #openstack-swift | 09:44 | |
| *** tamizh_geek has quit IRC | 09:48 | |
| *** krykowski has quit IRC | 09:49 | |
| slo_ | ho. I don't have them right now here but i will get back to you when i will have them/or will be accessible | 10:20 |
| *** krykowski has joined #openstack-swift | 10:22 | |
| ho | slo_: thanks for the response. today I will leave my office. have a nice day! | 10:23 |
| *** ho has quit IRC | 10:23 | |
| slo_ | ho: ok be good! hear you around | 10:29 |
| *** silor has joined #openstack-swift | 11:20 | |
| *** Anticimex has joined #openstack-swift | 11:38 | |
| *** esker has quit IRC | 11:47 | |
| *** clduser_ has joined #openstack-swift | 11:54 | |
| *** aix has joined #openstack-swift | 12:05 | |
| *** ppai has quit IRC | 12:09 | |
| *** dencaval has joined #openstack-swift | 12:11 | |
| *** proteusguy has quit IRC | 12:17 | |
| *** shakamunyi has quit IRC | 12:18 | |
| *** tamizh_geek has joined #openstack-swift | 12:33 | |
| *** tamizh_geek has quit IRC | 12:38 | |
| *** ppai has joined #openstack-swift | 12:43 | |
| *** rvasilets has joined #openstack-swift | 12:43 | |
| *** cdelatte has quit IRC | 12:48 | |
| *** links has joined #openstack-swift | 12:56 | |
| *** erlon_awaY has joined #openstack-swift | 13:15 | |
| *** erlon_awaY is now known as erlon | 13:27 | |
| *** ppai has quit IRC | 13:31 | |
| *** wbhuber has joined #openstack-swift | 13:38 | |
| *** NM has joined #openstack-swift | 13:41 | |
| *** ppai has joined #openstack-swift | 13:43 | |
| *** jeblair_ is now known as jeblair | 13:44 | |
| *** chlong has joined #openstack-swift | 13:57 | |
| *** jrichli has joined #openstack-swift | 14:02 | |
| openstackgerrit | Kamil Rykowski proposed openstack/swift: Replaced setting run_pause with standard interval https://review.openstack.org/171170 | 14:02 |
| *** fthiagogv has joined #openstack-swift | 14:05 | |
| *** chlong has quit IRC | 14:08 | |
| *** chlong has joined #openstack-swift | 14:26 | |
| *** vinsh has joined #openstack-swift | 14:49 | |
| *** NM has quit IRC | 14:50 | |
| *** rvasilets has quit IRC | 14:51 | |
| *** links has quit IRC | 14:56 | |
| *** aix has quit IRC | 15:13 | |
| *** aix has joined #openstack-swift | 15:15 | |
| *** ppai has quit IRC | 15:23 | |
| *** minwoob has joined #openstack-swift | 15:25 | |
| *** cdelatte has joined #openstack-swift | 16:04 | |
| *** delattec has joined #openstack-swift | 16:06 | |
| *** breitz has joined #openstack-swift | 16:11 | |
| *** proteusguy has joined #openstack-swift | 16:19 | |
| notmyname | good morning | 16:27 |
| *** krykowski has quit IRC | 16:31 | |
| *** jordanP has quit IRC | 16:33 | |
| *** NM has joined #openstack-swift | 16:41 | |
| *** delatte has joined #openstack-swift | 16:57 | |
| *** cdelatte has quit IRC | 17:00 | |
| *** cdelatte has joined #openstack-swift | 17:00 | |
| *** delattec has quit IRC | 17:01 | |
| *** zul has quit IRC | 17:06 | |
| *** zul has joined #openstack-swift | 17:07 | |
| *** delattec has joined #openstack-swift | 17:08 | |
| *** cdelatte has quit IRC | 17:11 | |
| *** delatte has quit IRC | 17:12 | |
| *** cdelatte has joined #openstack-swift | 17:12 | |
| *** jistr has quit IRC | 17:15 | |
| peluse | morning | 17:31 |
| *** geaaru has quit IRC | 17:32 | |
| *** bkopilov has quit IRC | 17:38 | |
| notmyname | peluse: hello | 17:47 |
| peluse | notmyname, just planning on doing reviews today - anything hot/exciting happenin? | 17:48 |
| notmyname | reviews are great, thanks. | 17:48 |
| peluse | and watching Jay Pipes cast on big Tent - any particular thoughts on this yourself? | 17:48 |
| notmyname | I'm about to take a look at the scheduling | 17:48 |
| peluse | https://www.youtube.com/watch?v=fdn5QAcmw-I&feature=youtu.be&t=9m | 17:49 |
| *** lpabon has joined #openstack-swift | 17:49 | |
| *** aix has quit IRC | 17:58 | |
| *** chlong has quit IRC | 18:02 | |
| *** EmilienM is now known as EmilienM|mtg | 18:05 | |
| *** sandywalsh_ has joined #openstack-swift | 18:05 | |
| *** sandywalsh has quit IRC | 18:06 | |
| *** zaitcev has joined #openstack-swift | 18:08 | |
| *** ChanServ sets mode: +v zaitcev | 18:08 | |
| *** tab has joined #openstack-swift | 18:09 | |
| *** tab is now known as Guest29382 | 18:09 | |
| notmyname | interesting question about the coverage module (via wbhuber) | 18:09 |
| notmyname | so there's a test that explicitly checks a condition | 18:09 |
| notmyname | but the coverage report says that code block isn't covered | 18:09 |
| notmyname | what's going on? | 18:09 |
| mtreinish | notmyname: eventlet? | 18:10 |
| notmyname | perhaps | 18:10 |
| mtreinish | there is a longstanding bug with coverage on eventlet (and similar threading mechanisms) where results are really weird | 18:10 |
| mtreinish | it's fixed in the current alpha release (4.x iirc) | 18:10 |
| notmyname | this is a for/else loop. the condition being tested is in the else | 18:10 |
| *** Guest29382 has quit IRC | 18:10 | |
| mtreinish | but in 3.7.x it's not | 18:11 |
| notmyname | 4.x of coverage? nose? eventlet? | 18:11 |
| mtreinish | coverage | 18:11 |
| notmyname | ok | 18:11 |
| *** tabb has joined #openstack-swift | 18:11 | |
| notmyname | this case is swift/obj/replicator.py:263, tested in test_reconstructor.py test_reconstruct_fa_errors_fails() | 18:12 |
| mtreinish | notmyname: but based on the description, probably not. I'd have to see the example to check | 18:12 |
| notmyname | yeah. I'm stumped | 18:13 |
| notmyname | I can prove that code block is getting exectued | 18:13 |
| * mtreinish spent way too much time debugging coverage in eventlet so he is a bit biased | 18:13 | |
| notmyname | but also that coverage shows the lines aren't executed | 18:13 |
| *** chlong has joined #openstack-swift | 18:14 | |
| notmyname | peluse: re "big tent". in general I think it's a great thing. I've pushed for more project autonomy in openstack for years, and I think that the big tent stuff is embracing reality of how projects are working instead of trying to shoehorn everything into a one-size-fits-all box. | 18:17 |
| peluse | notmyname, yeah, I didn't realize that a few governance changes had already been approved, I assume that means "in use" now? http://www.openstack.org/blog/2015/02/tc-update-project-reform-progress/ | 18:18 |
| tabb | One quick question. Having fresh install of Swift, creating/building rings I decide to change some parameter (commenting back bind_ip) within object/container/account-server.conf. Do I need to remake *.ring.gz files with swift-ring-builder rebalance? | 18:18 |
| notmyname | peluse: right. and the various tags are being worked through now AIUI | 18:19 |
| notmyname | tabb: the IP in the config needs to match what's in the ring | 18:19 |
| notmyname | if you change the config file you need to restart/reload the process. if you change the ring, it will get reloaded automatically | 18:20 |
| notmyname | tabb: IIRC the default for bind_ip is 0.0.0.0, so it shouldn't require a ring change, assuming the ring has routable IPs for each server in it | 18:21 |
| *** EmilienM|mtg is now known as EmilienM | 18:21 | |
| zaitcev | I think he's asking if rebalance is necessary or merely remove/add is enough. | 18:22 |
| zaitcev | And the answer is yes | 18:23 |
| notmyname | no it's not | 18:23 |
| notmyname | just changing the IP doesn't change data placement | 18:23 |
| zaitcev | what, really? I thought that rebalance writes out *.ring.gz, therefore all changes before that are confined to .builder. | 18:23 |
| notmyname | it does, but write_ring also writes out changes. but if you've already got an IP in the ring that can reach the nodes, changing an object server conf file from a specific IP to 0.0.0.0 (the default) wouldn't require deploying new rings | 18:24 |
| zaitcev | okay let me look at set_info source real quick | 18:25 |
| zaitcev | presumably tabb is going to use that rather than add/remove | 18:25 |
| notmyname | tabb: what was the old value in the config? | 18:26 |
| notmyname | the one you commented out | 18:26 |
| tabb | notmyname: i configured bind_ip to management IP (as said here: http://docs.openstack.org/juno/install-guide/install/yum/content/swift-install-storage-node.html), but I was having some trouble, so I commented ip_addr back as was default (so commented) | 18:26 |
| notmyname | tabb: and what's in the ring right now? and are those valid IPs for the object servers? | 18:26 |
| tabb | notmyname: i have a configuration with commeted ip_addr in all acc/cont/object-server.conf | 18:29 |
| *** jlk has left #openstack-swift | 18:29 | |
| zaitcev | tabb: you can answer notmyname's last question by running swift-ring-builder xxx.builder without other arguments, and check what it prints for the node in question | 18:31 |
| tabb | zaiticev: ok i will try it as i will have server available, this was my yesterday's work find outs... | 18:33 |
| tabb | as i recall the problem was, that I could not call swift stat command, so I find out on the internet someone having similar problems, so I changed/commeted ip_addr as it was, but than I was having trouble with uploading some file to swift object store (https://answers.launchpad.net/swift/+question/238269) | 18:37 |
| *** tsg has joined #openstack-swift | 18:50 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 18:51 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 18:51 |
| *** Nadeem has joined #openstack-swift | 18:53 | |
| *** tsg has quit IRC | 18:56 | |
| *** silor has quit IRC | 19:01 | |
| *** panbalag has quit IRC | 19:05 | |
| *** chlong has quit IRC | 19:07 | |
| *** sandywalsh_ has quit IRC | 19:09 | |
| *** panbalag has joined #openstack-swift | 19:14 | |
| *** sandywalsh has joined #openstack-swift | 19:16 | |
| *** fthiagogv has quit IRC | 19:24 | |
| clayg | so did everyone reviewing patch 178528 recognize that os.fsync(fileobject) works just as well as os.fsync(fileobject.fileno())? | 19:25 |
| patchbot | clayg: https://review.openstack.org/#/c/178528/ | 19:25 |
| clayg | oh, i guess the fcntl call won't interpolate like that | 19:25 |
| clayg | so, nm - good work | 19:25 |
| clayg | does anyone know what platform's fcntl will have a F_FULLSYNC attribute? mine has F_FULL*F*SYNC | 19:26 |
| openstackgerrit | Merged openstack/python-swiftclient: Use a socket timeout in get_auth https://review.openstack.org/177047 | 19:28 |
| *** lpabon has quit IRC | 19:31 | |
| *** panbalag has quit IRC | 19:37 | |
| *** dencaval has quit IRC | 19:45 | |
| peluse | clayg, can you take a 5 sec peek at this and tell if there's any preference between the two (and if for some reason they're not equivalent): https://gist.github.com/peluse/9518a8faf2d09f82f6cd | 19:52 |
| clayg | peluse: well there'd be a difference if some some reason some log adapter ever decided to return something (obviously) | 19:54 |
| clayg | I'd probably prefer to drop the return statement all together unless there is a *really* complelling reason to highlight the implicit return None | 19:55 |
| peluse | clayg, would it ever? that's what was sorta throwing me in a simple review - seems like the safer bet would be the 2nd way | 19:55 |
| clayg | yeah second looks much better | 19:55 |
| peluse | gracias | 19:55 |
| clayg | de nada | 19:56 |
| *** panbalag has joined #openstack-swift | 19:57 | |
| *** panbalag has quit IRC | 19:57 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 20:09 |
| *** Anticime1 has joined #openstack-swift | 20:14 | |
| *** breitz has quit IRC | 20:14 | |
| *** Anticimex has quit IRC | 20:15 | |
| *** breitz has joined #openstack-swift | 20:15 | |
| *** WBHuber_ has joined #openstack-swift | 20:16 | |
| *** wbhuber has quit IRC | 20:16 | |
| *** WBHuber_ has left #openstack-swift | 20:17 | |
| *** WBHuber_ has joined #openstack-swift | 20:17 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/173982 | 20:19 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/174343 | 20:19 |
| openstackgerrit | Michael Barton proposed openstack/swift: initial import of go code https://review.openstack.org/178851 | 20:22 |
| WBHuber_ | mtreinish: can you point me to the longstanding bug with coverage on eventlet? | 20:22 |
| *** zacksh has quit IRC | 20:24 | |
| *** zacksh has joined #openstack-swift | 20:25 | |
| notmyname | redbo: dfg_: around? | 20:30 |
| *** Trozz has quit IRC | 20:31 | |
| *** zacksh has quit IRC | 20:31 | |
| *** Trozz has joined #openstack-swift | 20:31 | |
| *** sudorandom has quit IRC | 20:32 | |
| *** zhill has joined #openstack-swift | 20:33 | |
| *** zacksh has joined #openstack-swift | 20:33 | |
| dfg_ | notmyname: whats up? | 20:35 |
| notmyname | dfg_: anything else you think needs to hit feature/hummingbird before the first revision lands? I was going to land it, but check with you in case there's something else to do before that happens | 20:35 |
| *** aerwin has joined #openstack-swift | 20:36 | |
| notmyname | dfg_: heh. seeing as you just put another +2 on it, I'm thinking the answer is "no" | 20:37 |
| *** sudorandom has joined #openstack-swift | 20:37 | |
| notmyname | dfg_: ok, button clicked | 20:37 |
| mtreinish | WBHuber_: sure one sec, I'll dig it up | 20:38 |
| notmyname | or it would seem glange is trying to race | 20:38 |
| dfg_ | ya- he wasn't logged into gerrit | 20:39 |
| dfg_ | so he lost :( | 20:39 |
| glange | it's not a competition | 20:39 |
| guitarzan | glange: that's what all the losers say | 20:39 |
| *** doxavore has joined #openstack-swift | 20:42 | |
| mtreinish | WBHuber_: I think this is the bug: https://bitbucket.org/ned/coveragepy/issue/149/coverage-gevent-looks-broken | 20:43 |
| WBHuber_ | mtreinish: looking... | 20:43 |
| doxavore | how would one speed up proxy behavior if seeing - as reported by statsd sinks - Object GET timing sub-10ms and proxy GET timing around 70ms? or is that pretty standard? (the proxies have a mix of slow and fast clients, but i've not been able to determine if that proxy GET timing is actually affected by slow clients) | 20:44 |
| clayg | guitarzan: !!!! | 20:47 |
| guitarzan | clayg: had to come to channel to troll glange :) | 20:47 |
| clayg | totally worth it | 20:47 |
| openstackgerrit | Merged openstack/swift: Fix intermittent container replicator test failure https://review.openstack.org/178116 | 20:48 |
| guitarzan | clayg: I saw your video. you're famous now | 20:49 |
| clayg | guitarzan: nah | 20:57 |
| clayg | swiftstack | 20:57 |
| clayg | wrong window :P | 20:57 |
| *** WBHuber_ has quit IRC | 21:05 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 21:12 |
| *** Nadeem has quit IRC | 21:16 | |
| *** zhill has quit IRC | 21:38 | |
| *** jrichli has quit IRC | 21:41 | |
| *** madgriff has joined #openstack-swift | 21:44 | |
| madgriff | hey guys, I am trying to hit an index file in my swift container from the browser. I have the ACLs set properly. If I try and hit an index file in the root directory (for lack of a better term) of my container I can. I can also hit index files that have approx 5 child directories, but there seems to be a limit | 21:46 |
| *** aix has joined #openstack-swift | 21:46 | |
| madgriff | I cannot hit index files that have say 10 '/' | 21:46 |
| clayg | madgriff: crazy, i'm sure you have the staticweb enabled if you're getting that far | 21:50 |
| clayg | madgriff: what do you get instead of the index file when you try to hit a psudeo dir? just a 404? or a listing? | 21:52 |
| clayg | madgriff: is it possilbe the psudeo dir object *exists* but doesn't have content-type of application/directory? | 21:53 |
| clayg | or whatever you set x-container-meta-web-directory-type too? | 21:53 |
| * clayg is just kinda trolling through the code | 21:53 | |
| clayg | ... I'm not s big staticweb user | 21:54 |
| madgriff | clayg: interestingly enough I get a 512 | 21:54 |
| madgriff | oops | 21:54 |
| madgriff | 412 | 21:54 |
| madgriff | 'Precondition Failed' | 21:54 |
| clayg | that *is* interesting - doesn't look like staticweb will raise that directly | 21:54 |
| clayg | i think utf-8/unicode issues sometimes look like 412'ish | 21:55 |
| madgriff | thats trying to hit something along the lines of <swift-url>/<container>/home/user/really/really/super/long/path/to/copied/content/from/linux/dir/index.html | 21:55 |
| madgriff | but if I upload a simple index.html file to my container and try to hit | 21:56 |
| clayg | oh so that's direct to the index.html | 21:56 |
| madgriff | <swift-url>/<container>/index.html Im good | 21:56 |
| clayg | so the issue isn't just psudeo-dir to index.html redirection - you're acctually failing to hit an object | 21:56 |
| madgriff | or if I even shorten the path | 21:56 |
| madgriff | yep | 21:56 |
| clayg | madgriff: weird - what if you try the url with curl | 21:56 |
| clayg | madgriff: does that object *exist* like you're expecint 2XX or 4XX | 21:57 |
| madgriff | good thinking, let me try that | 21:57 |
| madgriff | 'This server could not verify that you are authorized to access the document you requested.' | 21:58 |
| madgriff | Im basically running a script that copies files from a directory to my swift container, and then I have an application that redirects you to that directories index file in swift when you click on a link | 21:59 |
| madgriff | it seems to work on paths that are relatively short | 21:59 |
| clayg | crazy! and the acl on the container is .r:*? | 21:59 |
| madgriff | but when I hit it with paths that are long, I run into trouble | 21:59 |
| clayg | madgriff: yeah i'm sure that's a clue - i just don't know why it would be | 21:59 |
| madgriff | and Ive verified its the correct path | 21:59 |
| clayg | but 403 response seems like a useful clue too | 21:59 |
| madgriff | yes the acls have been set | 22:00 |
| clayg | like the path is getting translated to a listing or something - and you don't have .rlisting turned on | 22:00 |
| egon | Hmm.. | 22:00 |
| clayg | are you rolling the index.html on your own and uploading it? | 22:00 |
| egon | https://host/AUTH//static-test/1/2/3/4/5/6/7/8/9/0/11/test.html | 22:00 |
| egon | works on my deployment. | 22:00 |
| madgriff | interesting | 22:01 |
| madgriff | clayg: the only time I am manually uploading the index file is when I was trying to verify that I could still hit something in the container | 22:02 |
| madgriff | outside of my script | 22:02 |
| clayg | madgriff: ok, and so *can* you? | 22:02 |
| madgriff | yes | 22:02 |
| clayg | ok, so <swift-url>/<container>/home/user/really/really/super/long/path/to/copied/content/from/linux/dir/index.html is 412/403 dpending on browser/curl but <swift-url>/<container>/home/user/really/really/super/long/path/to/copied/content/from/linux/dir/foo.obj is 2XX everytime? | 22:03 |
| egon | home/user/really/really/super/long/path/to/copied/content/from/linux/dir/test.html actually works for me too. | 22:03 |
| clayg | egon: heh :) | 22:03 |
| madgriff | lol | 22:03 |
| egon | Container: static-test | 22:04 |
| egon | Objects: 5 | 22:04 |
| egon | Bytes: 119907 | 22:04 |
| egon | Read ACL: .r:*,.rlistings | 22:04 |
| egon | Write ACL: | 22:04 |
| egon | Sync To: | 22:04 |
| egon | Sync Key: | 22:04 |
| egon | Accept-Ranges: bytes | 22:04 |
| madgriff | clayg: I can hit <swift-url>/<container>/home/short/path/index.html and <swift-url>/<container>/index.html | 22:04 |
| egon | Connection: keep-alive | 22:04 |
| egon | X-Timestamp: 1430776549.75950 | 22:04 |
| egon | X-Trans-Id: tx2b639dc2a111432797fc0f84fd41c951 | 22:04 |
| egon | Content-Type: text/plain; charset=utf-8 | 22:04 |
| egon | I can change the ACL if that would help testing. | 22:05 |
| *** zhill has joined #openstack-swift | 22:05 | |
| madgriff | interesting when I run swift stat <container> Read ACL and Write ACL are blacnk | 22:05 |
| egon | Oh, that would cause problems. | 22:06 |
| clayg | how strange - i wonder why it works at all? | 22:06 |
| madgriff | not sure | 22:07 |
| madgriff | resetting acl's and now it shows up to 'Read ACL: | 22:07 |
| madgriff | ' | 22:07 |
| madgriff | but still cant hit it | 22:07 |
| clayg | some auth systems will accept auth tokens from cookies - did you "log in" when you first pointed your browser at your cluster? | 22:07 |
| clayg | madgriff: I think egon would suggest the read acl should be at least .r:* | 22:07 |
| madgriff | on the curl: '>The resource could not be found.' | 22:07 |
| madgriff | Read ACL: .r:* | 22:08 |
| clayg | ok, well a 404 makes a little more sense, the index.html does not exist - does that match what you were expecting? | 22:08 |
| madgriff | no :( | 22:08 |
| madgriff | if I list my container it shows up | 22:08 |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Bump up a timeout in a test https://review.openstack.org/179956 | 22:09 |
| clayg | madgriff: are you *sure* :) | 22:10 |
| *** zhill has quit IRC | 22:10 | |
| *** doxavore has quit IRC | 22:17 | |
| mattoliverau | Morning | 22:21 |
| notmyname | mattoliverau: hello! | 22:22 |
| notmyname | mattoliverau: how would you feel if I scheduled your working session at 9am? ie how much would you hate me? | 22:22 |
| notmyname | (nothing is set, just looking now) | 22:23 |
| madgriff | clayg: so I deleted my container, created a new one and set the acl. I then ran my script, tried curl on the url and recieved a 404. I then tried to download the object from the container to verify the object was valid and it worked | 22:30 |
| madgriff | clayg: I then manually created an index.html file, manually uploaded it to my container and hit it from a curl and it worked | 22:31 |
| madgriff | I believe it has to do with either the path or length of path | 22:31 |
| clayg | madgriff: in what respect? what's different about what the script uploaded and curl didn't like - and what you uploaded manullay and curl *did* like? | 22:32 |
| madgriff | nothing aside from the path | 22:34 |
| *** aerwin has quit IRC | 22:34 | |
| clayg | oh... why is the path different? | 22:34 |
| madgriff | i cd into a directory containing an index and upload it, vs the script that uploads it from /path/to/file | 22:35 |
| madgriff | so in the container you have: | 22:35 |
| madgriff | index.html | 22:35 |
| madgriff | and | 22:35 |
| madgriff | '/path/to/file/index.html' | 22:36 |
| clayg | hrmm.... so maybe try to manually upload to '/path/to/file/index.html' | 22:36 |
| mattoliverau | notmyname: someone will need to do a 9am working session so I'm happy to fall on my sword ;) it's fitting really, in SFO mine was first thing while fighting jet lag.. Why not again ;) I'll just need to make sure I'm well cafenated :p | 22:38 |
| notmyname | heh | 22:38 |
| notmyname | thanks :-) | 22:38 |
| clayg | notmyname: i'm not sure why it's only mattoliverau that has to eat the 9am problem - we're all going to have to be there? | 22:38 |
| madgriff | clayg: interesting still a 404 | 22:39 |
| mattoliverau | I have to be more coherent.. At least initially :p | 22:39 |
| notmyname | clayg: mattoliverau: actually, I kinda wanted to start off the day, and all the swift design sessions, with one that (IMO) is pretty interesting. and it's got mattoliverau's name on it | 22:40 |
| peluse | we should just stay up the night before | 22:40 |
| notmyname | also, knowing the individuals, I think mattoliverau would probably do better at 9am than torgomatic would | 22:40 |
| *** tabb has quit IRC | 22:40 | |
| notmyname | peluse: openstack summit...and you thought there was any sleep involved? | 22:40 |
| clayg | madgriff: that is interesting, you might have to do a paste of the commdline output... | 22:40 |
| torgomatic | notmyname: you overestimate my ability to even show up at 9 AM ;) | 22:40 |
| notmyname | torgomatic: exactly. in fact, I didn't overestimate. which is why I put yours later ;-) | 22:41 |
| mattoliverau | Lol | 22:41 |
| notmyname | peluse: don't laugh too much. you've got a 9am on thursday ;-) | 22:41 |
| peluse | argh | 22:42 |
| notmyname | so I discovered that there are _two_ design summit schedules. one, on the main site, has just the conference info | 22:51 |
| notmyname | the other has both the conference and the design summit stuff | 22:51 |
| notmyname | so you can just use one, but if you're like me you've already used the first and will have to migrate stuff around | 22:52 |
| notmyname | the combined summit schedule (now with swift) is at https://libertydesignsummit.sched.org | 22:52 |
| notmyname | well, ok it might take a few minutes for the swift stuff to sync | 22:52 |
| notmyname | http://libertydesignsummit.sched.org/type/design+summit/Swift#.VUf4hc4msyx | 22:54 |
| *** gsilvis has quit IRC | 22:54 | |
| notmyname | I'll be migrating my own schedule and then checking for any conflicts. please let me know if something needs to be changed | 22:55 |
| *** gsilvis has joined #openstack-swift | 23:01 | |
| mattoliverau | notmyname: great, I only knew about the non design track one. | 23:01 |
| *** madgriff has quit IRC | 23:03 | |
| *** evanjfraser has quit IRC | 23:03 | |
| *** zhill has joined #openstack-swift | 23:06 | |
| *** zhill has quit IRC | 23:11 | |
| notmyname | huh. so according to my calendar, the sunday before the summit (may 17) is Swift's birthday (when it was first in prod). find me in vancouver that night, and I'll buy you a drink | 23:18 |
| swifterdarrell | notmyname: swift isn't old enough to drink, you monster! | 23:18 |
| notmyname | you might also be interested in this schedule for the summit: the vBrownBag talks http://openstack.prov12n.com/vbrownbag-techtalks-at-vancouver-summitthe-schedule/ | 23:26 |
| *** chlong has joined #openstack-swift | 23:30 | |
| *** jrichli has joined #openstack-swift | 23:32 | |
| notmyname | hmm...looks like I'll have to do some schedule adjustment. minor | 23:36 |
| notmyname | I had to swap 2 working sessions so we wouldn't overlap with zaqar for the metadata search one | 23:40 |
| *** panbalag has joined #openstack-swift | 23:46 | |
| openstackgerrit | Merged openstack/swift: initial import of go code https://review.openstack.org/178851 | 23:46 |
| jrichli | notmyname: wasn't "Openstack is Doomed ..." one of them you wanted to see? | 23:47 |
| notmyname | yup. unfortunately, that's during one of the fishbowl sessions, so I'll have to watch the recording | 23:48 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!