| *** dmorita has joined #openstack-swift | 00:01 | |
| *** vint_bra has quit IRC | 00:06 | |
| *** jamielennox is now known as jamielennox|away | 00:58 | |
| *** sams-gleb has joined #openstack-swift | 01:00 | |
| *** sams-gleb has quit IRC | 01:04 | |
| *** jamielennox|away is now known as jamielennox | 01:12 | |
| *** klrmn has quit IRC | 01:16 | |
| kota_ | morning | 01:27 |
|---|---|---|
| kota_ | notmyname: try to make a time to look at that patch, but i'm not sure if i can do, looking at my schedule today, a bunch of slots filled by some meetings :,( | 01:29 |
| *** JimCheung has quit IRC | 01:30 | |
| *** zhurong has joined #openstack-swift | 01:30 | |
| mattoliverau | kota_: morning | 01:31 |
| mattoliverau | kota_: I'm looking now :) | 01:31 |
| mattoliverau | just trying to test on a keystone server | 01:31 |
| kota_ | mattoliverau: nice! thanks | 01:32 |
| *** vint_bra has joined #openstack-swift | 01:35 | |
| *** m_kazuhiro has joined #openstack-swift | 01:37 | |
| m_kazuhiro | good morning. | 01:38 |
| kota_ | m_kazuhiro: morning | 01:38 |
| m_kazuhiro | kota_: morning! | 01:39 |
| *** lespaul has joined #openstack-swift | 01:40 | |
| mattoliverau | m_kazuhiro: o/ | 01:48 |
| m_kazuhiro | mattoliverau: o/ | 01:51 |
| lespaul | Hello. To install Swift components from the source rather than Ubuntu repo, do I just use the Github Swift repo for both the proxy and storage nodes? | 01:53 |
| mattoliverau | lespaul: yup | 01:54 |
| *** dmorita has quit IRC | 01:55 | |
| mattoliverau | You can, pip install -r requirements.txt; pip install -e . (For editable mode [Dev work]) | 01:55 |
| *** JimCheung has joined #openstack-swift | 01:55 | |
| mattoliverau | Or without the -e | 01:56 |
| mattoliverau | Then you should have it installed so you can use the Swift commands (swift-init, swift-ring-builder etc) | 01:57 |
| lespaul | Thanks. | 01:57 |
| *** JimCheung has quit IRC | 02:00 | |
| *** vint_bra has quit IRC | 02:00 | |
| *** dmorita has joined #openstack-swift | 02:13 | |
| *** tanee_away is now known as tanee | 02:13 | |
| *** dmorita has quit IRC | 02:17 | |
| *** klrmn has joined #openstack-swift | 02:42 | |
| *** sams-gleb has joined #openstack-swift | 03:02 | |
| *** sams-gleb has quit IRC | 03:06 | |
| *** dmorita has joined #openstack-swift | 03:14 | |
| *** dmorita has quit IRC | 03:19 | |
| *** chsc has joined #openstack-swift | 03:33 | |
| *** janonymous has joined #openstack-swift | 03:37 | |
| *** chsc has quit IRC | 03:38 | |
| *** gkadam has joined #openstack-swift | 03:43 | |
| *** dmorita has joined #openstack-swift | 03:56 | |
| *** m_kazuhiro has quit IRC | 03:57 | |
| *** dmorita has quit IRC | 04:00 | |
| *** m_kazuhiro has joined #openstack-swift | 04:03 | |
| zaitcev | Gerrit is dead? | 04:09 |
| mattoliverau | yeah, seems to be | 04:09 |
| zaitcev | oh well, time for some anime and sleep then | 04:09 |
| mattoliverau | zaitcev: :) | 04:10 |
| notmyname | mattoliverau: don't know, but I jsut got four gerrit emails with your same review comment | 04:11 |
| mattoliverau | notmyname: ok, so it down, but post must be working.. opps | 04:11 |
| mattoliverau | anyway, after finally deploying a local keystone server (v3), i can still auth with swiftclient as v2 or v3 with the new patch, but joel says he's having trouble.. so wanted to touch base before I do anything. | 04:12 |
| mattoliverau | also, I took notes, so might write a script to deploy keystone with just a swift setup for anyone interested for keystone + swift testing (both from source code). | 04:14 |
| notmyname | nnice | 04:14 |
| zaitcev | there's a script in swift3 that seems nice for that, better than my old one for sure | 04:15 |
| zaitcev | https://github.com/openstack/swift3/blob/master/swift3/test/functional/setup_keystone | 04:16 |
| mattoliverau | zaitcev: oh really, cool. | 04:19 |
| zaitcev | I'm not linking you my script because this is clearly better | 04:19 |
| mattoliverau | I was following using uwsgi as they recommend on the keystone dev documentation.. but wow, that couldn've saved me some time :P But on the plus side now, I know how to do it. And my notes goes into more details like setting up the catalog etc. | 04:21 |
| mattoliverau | ie I mean multi endpoints etc. so yeah maybe the swift3 one would have been _way_ more then enough :P | 04:22 |
| *** psachin has joined #openstack-swift | 04:25 | |
| *** links has joined #openstack-swift | 04:35 | |
| *** JimCheung has joined #openstack-swift | 04:36 | |
| *** JimCheung has quit IRC | 04:41 | |
| *** sams-gleb has joined #openstack-swift | 05:04 | |
| *** sams-gleb has quit IRC | 05:08 | |
| *** dmorita has joined #openstack-swift | 05:16 | |
| *** dmorita has quit IRC | 05:20 | |
| *** JimCheung has joined #openstack-swift | 05:22 | |
| *** JimCheung has quit IRC | 05:27 | |
| *** janonymous has quit IRC | 05:44 | |
| *** adriant has quit IRC | 06:11 | |
| *** bkopilov has joined #openstack-swift | 06:14 | |
| *** JimCheung has joined #openstack-swift | 06:16 | |
| *** JimCheung has quit IRC | 06:20 | |
| *** gyee has joined #openstack-swift | 06:45 | |
| *** gyee has quit IRC | 06:46 | |
| *** gyee has joined #openstack-swift | 06:49 | |
| *** gyee has quit IRC | 06:50 | |
| *** tone_zrt has joined #openstack-swift | 06:56 | |
| *** m_kazuhiro has quit IRC | 07:03 | |
| *** sams-gleb has joined #openstack-swift | 07:06 | |
| *** sams-gleb has quit IRC | 07:10 | |
| openstackgerrit | Béla Vancsics proposed openstack/swift master: Reduced the complexity of the _response_iter method https://review.openstack.org/426782 | 07:11 |
| *** tesseract has joined #openstack-swift | 07:20 | |
| openstackgerrit | Merged openstack/swift master: Update calling super class constructor style in proxy controllers https://review.openstack.org/444748 | 07:38 |
| *** hseipp has joined #openstack-swift | 07:38 | |
| *** sams-gleb has joined #openstack-swift | 07:50 | |
| *** pcaruana has joined #openstack-swift | 07:50 | |
| *** dmorita has joined #openstack-swift | 08:28 | |
| *** dmorita has quit IRC | 08:33 | |
| *** furlongm has quit IRC | 08:35 | |
| *** amoralej|off is now known as amoralej | 08:39 | |
| *** murugesh has joined #openstack-swift | 08:45 | |
| murugesh | Hi There | 08:48 |
| murugesh | I have issue in accessing the swift in UI | 08:48 |
| murugesh | when i click on object-store and then containers it is redirected to 404 not found page | 08:50 |
| murugesh | The requested URL /auth/logout was not found on this server. | 08:50 |
| *** janonymous has joined #openstack-swift | 08:56 | |
| *** gkadam_ has joined #openstack-swift | 09:07 | |
| *** gkadam has quit IRC | 09:07 | |
| *** chmouel_ is now known as chmouel | 09:08 | |
| *** gkadam_ is now known as gkadam | 09:08 | |
| *** gkadam is now known as Guest58076 | 09:09 | |
| *** Guest58076 is now known as gkadam | 09:09 | |
| acoles | good morning | 09:09 |
| *** oshritf_ has joined #openstack-swift | 09:13 | |
| *** oshritf has quit IRC | 09:16 | |
| *** geaaru has joined #openstack-swift | 09:32 | |
| *** cbartz has joined #openstack-swift | 09:48 | |
| *** zhurong has quit IRC | 09:54 | |
| *** gkadam_ has joined #openstack-swift | 10:10 | |
| *** gkadam has quit IRC | 10:13 | |
| *** SJ has joined #openstack-swift | 10:14 | |
| SJ | ?join #openstack-swift | 10:15 |
| SJ | ?join #openstack | 10:16 |
| *** dmorita has joined #openstack-swift | 10:16 | |
| murugesh | Hi +acoles | 10:17 |
| murugesh | good morning | 10:18 |
| *** dmorita has quit IRC | 10:21 | |
| *** neonpastor has quit IRC | 10:25 | |
| *** sams-gle_ has joined #openstack-swift | 10:28 | |
| *** sams-gleb has quit IRC | 10:30 | |
| *** JimCheung has joined #openstack-swift | 10:33 | |
| *** JimCheung has quit IRC | 10:37 | |
| *** dmorita has joined #openstack-swift | 10:40 | |
| *** dmorita has quit IRC | 10:45 | |
| *** acoles_ has joined #openstack-swift | 11:08 | |
| *** sams-gle_ has quit IRC | 11:08 | |
| *** sams-gleb has joined #openstack-swift | 11:09 | |
| *** acoles_ has quit IRC | 11:10 | |
| *** acoles_ has joined #openstack-swift | 11:10 | |
| *** ChanServ sets mode: +v acoles_ | 11:10 | |
| *** sams-gleb has quit IRC | 11:13 | |
| mahatic | acoles: o/ | 11:23 |
| acoles_ | mahatic: hi | 11:23 |
| mahatic | you have two nicks! | 11:24 |
| mahatic | active | 11:24 |
| acoles_ | yeah! :) | 11:24 |
| mahatic | lol, isn't this your away nick? ;) | 11:24 |
| acoles_ | I'm trying out a different irc client. I was using xchat but switched to OSX and the xchat client has a couple of annoying quirks, so i am trying textual | 11:25 |
| mahatic | "acoles is now acoles_" is what I used to read :P | 11:25 |
| mahatic | oic | 11:25 |
| acoles_ | so you should never see "acoles is now acoles_" anymore, I learnt that away nicks are deprecated ! | 11:26 |
| acoles_ | so switched that off | 11:26 |
| mahatic | oh, what bouncer do you use? | 11:26 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 11:29 |
| *** sams-gleb has joined #openstack-swift | 11:30 | |
| *** acoles_ has quit IRC | 11:32 | |
| *** acoles_ has joined #openstack-swift | 11:33 | |
| *** ChanServ sets mode: +v acoles_ | 11:33 | |
| *** gkadam_ is now known as gkadam | 11:34 | |
| *** links has quit IRC | 11:36 | |
| *** gkadam_ has joined #openstack-swift | 11:37 | |
| *** gkadam has quit IRC | 11:39 | |
| *** links has joined #openstack-swift | 11:48 | |
| *** dmorita has joined #openstack-swift | 11:52 | |
| *** dmorita has quit IRC | 11:57 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Add more doc and test for cors_expose_headers option https://review.openstack.org/445021 | 11:58 |
| *** gkadam__ has joined #openstack-swift | 12:01 | |
| *** gkadam_ has quit IRC | 12:03 | |
| *** gabor_antal has joined #openstack-swift | 12:07 | |
| *** vint_bra has joined #openstack-swift | 12:12 | |
| *** janonymous has quit IRC | 12:14 | |
| murugesh | Hi +acoles | 12:19 |
| *** vint_bra has quit IRC | 12:22 | |
| *** hseipp has quit IRC | 12:29 | |
| acoles_ | murugesh: hi | 12:29 |
| murugesh | I have installed and configured swift-proxy and nodes successfully this time | 12:30 |
| acoles_ | great! | 12:30 |
| murugesh | and i am able to create containers and upload objects into containers | 12:31 |
| murugesh | but the problem is when i refresh the web UI its gone and not appearing | 12:32 |
| murugesh | when i tail the log "/var/log/swift/swift.log i get the following error | 12:32 |
| patchbot | Error: No closing quotation | 12:32 |
| murugesh | http://paste.openstack.org/show/602673/ | 12:33 |
| murugesh | could you pls help me with this?? | 12:34 |
| acoles_ | I can't see an error in that paste ^^ | 12:34 |
| murugesh | What does it mean from the paste "Deferring reject downstream" | 12:43 |
| *** kei_yama has quit IRC | 12:43 | |
| acoles_ | It means that authtoken has failed to authorize the request but is giving an opportunity for other auth mechanism to authorize such as ACLs. Controlled by this option https://github.com/openstack/swift/blob/a7fa7900fc1ab0ed393b6dc30714213aedc5c15c/etc/proxy-server.conf-sample#L352-L355 | 12:46 |
| *** klamath has joined #openstack-swift | 12:47 | |
| *** klamath has quit IRC | 12:47 | |
| *** klamath has joined #openstack-swift | 12:47 | |
| *** links has quit IRC | 12:52 | |
| murugesh | so how can i fix this issue?? | 12:52 |
| *** links has joined #openstack-swift | 12:53 | |
| *** amoralej is now known as amoralej|lunch | 12:53 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift master: Limit number of revert tombstone SSYNC requests https://review.openstack.org/439572 | 12:55 |
| murugesh | +acoles, The option in proxy-server.conf already set like this "delay_auth_decision = True" | 12:58 |
| *** catintheroof has joined #openstack-swift | 12:58 | |
| *** catintheroof has quit IRC | 12:59 | |
| *** catintheroof has joined #openstack-swift | 12:59 | |
| acoles_ | murugesh: it's not a problem - you have the correct config and the log message is not an error | 13:01 |
| murugesh | The actual problem is i am able to create the container,folder and upload the files successfully | 13:02 |
| murugesh | but when i refresh the web UI(horizon) it disappears | 13:02 |
| murugesh | what might be the problem here?? | 13:02 |
| acoles_ | murugesh: sorry, I am not familiar with the horizon UI | 13:04 |
| *** chlong has joined #openstack-swift | 13:04 | |
| *** dmorita has joined #openstack-swift | 13:04 | |
| acoles_ | maybe try #openstack-horizon | 13:05 |
| *** silor has joined #openstack-swift | 13:05 | |
| robcresswell | murugesh: It probably just means that the UI believes the request to upload was successful, but the async nature of most resource creation means its not easy to track | 13:05 |
| robcresswell | So probably just gets a response from Swift that says "yeah we accept this request" and shows it in the UI. | 13:05 |
| robcresswell | When you refresh its checking the actual list and therefore nothing is there | 13:06 |
| murugesh | yeah | 13:06 |
| murugesh | what am i doing wrong here? | 13:06 |
| murugesh | #robcresswell, if i did not refresh also it disappears in sometime | 13:07 |
| acoles_ | murugesh: are you able to do a curl GET request to the container? if so do you get back a listing with your object? or using the swiftclient CLI 'swift list <containername>' | 13:07 |
| acoles_ | oh, maybe I misunderstood, so you see the object listed in the container, but then is disappears? | 13:08 |
| murugesh | yeah total container itself gone | 13:09 |
| robcresswell | murugesh: Yeah, because it doesnt actually exist in Swift | 13:09 |
| robcresswell | tries to create, fails, shows nothing | 13:09 |
| *** dmorita has quit IRC | 13:09 | |
| murugesh | +acoles, when i do "swift list" i get the container list first time and if i rerun the command i don't get any response | 13:10 |
| murugesh | #robcresswell, I don't see anything in the /var/log/swift/swift.log other than this | 13:11 |
| murugesh | http://paste.openstack.org/show/602673/ | 13:11 |
| acoles_ | murugesh: does "don't get any response" mean the command never returns?? or it returns an empty list? | 13:13 |
| *** silor1 has joined #openstack-swift | 13:13 | |
| *** silor has quit IRC | 13:13 | |
| *** silor1 is now known as silor | 13:13 | |
| murugesh | its an empty list | 13:13 |
| murugesh | the time i get empty list through command line i don't see the container in web UI too | 13:14 |
| acoles_ | murugesh: each time you do a list (container GET) the request may be forwarded to a different backend container server (assuming you have more than one replica), so it could be that one of your container servers is not getting object updates, whereas the other is - if you repeated 'swift list' then sometimes you'd see the object , sometimes not. | 13:16 |
| * acoles_ away | 13:16 | |
| murugesh | yeah we have 3 storage nodes for swift | 13:17 |
| murugesh | I don't find any solution from googling this issue | 13:21 |
| murugesh | do you any ideas +acoles & robcresswell?? | 13:22 |
| robcresswell | Nope, I'm a Horizon guy, not a Swift guy :) | 13:23 |
| murugesh | Thanks for your help robcresswell. | 13:26 |
| thurloat | murugesh do you never get the list again after running once? or is it deterministically every second time? | 13:27 |
| *** vint_bra has joined #openstack-swift | 13:28 | |
| thurloat | if you run `swift list` 10 times, what is the result? if it's on and off again -- most likely acoles_ is correct. one of your container servers is upset. | 13:28 |
| thurloat | so maybe check the rsyncd log for replication activity | 13:28 |
| murugesh | thurloat, As i told earlier if did not refresh also the entire containers,folders,files disappears after sometime of creation | 13:28 |
| murugesh | i am checking on rsyncd log on all three storage nodes | 13:30 |
| *** JimCheung has joined #openstack-swift | 13:31 | |
| *** JimCheung has quit IRC | 13:35 | |
| murugesh | @thurload, I don't get any logs in /var/log/rysncd.log in storage nodes | 13:36 |
| murugesh | but i see some 404 GET response in /var/log/swift/swift.log on swift proxy server | 13:37 |
| murugesh | please find the paste link below | 13:37 |
| thurloat | can you paste.ec one of your rsyncd.conf files? | 13:38 |
| murugesh | http://paste.openstack.org/show/602685/ | 13:38 |
| murugesh | Sure will paste the rysnc.conf file of first node | 13:38 |
| murugesh | @thurloat, here is the rsync.conf file | 13:41 |
| murugesh | http://paste.openstack.org/show/602686/ | 13:41 |
| *** catinthe_ has joined #openstack-swift | 13:41 | |
| -openstackstatus- NOTICE: Gerrit is going to be restarted due to performance problems | 13:41 | |
| *** ChanServ changes topic to "Gerrit is going to be restarted due to performance problems" | 13:41 | |
| *** catintheroof has quit IRC | 13:41 | |
| thurloat | and silly question: your rsync service is running? | 13:42 |
| murugesh | yeah thurload rsync service is up and running | 13:44 |
| *** catintheroof has joined #openstack-swift | 13:45 | |
| *** catinthe_ has quit IRC | 13:45 | |
| -openstackstatus- NOTICE: Gerrit has been successfully restarted | 13:48 | |
| *** ChanServ changes topic to "Gerrit has been successfully restarted" | 13:48 | |
| *** SJ has quit IRC | 13:50 | |
| *** hseipp has joined #openstack-swift | 13:52 | |
| thurloat | quite strange, there should be rsync activity after writing data to either account/container/object storage from the server that accepted the write. | 13:53 |
| *** ChanServ changes topic to "Gerrit is going to be restarted due to performance problems" | 13:54 | |
| -openstackstatus- NOTICE: Gerrit has been successfully restarted | 13:54 | |
| *** links has quit IRC | 13:57 | |
| *** gkadam__ has quit IRC | 14:01 | |
| *** jaosorior has joined #openstack-swift | 14:03 | |
| murugesh | I am also confused with this thurloat | 14:03 |
| jaosorior | Joel Wright, are you around? | 14:04 |
| *** silor has quit IRC | 14:06 | |
| *** baffle has joined #openstack-swift | 14:06 | |
| *** amoralej|lunch is now known as amoralej | 14:08 | |
| baffle | Hi, I have a problem retrieving a DLO. It is 131G combined of 13000+ segments; First a stat on the DLO would say "Content Length: 0". After looking around, we changed the swift-constraint -> container_listing_limit to 200000 from the default 10000. Now Content Length reports as 104857600000, but not the full size of the DLO. Ranged requests over 104857600000 fails with 416. Before we changed | 14:11 |
| baffle | container_listing_limit, requests would return 200 and the whole object instead. Ideas? | 14:11 |
| *** joeljwright has joined #openstack-swift | 14:14 | |
| *** ChanServ sets mode: +v joeljwright | 14:14 | |
| openstackgerrit | Merged openstack/swift master: Fixed indentation in common/ and obj/ https://review.openstack.org/429750 | 14:18 |
| baffle | 1024*1024 is 1048576 and the Content Length is magically capped at 104857600000. I feel some magic here. :) | 14:21 |
| *** murugesh has quit IRC | 14:30 | |
| openstackgerrit | Adam Thurlow proposed openstack/swift master: Un-delete accounts that were previously marked as Deleted for the reaper. https://review.openstack.org/445160 | 14:30 |
| timss | baffle: Strange indeed! I don't know too much about these internals, but this part of the DLO middleware seem to possibly match your 200/416 scenario, or? | 14:39 |
| timss | https://github.com/openstack/swift/blob/e8a80e874a5086e94c5ae93e3a9191cb813d1631/swift/common/middleware/dlo.py#L265 | 14:39 |
| *** sams-gleb has quit IRC | 14:41 | |
| *** sams-gleb has joined #openstack-swift | 14:42 | |
| *** sams-gleb has quit IRC | 14:46 | |
| *** sams-gleb has joined #openstack-swift | 14:59 | |
| baffle | timss: I figured it out. DLO middleware was the initial hurdle, but then swift-container-server also used the same limit. | 15:23 |
| *** psachin has quit IRC | 15:25 | |
| *** SkyRocknRoll has joined #openstack-swift | 15:32 | |
| notmyname | good morning | 15:37 |
| thurloat | morning notmyname | 15:39 |
| thurloat | actually, afternoon :) | 15:39 |
| *** bikmak has joined #openstack-swift | 15:42 | |
| notmyname | joeljwright: jaosorior: hello. hoped to catch you before the end of the day | 15:43 |
| joeljwright | hello | 15:43 |
| jaosorior | notmyname: hey | 15:43 |
| notmyname | joeljwright: jaosorior: although I haven't dug into the code, I've been following the conversation on https://review.openstack.org/#/c/443104/ | 15:43 |
| patchbot | patch 443104 - python-swiftclient - Use generic keystone client instead of versioned one | 15:43 |
| timss | baffle: Ah ok, nice to know. Would that happen to explain the 1024*1024 magic though? | 15:43 |
| joeljwright | notmyname: yeah, we had a separate IRC chat too | 15:44 |
| notmyname | joeljwright: jaosorior: and despite any (mis?) config issues in joeljwright's prod keystone, it seems that's actually a thing, and working for him since this is the first issue he's had. so I'm happy to see that jaosorior is looking at adding a code workaround | 15:45 |
| notmyname | joeljwright: great :-) | 15:45 |
| joeljwright | I'm guessing I'm not the only person with an old/misconfigured keystone | 15:45 |
| notmyname | hah! unlikely :-) | 15:45 |
| notmyname | oh, sorry. rgiht | 15:45 |
| thurloat | joeljwright how old is your keystone? | 15:46 |
| notmyname | missed the "not" | 15:46 |
| jaosorior | notmyname: I'm trying to come up with a workaround in keystoneclient. I can mark that swiftclient commit as WIP in the meantime | 15:46 |
| baffle | timss: Yes. At first it just failed at calculating any kind of content length, due to there being more then 10000 objects in the DLO, thus reporting 0. That might acutally indicate infinity, not sure. :) Then, the DLO middleware could parse more than 10000 objects, but container server only returned 10000 of the 13000. The objects were exactly 10485760 in length each. So it was a bit confusing. | 15:46 |
| joeljwright | thurloat: I think it's icehouse | 15:46 |
| notmyname | jaosorior: ok. probably a good idea. thanks | 15:46 |
| thurloat | joeljwright juno here :) | 15:47 |
| joeljwright | :) | 15:47 |
| baffle | Juno or Kilo Keystone live, Mitaka in testing. | 15:47 |
| baffle | And PKI tokens. | 15:47 |
| baffle | \o/ | 15:48 |
| * thurloat running nova grizzly *flexes* | 15:48 | |
| joeljwright | jaosorior: I hope we're not depressing you too much :S | 15:48 |
| *** tonanhngo has joined #openstack-swift | 15:48 | |
| jaosorior | haha well, you can still workaround this issue. At that point keystone had another config option | 15:48 |
| jaosorior | other than the oslo.middleware one | 15:48 |
| jaosorior | public_endpoint I think it was called. And that should contain the proper scheme | 15:49 |
| baffle | Has anyone looked at implementing ACLs enables features like "Allow read/write from clients connecting from <neutron-network-id>"? | 15:51 |
| timss | baffle: Aha makes sense, thanks for the explaination. Looking forward to the day this happens to my system too now ..! :) | 15:52 |
| baffle | timss: It's not that fun. It hit a customer that used our Swift for image backups of servers using a backup client; Their customer had a cryptolocker-party on one of their main storage servers. Pretty important data as well, and they needed a restore *now*. \o/ | 15:54 |
| baffle | timss: Think HIPAA. | 15:55 |
| notmyname | baffle: we've got ACLs that look at the referer header. and expanding that to include a source IP whitelist has been discussed but not implemented | 15:56 |
| zaitcev | referrer is not reliable | 16:07 |
| zaitcev | er, not secure | 16:07 |
| notmyname | zaitcev: totally not! it's user-set | 16:07 |
| notmyname | timss: baffle: I'm just catching up on my IRC buffer playback from overnight. the issue is a large DLO (lots of segments) isn't letting you get the full object back? | 16:07 |
| zaitcev | So, I would dedicate a proxy to this function and call its region "HIPAA" in Keystone, then firewall it from the outside world, and only allow access from that Neutron network. | 16:08 |
| notmyname | zaitcev: +1 | 16:08 |
| baffle | notmyname: Yeah, there has to be some magic there, in how proxy magically knows what neutron net the request is coming from. Maybe investigate network-local services instead? I.e. like how the metadata services work today. :) | 16:10 |
| baffle | notmyname: Solved the DLO issue allready, and also informed the vendor about the defaults that others probably will hit at one point in time. | 16:10 |
| notmyname | baffle: oh wow. so you want to restrict it to a sertain subset of IPs but that subset is dynamic based on what's been provisioned in neutron? | 16:11 |
| zaitcev | yeah, I didn't expect that | 16:11 |
| notmyname | baffle: ah, good to hear about the DLOs. (personally, I'd suggest using SLOs instead) | 16:11 |
| joeljwright | notmyname: +1 on SLO :) | 16:12 |
| baffle | I guess something like this: https://aws.amazon.com/blogs/aws/new-vpc-endpoint-for-amazon-s3/ | 16:13 |
| baffle | notmyname: Actually, DLOs work great for this purpose. I was a bit amazed. | 16:13 |
| *** SkyRocknRoll has quit IRC | 16:13 | |
| notmyname | baffle: yeah, DLOs can work great :-) | 16:14 |
| notmyname | baffle: but they can have some odd downsides too, and normally the more explicit SLOs make everyone happier | 16:14 |
| timss | baffle: Oh, ouch. Curious now how one would tackle even larger objects/more segments, and where the pain point is with regards to listing/fetching tens of thousands of segments. Larger segments is one obvious alternative of course, but what if one lets the client/customer "go wild"? :) | 16:15 |
| jaosorior | joeljwright, notmyname: So, this is the WIP patch I have right now for keystoneclient https://review.openstack.org/445559 I'm poking the keystone folks to see what they think. And still need to add tests and such. | 16:15 |
| patchbot | patch 445559 - python-keystoneclient - Workaround for unmatching scheme in discovery | 16:15 |
| notmyname | timss: you can nest SLOs | 16:16 |
| *** SkyRocknRoll has joined #openstack-swift | 16:17 | |
| timss | notmyname: Aye, but if we're talking DLO | 16:19 |
| notmyname | researching code before I talk, because I always get this part wrong ... | 16:20 |
| *** JimCheung has joined #openstack-swift | 16:23 | |
| notmyname | timss: yeah, DLOs can't be nested | 16:27 |
| *** bikmak has quit IRC | 16:28 | |
| notmyname | timss: here's my attempt against my dev box https://gist.github.com/notmyname/6cae683c43816384a3488050e0f99d41 | 16:28 |
| jaosorior | joeljwright, notmyname: So... one keystone folk's answer is "fix your keystone" (which you could do using the public_endpoint and admin_endpoint configuration parameters) | 16:32 |
| jaosorior | however | 16:32 |
| jaosorior | he hasn't -1ed the patch :/ .... so, maybe there's still hope? | 16:33 |
| *** dmorita has joined #openstack-swift | 16:33 | |
| *** chsc has joined #openstack-swift | 16:35 | |
| *** chsc has joined #openstack-swift | 16:35 | |
| timburke | i'm trying to think of a case where you'd really *need* DLOs to be nested. it's not like we have the same length limits that we have for SLOs; even if it takes more than one container listing, we'll stream out all the objects with that prefix... | 16:36 |
| notmyname | timburke: yeah, I was just getting to that same conclusion myself | 16:37 |
| notmyname | however, if timss *isn't* seeing all the segments after the first page of listing results... | 16:37 |
| notmyname | it looks like the code should be handling more than one page of results | 16:38 |
| *** acoles_ has quit IRC | 16:44 | |
| timss | notmyname: It wasn't my intention to suggest nested DLOs, but cool to see what happened anyway! :) | 16:46 |
| notmyname | timss: :-) | 16:46 |
| timss | baffle were the one who the large DLO object btw, but I'm expecting I'm gonna hit something similar one day as well. | 16:46 |
| timss | Didn't seem like the client and/or container was able to handle pages of listings though, from what he described | 16:47 |
| *** acoles_ has joined #openstack-swift | 16:50 | |
| *** ChanServ sets mode: +v acoles_ | 16:50 | |
| *** acoles_ has quit IRC | 16:51 | |
| *** acoles_ has joined #openstack-swift | 16:52 | |
| *** ChanServ sets mode: +v acoles_ | 16:52 | |
| cbartz | Hello, any core reviewer (except timburke who has already reviewed) here for patch 414232 ? Would be great. | 16:55 |
| patchbot | https://review.openstack.org/#/c/414232/ - swift - Exclude containers for account quota | 16:55 |
| *** d0ugal has quit IRC | 16:55 | |
| *** _JZ_ has joined #openstack-swift | 16:56 | |
| *** cbartz has quit IRC | 16:59 | |
| joeljwright | jaosorior: I thought that might be the answer | 17:08 |
| jaosorior | joeljwright: I hope that the dude that has more experience with the discovery side in keystone will be more reasonable. | 17:09 |
| joeljwright | let's hope! | 17:10 |
| jaosorior | joeljwright: in the meantime, have you checked on the values in public_endpoint and admin_endpoint in your keystone configuration? Just so this doesn't affect you in other components. | 17:11 |
| *** chsc has quit IRC | 17:12 | |
| *** chsc has joined #openstack-swift | 17:16 | |
| openstackgerrit | Merged openstack/swift master: refactor some common code from crypto https://review.openstack.org/439058 | 17:22 |
| *** hseipp has quit IRC | 17:23 | |
| *** bkopilov has quit IRC | 17:27 | |
| *** oshritf__ has joined #openstack-swift | 17:31 | |
| *** gabor_antal_ has joined #openstack-swift | 17:31 | |
| *** gabor_antal has quit IRC | 17:31 | |
| *** oshritf_ has quit IRC | 17:34 | |
| *** bkopilov has joined #openstack-swift | 17:38 | |
| joeljwright | timburke: did you have a chance to look at the SLO pre/postamble patch? | 17:44 |
| *** jaosorior has quit IRC | 17:45 | |
| timburke | joeljwright: ugh. i did not :-( sorry | 17:45 |
| joeljwright | timburke: no problem | 17:47 |
| joeljwright | timburke: I'm going to keep an eye on any updates, but I'm away from the office for the next few weeks | 17:49 |
| *** tesseract has quit IRC | 17:49 | |
| *** d0ugal has joined #openstack-swift | 17:59 | |
| acoles_ | joeljwright: skiing? | 18:00 |
| *** pcaruana has quit IRC | 18:03 | |
| *** d0ugal has quit IRC | 18:04 | |
| *** mgagne_ is now known as mgagne | 18:05 | |
| *** geaaru has quit IRC | 18:08 | |
| *** gabor_antal_ is now known as gabor_antal | 18:13 | |
| joeljwright | acoles_: shhh! | 18:20 |
| joeljwright | don't tell everyone! | 18:20 |
| joeljwright | acoles_: yeah, I'm ski guiding for 3 weeks :) | 18:21 |
| acoles_ | life's tough :) | 18:22 |
| joeljwright | :) | 18:26 |
| *** joeljwright has quit IRC | 18:26 | |
| *** spotz is now known as spotz_zzz | 18:32 | |
| *** spotz_zzz is now known as spotz | 18:33 | |
| baffle | timss: Yeah, I contacted the vendor and asked them to look into changing the default segment size, this can't be that uncommon. | 18:34 |
| *** acoles_ has quit IRC | 18:48 | |
| *** dmorita has quit IRC | 18:52 | |
| *** dmorita has joined #openstack-swift | 18:56 | |
| *** dmorita has quit IRC | 19:04 | |
| *** ganders has joined #openstack-swift | 19:28 | |
| *** amoralej is now known as amoralej|off | 19:43 | |
| notmyname | reminder (mostly for those in the USA): tomorrows meeting is still at 2100UTC. which might be different for you this week. UTC doesn't do time changes. it's at 2pm pacific, 5pm eastern | 19:46 |
| *** dmorita has joined #openstack-swift | 19:48 | |
| *** ganders has quit IRC | 19:55 | |
| MooingLemur | yay, Arizona | 20:09 |
| notmyname | yeah, I have no idea about what time it is in AZ. ;-) | 20:18 |
| *** lespaul_ has joined #openstack-swift | 20:18 | |
| tdasilva | i think they don't observe daylight savings, so meeting time never changes for them??? | 20:19 |
| lespaul_ | Hello. Before installing Swift from the source/Github for the proxy and storage nodes, any dependencies that needs to be met first? How about pyeclib and liberasurecode? I'm getting a dnspython error when running the Swift setup.py install, hence the question. Thanks. | 20:28 |
| notmyname | lespaul_: there's the requirements.txt file which has dependencies that can be found on pypi. then there's bindep.txt that lists non-python dependencies | 20:30 |
| lespaul_ | notmyname: Got it, thanks. | 20:35 |
| *** tonanhngo has quit IRC | 21:04 | |
| *** _JZ_ has quit IRC | 21:13 | |
| *** silor has joined #openstack-swift | 21:14 | |
| *** silor has quit IRC | 21:20 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** Jeffrey4l has joined #openstack-swift | 21:35 | |
| mattoliverau | Morning | 21:44 |
| *** vint_bra has quit IRC | 21:46 | |
| *** catintheroof has quit IRC | 21:46 | |
| thurloat | notmyname added some tests to patch 445160 and moved it out of WIP (i think) | 22:13 |
| patchbot | https://review.openstack.org/#/c/445160/ - swift - Un-delete accounts that were previously marked as ... | 22:13 |
| *** ChanServ changes topic to "Topic: Let's talk, we're nice. | PTG etherpad: https://etherpad.openstack.org/p/swift-ptg-pike | Logs: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/ | Meetings: https://wiki.openstack.org/wiki/Meetings/Swift | Priority Reviews: https://wiki.openstack.org/wiki/Swift/PriorityReviews" | 22:42 | |
| *** sams-gleb has quit IRC | 22:53 | |
| *** Renich has joined #openstack-swift | 23:01 | |
| Renich | o/ | 23:01 |
| Renich | I need some help translating v3 access into v1. I know how to access my Object Storage via V3, but a user requires PHP: https://github.com/php-opencloud/openstack/blob/master/samples/object_store/v1/containers/list.php and the only example there is using V1. | 23:02 |
| Renich | Would somebody help me making the translation? | 23:02 |
| Renich | I need to know how to connect using v1 | 23:03 |
| *** klamath has quit IRC | 23:06 | |
| timburke | Renich: the V1 there is referring to the object-store API version (i believe) -- the auth API version (which i think is what you're looking for?) is separate | 23:11 |
| notmyname | Renich: yeah, i think you're looking for something under https://github.com/php-opencloud/openstack/tree/master/samples/identity | 23:12 |
| timburke | that said, the auth details look like v3 auth already, +/- some domains | 23:12 |
| notmyname | timburke: you mean "v2" ;-) | 23:12 |
| Renich | timburke: ah! I missed that... sorry :S | 23:12 |
| notmyname | timburke: isn't v3 just v2 + domains? :-) | 23:12 |
| timburke | notmyname: well, there's also the project vs. tenant thing. with no domain, it falls back to some default one, right? | 23:13 |
| notmyname | probably? | 23:14 |
| Renich | timburke: well, that doesn't seem to be the answer. Those examples are from CRUD users and stuff. | 23:14 |
| Renich | timburke: I will try and dig the code to see if I can learn something | 23:14 |
| Renich | Code is very well documented... phew | 23:16 |
| timburke | hmm. the auth details for https://github.com/php-opencloud/openstack/blob/master/samples/identity/v3/users/list_users.php look similar... but the example doesn't work with v3? | 23:16 |
| Renich | timburke: I dunno what to set in region. Nor if "project name" works as "project id" | 23:17 |
| Renich | lines 5 through 17 | 23:17 |
| Renich | How can a user get info on his project/account? openstack project list doesn't work for a normal user | 23:19 |
| timburke | you can probably leave out region; i would expect it to select one automatically (likely just the first one from the service catalog). you're probably right about project name vs. id -- but i'm willing to be you could have something like `'scope' => ['project' => ['name' => '{projectName}', 'domain' => <something>]]` | 23:19 |
| Renich | timburke: probablly... will do trian and error, hehe | 23:20 |
| *** kei_yama has joined #openstack-swift | 23:29 | |
| *** chsc has quit IRC | 23:30 | |
| *** tonanhngo has joined #openstack-swift | 23:35 | |
| mattoliverau | fyi I wrote down how I setup keystone in a SAIO in case others want to ever do that thing. https://oliver.net.au/?p=313 it sets it up with keystone admin and public ports etc. Mainly did it for myself so I don't have to re-learn later :P | 23:39 |
| tdasilva | mattoliverau: awesome!!! now we just need to do that in swift-ansible | 23:45 |
| mattoliverau | tdasilva: yeah, now that I have the steps down, shouldn't be too hard ;) | 23:46 |
| tdasilva | mattoliverau: thanks for sharing that | 23:47 |
| mattoliverau | its also a little rough, probably should have given it a read before I hit publish but meh :P | 23:48 |
| tdasilva | mattoliverau: no worries, i know where to find the author if a question comes up :) | 23:49 |
| mattoliverau | :P | 23:49 |
| *** sams-gleb has joined #openstack-swift | 23:53 | |
| *** adriant has joined #openstack-swift | 23:58 | |
| *** sams-gleb has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!