| *** tgohad has quit IRC | 00:03 | |
| *** tkay has quit IRC | 00:10 | |
| *** gyee has quit IRC | 00:12 | |
| *** dmorita has joined #openstack-swift | 00:29 | |
| *** kyles_ne has quit IRC | 00:41 | |
| *** kyles_ne has joined #openstack-swift | 00:41 | |
| *** oomichi has joined #openstack-swift | 00:42 | |
| *** kyles_ne has quit IRC | 00:46 | |
| *** shri has left #openstack-swift | 00:48 | |
| *** hhuang has joined #openstack-swift | 01:05 | |
| *** mkollaro has quit IRC | 01:18 | |
| *** NM has quit IRC | 01:38 | |
| *** oomichi has quit IRC | 01:56 | |
| *** addnull has joined #openstack-swift | 02:00 | |
| *** NM has joined #openstack-swift | 02:00 | |
| *** openstackgerrit has quit IRC | 02:01 | |
| *** haomai___ has quit IRC | 02:03 | |
| *** haomaiwang has joined #openstack-swift | 02:03 | |
| *** NM has quit IRC | 02:05 | |
| *** haomai___ has joined #openstack-swift | 02:06 | |
| *** haomaiwang has quit IRC | 02:08 | |
| *** haomai___ has quit IRC | 02:12 | |
| *** haomaiwang has joined #openstack-swift | 02:12 | |
| *** tsg has quit IRC | 02:13 | |
| *** haomaiw__ has joined #openstack-swift | 02:18 | |
| *** openstackgerrit has joined #openstack-swift | 02:22 | |
| *** haomaiwang has quit IRC | 02:22 | |
| *** haomaiw__ has quit IRC | 02:26 | |
| *** haomaiwang has joined #openstack-swift | 02:27 | |
| *** haomai___ has joined #openstack-swift | 02:33 | |
| openstackgerrit | Pete Zaitcev proposed a change to openstack/python-swiftclient: Stop creating extraneous directories https://review.openstack.org/122301 | 02:35 |
|---|---|---|
| *** NM has joined #openstack-swift | 02:35 | |
| *** haomaiwang has quit IRC | 02:36 | |
| *** NM has quit IRC | 02:45 | |
| *** NM has joined #openstack-swift | 02:58 | |
| *** kyles_ne has joined #openstack-swift | 02:59 | |
| *** addnull has quit IRC | 03:01 | |
| *** addnull has joined #openstack-swift | 03:05 | |
| *** tsg has joined #openstack-swift | 03:11 | |
| *** kyles_ne has quit IRC | 03:22 | |
| *** kyles_ne has joined #openstack-swift | 03:22 | |
| *** kyles_ne has quit IRC | 03:27 | |
| *** NM has quit IRC | 03:32 | |
| *** Kbee has joined #openstack-swift | 03:52 | |
| notmyname | python-swiftclient 2.3.0 released: https://pypi.python.org/pypi/python-swiftclient/2.3.0 | 03:53 |
| *** oomichi has joined #openstack-swift | 03:53 | |
| *** tsg has quit IRC | 04:00 | |
| mattoliverau | notmyname: yay! | 04:08 |
| openstackgerrit | Matthew Oliver proposed a change to openstack/swift: Add concurrent reads option to proxy https://review.openstack.org/117710 | 04:11 |
| notmyname | mattoliverau: how's performance compare on that patch? | 04:14 |
| *** tkay has joined #openstack-swift | 04:20 | |
| mattoliverau | notmyname: that's what I'm attempting to test now, but would like to test it on a multi-region cluster to better test latency, or at least one with multiple storage nodes :) (SAIO isn't very imformative) So building up a better test environment in the cloud. | 04:34 |
| notmyname | cool | 04:34 |
| *** zaitcev has quit IRC | 04:41 | |
| *** k4n0 has joined #openstack-swift | 04:46 | |
| *** bkopilov has joined #openstack-swift | 05:02 | |
| *** kopparam has joined #openstack-swift | 05:06 | |
| *** HenryG has quit IRC | 05:09 | |
| *** tkay has quit IRC | 05:12 | |
| *** HenryG has joined #openstack-swift | 05:21 | |
| notmyname | mattoliverau: I think you might need some updates in your abandon page | 05:25 |
| *** tkay has joined #openstack-swift | 05:25 | |
| notmyname | mattoliverau: eg https://review.openstack.org/#/c/62494 shows up but it had activity yesterday | 05:25 |
| mattoliverau | notmyname: really.. hmm, let me have a look | 05:26 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: Add unit test using real sockets https://review.openstack.org/97898 | 05:34 |
| notmyname | mattoliverau: FYI ^ on stuff like that where there was a small nit (eg pep8 line too long in this case) feel free to push a patch set over it if that's the only thing you see. | 05:35 |
| notmyname | I noticed it because of the abandoner, but the only -1 was because of that one line-too-long pep8 error | 05:36 |
| notmyname | mattoliverau: whoa the list just got a lot shorter. did you tickle something? | 05:37 |
| mattoliverau | notmyname: ahh right, yeah I've -1ed for that. Thanks, will do that in the future :) | 05:37 |
| mattoliverau | yup | 05:37 |
| mattoliverau | notmyname: fixed it. the html page had the wrong owership and the abandon user couldn't update the file... must've been me when I played with it yesterday to make sure it was working.. sorry about that. | 05:38 |
| notmyname | ah, cool | 05:38 |
| notmyname | how often will it check gerrit for activity? | 05:38 |
| notmyname | eg I just abandoned another one | 05:38 |
| mattoliverau | mental note, when I want to force a run use sudo -u abandon :P | 05:39 |
| mattoliverau | it's croned to run every 15 mins. | 05:39 |
| notmyname | mattoliverau: can you add a whitelist of stuff that won't get alerted on? | 05:39 |
| mattoliverau | I can speed it up now that it looks like its working | 05:39 |
| notmyname | meh whatever (to speeding up the cron) | 05:40 |
| mattoliverau | sure, whilte list on user? | 05:40 |
| notmyname | no, patches. https://review.openstack.org/#/c/92196 and https://review.openstack.org/#/c/77091 | 05:41 |
| notmyname | and https://review.openstack.org/#/c/101758/ and https://review.openstack.org/#/c/89250/ | 05:42 |
| *** mahatic has joined #openstack-swift | 05:42 | |
| mattoliverau | notmyname: sure, I can do it on change number, owner, email, username or match something in the subject. | 05:42 |
| notmyname | cool | 05:43 |
| notmyname | those 4 change numbers, please | 05:43 |
| notmyname | and that should clear out the abandon list (and prevent a couple more from getting there) | 05:43 |
| mattoliverau | sir, yes sir! | 05:43 |
| notmyname | lol | 05:43 |
| notmyname | mattoliverau: isn't that what you're supposed to say to mikal? | 05:43 |
| mattoliverau | I could white list the proposal bot | 05:43 |
| mattoliverau | notmyname: he would like that | 05:44 |
| notmyname | ya, whitelisting the proposal bot would probably be better | 05:44 |
| Kbee | notmyname: I had filed a bug, on which i'm working too. ->https://bugs.launchpad.net/swift/+bug/1370680, how does importance, and getting into release gets decided ? | 05:47 |
| notmyname | Kbee: since swift releases stable releases fairly often (we currently choose to release more often than the integrated release), there isn't a particular "schedule" to get on to get in a certain release | 05:49 |
| *** hhuang has quit IRC | 05:49 | |
| Kbee | notmyname: cool.. and setting bug importance.. does it happen in weekly meetings or just offline by core team.. Just understanding the process.. forgive my ignorance | 05:50 |
| notmyname | Kbee: as to priority, reviewers take a look at gerrit (there's a sample swift dashboard linked in the topic). but since there are so many patches to review, things deemed more important get put on the priority reviews page (again, see the topic for the link) | 05:51 |
| notmyname | Kbee: no no. great questions | 05:51 |
| notmyname | Kbee: if something isn't getting looked at (unfortunately it happens), then bringing it up in here in IRC is a good first step. if it's still not getting addressed, then bringing it up in a meeting is good | 05:52 |
| Kbee | notmyname: ok.. thanks.. | 05:52 |
| Kbee | this bug -> https://bugs.launchpad.net/swift/+bug/1370680 | 05:53 |
| notmyname | Kbee: quickly glancing at that bug, I'm not going to bump other things on the current priority review list for Juno since it's not something that has end-user impact. but it looks like a great improvement! thanks for working on it! | 05:53 |
| Kbee | notmyname: Sure.. Thanks.. | 05:54 |
| notmyname | Kbee: make sense? any other questions about that? | 05:55 |
| Kbee | notmyname: Yes it does.. no more.. thnx | 05:55 |
| *** tkay has quit IRC | 05:55 | |
| notmyname | Kbee: ok, cool. feel free to ask if you have other questions | 05:56 |
| notmyname | but I probably won't answer anything else tonight. it's getting late and I need to wind down for the day ;-) | 05:56 |
| Kbee | notmyname: Sure.. thnx.. | 05:56 |
| notmyname | mattoliverau: thanks again for the abandoner (the experiment continues!) | 05:57 |
| * notmyname signs out for the night | 05:57 | |
| mattoliverau | notmyname: my pleasure, sleep well | 05:58 |
| *** morganfainberg has quit IRC | 05:58 | |
| *** ttrumm has joined #openstack-swift | 06:08 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/swift: Imported Translations from Transifex https://review.openstack.org/122327 | 06:12 |
| *** morganfainberg has joined #openstack-swift | 06:13 | |
| openstackgerrit | Christian Schwede proposed a change to openstack/swift: Limit partition movement when adding a new tier https://review.openstack.org/121422 | 06:13 |
| *** ppai has joined #openstack-swift | 06:14 | |
| *** ttrumm_ has joined #openstack-swift | 06:19 | |
| *** sungju has joined #openstack-swift | 06:20 | |
| *** nshaikh has joined #openstack-swift | 06:21 | |
| *** ttrumm has quit IRC | 06:22 | |
| openstackgerrit | Christian Schwede proposed a change to openstack/swift: Limit partition movement when adding a new tier https://review.openstack.org/121422 | 06:32 |
| *** ttrumm_ has quit IRC | 06:32 | |
| *** ttrumm has joined #openstack-swift | 06:33 | |
| *** ttrumm has quit IRC | 06:51 | |
| *** ttrumm has joined #openstack-swift | 06:52 | |
| *** mikehn has quit IRC | 06:53 | |
| *** mikehn has joined #openstack-swift | 06:59 | |
| *** hhuang has joined #openstack-swift | 07:02 | |
| *** ttrumm has quit IRC | 07:06 | |
| *** kopparam has quit IRC | 07:07 | |
| *** LarsN has quit IRC | 07:14 | |
| *** LarsN has joined #openstack-swift | 07:14 | |
| *** echevemaster has quit IRC | 07:15 | |
| *** mahatic has quit IRC | 07:20 | |
| *** mahatic has joined #openstack-swift | 07:21 | |
| *** TaiSHi has quit IRC | 07:28 | |
| *** TaiSHi has joined #openstack-swift | 07:29 | |
| *** geaaru has joined #openstack-swift | 07:34 | |
| *** morganfainberg has quit IRC | 07:35 | |
| *** TaiSHi has quit IRC | 07:36 | |
| *** TaiSHi has joined #openstack-swift | 07:37 | |
| *** mikehn has quit IRC | 07:37 | |
| *** mikehn has joined #openstack-swift | 07:38 | |
| *** morganfainberg has joined #openstack-swift | 07:39 | |
| openstackgerrit | Christian Schwede proposed a change to openstack/swift: Limit partition movement when adding a new tier https://review.openstack.org/121422 | 07:43 |
| *** Kbee has quit IRC | 07:49 | |
| *** mikehn has quit IRC | 07:54 | |
| *** mikehn has joined #openstack-swift | 07:55 | |
| *** occup4nt has joined #openstack-swift | 08:03 | |
| *** occupant has quit IRC | 08:06 | |
| *** kopparam has joined #openstack-swift | 08:06 | |
| *** kopparam has quit IRC | 08:11 | |
| *** bkopilov has quit IRC | 08:12 | |
| *** mkollaro has joined #openstack-swift | 08:15 | |
| mattoliverau | I'm calling it a night too. Night all. | 08:16 |
| acoles | mattoliverau: g'night | 08:16 |
| *** Kbee has joined #openstack-swift | 08:21 | |
| *** kopparam has joined #openstack-swift | 08:37 | |
| *** k4n0 has quit IRC | 08:37 | |
| *** Kbee has quit IRC | 08:42 | |
| *** navid__ has joined #openstack-swift | 08:45 | |
| *** nshaikh has quit IRC | 08:46 | |
| *** aix has joined #openstack-swift | 08:51 | |
| *** k4n0 has joined #openstack-swift | 08:52 | |
| *** bkopilov has joined #openstack-swift | 08:54 | |
| *** bkopilov has quit IRC | 08:56 | |
| *** bkopilov has joined #openstack-swift | 08:56 | |
| *** Kbee has joined #openstack-swift | 08:56 | |
| *** sungju has quit IRC | 09:04 | |
| *** bkopilov has quit IRC | 09:07 | |
| *** bkopilov has joined #openstack-swift | 09:08 | |
| *** k4n0 has quit IRC | 09:22 | |
| *** ppai has quit IRC | 09:31 | |
| *** k4n0 has joined #openstack-swift | 09:35 | |
| *** ppai has joined #openstack-swift | 09:47 | |
| *** jamiehannaford has joined #openstack-swift | 10:13 | |
| *** hhuang has quit IRC | 10:13 | |
| *** ppai has quit IRC | 10:17 | |
| *** ppai has joined #openstack-swift | 10:30 | |
| *** mahatic has quit IRC | 10:30 | |
| *** oomichi has quit IRC | 10:37 | |
| *** addnull has quit IRC | 10:37 | |
| *** jamiehannaford has quit IRC | 10:44 | |
| *** mahatic has joined #openstack-swift | 10:47 | |
| *** Kbee has quit IRC | 10:49 | |
| *** aix has quit IRC | 11:08 | |
| *** ppai has quit IRC | 11:17 | |
| *** kopparam has quit IRC | 11:19 | |
| *** kopparam has joined #openstack-swift | 11:20 | |
| *** kopparam has quit IRC | 11:24 | |
| *** kopparam has joined #openstack-swift | 11:26 | |
| *** ppai has joined #openstack-swift | 11:33 | |
| *** NM has joined #openstack-swift | 11:35 | |
| *** aix has joined #openstack-swift | 11:38 | |
| *** jamiehannaford has joined #openstack-swift | 11:42 | |
| *** NM has quit IRC | 11:54 | |
| *** hhuang has joined #openstack-swift | 11:55 | |
| *** dmorita has quit IRC | 11:56 | |
| *** jamiehannaford has quit IRC | 12:00 | |
| *** k4n0 has quit IRC | 12:01 | |
| *** ppai has quit IRC | 12:14 | |
| *** viklund has joined #openstack-swift | 12:19 | |
| *** tsg has joined #openstack-swift | 12:19 | |
| viklund | is it possible to move a swift container from one tenant to another? | 12:20 |
| *** kopparam has quit IRC | 12:28 | |
| *** ppai has joined #openstack-swift | 12:28 | |
| *** sandywalsh has joined #openstack-swift | 12:39 | |
| *** miqui has joined #openstack-swift | 12:41 | |
| *** bkopilov has quit IRC | 12:45 | |
| *** bkopilov has joined #openstack-swift | 12:46 | |
| *** bkopilov has quit IRC | 12:46 | |
| *** annegentle has joined #openstack-swift | 12:54 | |
| *** ppai has quit IRC | 12:58 | |
| goodes | you have to copy it | 13:03 |
| goodes | viklund: there is no such concept as 'move' in swift | 13:03 |
| viklund | ok | 13:04 |
| viklund | so that's supported by some api or do I have to do it with a script and iterate over the objects? | 13:05 |
| goodes | viklund: there is an offload copy command (i.e. the copy will happen internally within swift rather) however I believe that it must be within the same account as the syntax is X-Copy-From: /container/object | 13:06 |
| goodes | which makes sense from a security point of view | 13:07 |
| viklund | goodes: thanks | 13:07 |
| viklund | I've been looking at the apis to see what is possible, it's not always so easy | 13:07 |
| viklund | to understand | 13:07 |
| goodes | sadly true | 13:07 |
| goodes | the swift client allows you to download all the files in a container at once and then upload them all at the same time | 13:08 |
| viklund | probably not a good solution for large files | 13:09 |
| goodes | the only other efficient way would be (if security is not too much of a concern) to write some proxy middleware to do it | 13:09 |
| goodes | it would save on the round trip from proxy -> swift client -> proxy | 13:09 |
| viklund | security is THE concern here | 13:10 |
| goodes | so it would not make sense for it to happen with swift | 13:11 |
| viklund | yes | 13:11 |
| goodes | either you would need to authenticate for source and target access which is not currently supported as current logic only supports a single authentication context | 13:12 |
| viklund | we will probably make sure that we don't have to move data between containers instead | 13:12 |
| goodes | again you could write some middleware that sat infront if authenentication middleware, accepted two tokens, performed an authenticated read and then an authenticated write | 13:13 |
| goodes | it think that it would be a few milliseconds slower then doing it within the account | 13:13 |
| goodes | as the cost would just be the additional authentication and the (miminal) cost of the data moving through the middleware | 13:14 |
| *** annegentle has quit IRC | 13:14 | |
| *** ppai has joined #openstack-swift | 13:14 | |
| viklund | thanks again goodes, I will look into that as well | 13:15 |
| *** NM has joined #openstack-swift | 13:18 | |
| *** tdasilva has joined #openstack-swift | 13:19 | |
| *** jamiehannaford has joined #openstack-swift | 13:20 | |
| *** jamiehannaford has quit IRC | 13:21 | |
| acoles | viklund: goodes: fyi, there is a 'copy between accounts' feature that has just merged in upstream swift https://review.openstack.org/#/c/72157/ (i have not studied it so can't offer much more than the link) | 13:26 |
| acoles | just object copy by looks of it | 13:27 |
| *** vr2 has joined #openstack-swift | 13:27 | |
| *** navid__ has quit IRC | 13:30 | |
| *** nshaikh has joined #openstack-swift | 13:31 | |
| *** dmsimard_away is now known as dmsimard | 13:32 | |
| tdasilva | acoles, viklund, goodes : I tried the copy between accounts recently as part of the migration patch and it worked well, you just need to provide the 'Destination-Account' header as part of the COPY request | 13:34 |
| *** Kbee has joined #openstack-swift | 13:34 | |
| tdasilva | i don't think it does any special handling of auth tokens, so it just expects that the token being used has access to both accounts | 13:35 |
| dmsimard | Hey guys. I feel like I'm missing something basic here.. but how do I compute the total size in used by each tenant ? Using the swiftclient library, it doesn't look like get_account or head_account provides all the space in use. I tried iterating over the containers provided by get_account but I end up with the same size. | 13:37 |
| *** annegentle has joined #openstack-swift | 13:37 | |
| dmsimard | For instance, there is a tenant that comes up as bytes used 0 on get_account (or "swift stat") but he actually has stuff. | 13:38 |
| dmsimard | Swift is backed by keystone, I'm doing this by authenticating as admin and retrieving the stats from inside the tenant. | 13:41 |
| openstackgerrit | Lorcan Browne proposed a change to openstack/swift: Add "--no-overlap" option to swift-dispersion populate https://review.openstack.org/118411 | 13:41 |
| *** Manifesto has joined #openstack-swift | 13:42 | |
| vr2 | Manifesto hello | 13:48 |
| Manifesto | hello | 13:51 |
| Manifesto | I tried to use the SAIO guideline but when running swift-init I got this error : | 13:52 |
| Manifesto | ImportError: No module named middleware.recon | 13:53 |
| Manifesto | and | 13:53 |
| Manifesto | ImportError: No module named catch_errors | 13:53 |
| Manifesto | any idea ? | 13:53 |
| openstackgerrit | Alistair Coles proposed a change to openstack/swift: Early checks for data migration config and headers https://review.openstack.org/121831 | 13:54 |
| goodes | tdasilva: is there a separate channel for ec work | 13:54 |
| tdasilva | goodes: i don't think so | 13:55 |
| tdasilva | goodes: I know there is a trello board for keeping a list of tasks | 13:55 |
| tdasilva | but peluse_, tsg would know more for sure | 13:55 |
| tsg | goodes: #link https://trello.com/b/LlvIFIQs/swift-erasure-codes | 13:57 |
| goodes | tsg: I was thinking about doing the SAIO ticket | 13:57 |
| tsg | it is up-to-date, at least as of last week | 13:57 |
| tsg | goodes: sure, you can start looking at it .. just that it may be a bit early for that at the moment | 13:58 |
| goodes | however, I tried downloading the EC branch and had isssues with it parsing the EC info in the conf file and was wondering if this is the forum or is there a dedicate EC forum, mail | 13:58 |
| tsg | goodes: EC stuff is still heavily under development | 13:59 |
| tsg | only 2 base patches have merged | 13:59 |
| tsg | and 3 more to go before you can try anything useful | 13:59 |
| goodes | so better topull from gerrit? | 13:59 |
| tsg | goodes: if you want to play around with the current state of the tree, yes | 14:00 |
| goodes | tdasilva: it seems to me that 'Destination-Account' is a security issue as it is only inspected after auth | 14:00 |
| goodes | I could go an wipe out all your files with an empty file (assuming that I know the container/object_name of all your objects) | 14:01 |
| tdasilva | goodes: how so? | 14:01 |
| goodes | tdasilva: am I correct in assuming that once you get into proxy/server.py there are no more access checks performed? | 14:03 |
| tdasilva | goodes: but only if you had access to both accounts | 14:03 |
| tdasilva | goodes: I remember this same question being raised during the review process and i'm looking for the answer here: https://review.openstack.org/#/c/72157 | 14:04 |
| goodes | so who checks that I have access to the 'Destination-Account'? | 14:04 |
| tdasilva | the PUT method as far as I understand in the proxy/controller/obj.py | 14:07 |
| tdasilva | it calls authorize() | 14:08 |
| tdasilva | and again, when doing the GET for the source | 14:08 |
| tdasilva | goodes: does that make sense? basically when you send a COPY request, it becomes a PUT request, the PUT method will call authorize(). Also in the PUT request it will look at the X-Copy-From headers and issue a GET request, which will again call authorize() | 14:11 |
| tdasilva | goodes: that's my understanding, just not 100% sure how @delay_denial factors into all this. acoles, chmouel, portante might be able to provide more details | 14:13 |
| goodes | makes sense | 14:16 |
| goodes | just looking at the flow now | 14:16 |
| acoles | tdasilva: I think that is correct | 14:16 |
| acoles | the call to authorize() is 'delayed' (callback form the controllers) so that the controller can fetch any ACLs that may be on a container and put in in the env for keystoneauth (or other auth) to then use | 14:17 |
| *** jseiler has joined #openstack-swift | 14:17 | |
| *** jseiler has left #openstack-swift | 14:18 | |
| goodes | its now clear to me when its doing | 14:21 |
| Manifesto | no idea on this ImportError: No module named catch_errors ? | 14:23 |
| tdasilva | acoles: thanks! i see how it gets delayed here: https://github.com/openstack/swift/blob/master/swift/proxy/server.py#L367 | 14:28 |
| *** tongli has joined #openstack-swift | 14:28 | |
| *** tdasilva has quit IRC | 14:35 | |
| *** elambert has quit IRC | 14:46 | |
| *** Trixboxer has joined #openstack-swift | 14:48 | |
| *** nshaikh has quit IRC | 15:02 | |
| *** Kbee has quit IRC | 15:02 | |
| *** Kbee has joined #openstack-swift | 15:08 | |
| *** ppai has quit IRC | 15:09 | |
| *** kevinc has joined #openstack-swift | 15:11 | |
| *** nshaikh has joined #openstack-swift | 15:16 | |
| openstackgerrit | Lorcan Browne proposed a change to openstack/swift: Add "--no-overlap" option to swift-dispersion populate https://review.openstack.org/118411 | 15:17 |
| *** bill_az has joined #openstack-swift | 15:25 | |
| *** Kbee has quit IRC | 15:32 | |
| *** openstackgerrit has quit IRC | 15:33 | |
| *** annegentle has quit IRC | 15:39 | |
| *** annegentle has joined #openstack-swift | 15:39 | |
| dmsimard | Trying my question again - through the swiftclient library - how would I calculate the total size of an account/tenant ? | 15:40 |
| *** annegentle has quit IRC | 15:44 | |
| *** vr2 has quit IRC | 15:45 | |
| *** vr2 has joined #openstack-swift | 15:45 | |
| dmsimard | Nevermind, had an issue - fixed now. | 15:46 |
| *** vr2 has quit IRC | 15:48 | |
| *** vr3 has joined #openstack-swift | 15:48 | |
| *** vr3 has quit IRC | 15:57 | |
| *** vr2 has joined #openstack-swift | 15:57 | |
| *** openstackgerrit has joined #openstack-swift | 16:03 | |
| notmyname | good morning | 16:06 |
| notmyname | dmsimard: you got it? (the answer is to look at the stat response (which does a HEAD to swift) | 16:06 |
| *** annegentle has joined #openstack-swift | 16:07 | |
| *** nshaikh has quit IRC | 16:09 | |
| *** tsg has quit IRC | 16:10 | |
| mahatic | notmyname, good morning. The opw is on https://wiki.gnome.org/OutreachProgramForWomen/2014/DecemberMarch/#Participating_Organizations | 16:12 |
| mahatic | notmyname, is Swift participating? | 16:12 |
| *** kyles_ne has joined #openstack-swift | 16:13 | |
| openstackgerrit | A change was merged to openstack/swift: Move multipart MIME parser into utils https://review.openstack.org/120652 | 16:19 |
| *** mwstorer has joined #openstack-swift | 16:22 | |
| *** annegentle has quit IRC | 16:24 | |
| *** annegentle has joined #openstack-swift | 16:24 | |
| *** vr2 has quit IRC | 16:27 | |
| *** ZBhatti has joined #openstack-swift | 16:27 | |
| notmyname | chmouel: now that the swift bind_port is explicit, can we get swift in devstack turned on by default? | 16:33 |
| chmouel | notmyname: humm perhaps, | 16:34 |
| chmouel | notmyname: but does that really matter to be by default? | 16:34 |
| notmyname | mahatic: I talked to anteaya about it recently. I think you've been talking to her, right? I offered to work with someone with it | 16:34 |
| notmyname | chmouel: well, that was one objection I heard. something like "we don't want to use non-default ports in devstack" (which I think is kinda silly) | 16:35 |
| notmyname | chmouel: the other option is to wait until swift changes the recommended port. (I saw something in the 62xx range proposed.) but that can't happen until after the juno release | 16:36 |
| mahatic | notmyname, oh okay. Nope I haven't actually spoken to her in a while. Last was before i posted to the openstack mailing list. | 16:36 |
| *** Manifesto has quit IRC | 16:36 | |
| chmouel | notmyname: well so the story back in the day we had it enabled by default and it was breaking those SSH X forwarding for the lnux desktop users, | 16:36 |
| notmyname | chmouel: right because it conflicts with X. also somethign about SELinux policies | 16:37 |
| mahatic | notmyname, Are you talking about the mail you sent on the mailing list that you or someone else would work? Or has any other conversation/update happened lately? | 16:37 |
| chmouel | notmyname: so maybe we should wait for the 62xx bind port by default get merged and try it | 16:37 |
| notmyname | chmouel: and when I recommended "just change the bind_port" that went over like kicking kittens | 16:37 |
| notmyname | chmouel: but the point is that there isn't and won't be a default any more | 16:37 |
| *** Kbee has joined #openstack-swift | 16:38 | |
| notmyname | mahatic: ya, probably that mail. well, I also talked to anteaya privately about it (she convinced me to reply ;-). I haven't seen anything more recent | 16:38 |
| dmsimard | notmyname: Yeah.. for some reason I had timeout issues within the clusters. I fixed it and got the total size as expected. I would get zero - perhaps due to the timeout. | 16:38 |
| notmyname | chmouel: we'll change what we recommend, of course. but technically it's not a default | 16:38 |
| notmyname | dmsimard: ok. the account stats are updated asynchronously via the container-updater. so if that is having issues, then you won't get anything in the account status | 16:39 |
| anteaya | mahatic: I suggest that you not worry too much about opw for now | 16:39 |
| notmyname | there. talk to anteaya :-) | 16:39 |
| chmouel | notmyname: thinking from the end user point of view, if the guy start to spin up just devstack and have all those nova stuff + swift enabled by default would be pretty heavy on its vm | 16:39 |
| anteaya | mahatic: spend time in the channel and learn how to make a contribution | 16:39 |
| dmsimard | notmyname: Makes sense. Does /healthcheck take into account the status of container-updater ? | 16:40 |
| mahatic | notmyname, :) anteaya, alright | 16:40 |
| notmyname | chmouel: sure sure. that is one problem with devstack anyway. of course, I think I've seen -infra conversation about trying to solve that somehow | 16:40 |
| chmouel | notmyname: well i think the question should be what the default in devstack should be for a guy trying that openstack thing for the first time, an object storage or a compute api system | 16:41 |
| notmyname | dmsimard: no. /healthcheck basically hits the proxy and returns "OK". if you don't get "OK" then you can't get to the proxy server. mostly it's used for loadbalancer backpressure | 16:41 |
| notmyname | chmouel: devstack should never be for "a guy trying that openstack thing". kinda how the SAIO isn't good for people just wanting to try out swift. it's a dev environment | 16:42 |
| *** annegentle has quit IRC | 16:43 | |
| dmsimard | notmyname: There is a /healthcheck endpoint for proxy-server but also container-server, object-server and account-server, though. | 16:44 |
| chmouel | notmyname: yep agree but even so if it's a dev environement having too much by default started would be bad for the experience of that (beginner) dev, i think we may need to make it easy to switch on/off features to make it as we want | 16:46 |
| notmyname | dmsimard: well, ok :-) | 16:46 |
| notmyname | dmsimard: healthcheck is a wsgi middleware you can put in any pipeline, so yeah you can put it in front of the storage servers too :-) | 16:46 |
| dmsimard | Right. | 16:46 |
| goodes | does anyone know, is list_endpoints just for sahara? | 16:47 |
| notmyname | chmouel: yeah. I agree. but I also don't like the common openstack perception of "use devstack. oh swift is different". I want to see swift available and considered by all openstack devs | 16:47 |
| notmyname | goodes: "just". my favorite word :-) | 16:47 |
| goodes | was that was 'drove | 16:48 |
| goodes | list_endpoints? | 16:48 |
| goodes | are they using it to work out which region/zone to run the job or is it deeper, i.e have the hadoop nodes direct access the object servers to reduce latency? | 16:50 |
| chmouel | notmyname: yeah right, i think it's just the default tailored for nova, all others integrated project are like swift they need to be explicitely enabled in devstack (i.e: heat, ceilo, trove and so on) | 16:50 |
| chmouel | unfortunately i think people still assume cloud == compute :( | 16:51 |
| chmouel | or iphone storage fwiw | 16:51 |
| notmyname | chmouel: which should be swift, of course | 16:51 |
| notmyname | goodes: I think those are the same thing. or one is a more specific example of the other. in both cases, "yes". it's to ensure that tehy can move compute close to the data | 16:51 |
| goodes | I am trying to understand if sahara/hadoop will thus bypass the proxy server (and all middleware, stats etc.) | 16:53 |
| *** gyee has joined #openstack-swift | 16:53 | |
| openstackgerrit | Keshava Bharadwaj proposed a change to openstack/swift: Provides proper error handling on builder unpickle https://review.openstack.org/122225 | 16:54 |
| notmyname | goodes: yes. that would be the only useful reason to use list_endpoints. if the thing requesting info from list_endpoints doesn't have access to the actual storage nodes that are returned, then there is no point to getting the info | 16:55 |
| goodes | unless you wanted to know which room or rack to run it in | 16:56 |
| goodes | which you could also infer from the info without using a direct client | 16:56 |
| goodes | just seems (to borrow someone elses term) a horrible kludge | 16:56 |
| openstackgerrit | Lorcan Browne proposed a change to openstack/swift: Add new features to swift-drive-audit https://review.openstack.org/122468 | 17:01 |
| notmyname | goodes: if all you care about is getting close enough, but not running on the storage nodes, then you can use something else and don't need list_endpoints | 17:05 |
| *** elambert has joined #openstack-swift | 17:05 | |
| *** tsg has joined #openstack-swift | 17:06 | |
| *** tkay has joined #openstack-swift | 17:07 | |
| goodes | trying to understand their thinking. Is this a short term kludge and they will act in more classic way | 17:10 |
| goodes | in the future, or is this seriously how they intend to continue | 17:10 |
| notmyname | "they" == sahara? I have no diea | 17:11 |
| smart_developer | Has anyone experienced the uncontrollable, swift-object auditor creation issue when using Swift ? | 17:12 |
| *** shri has joined #openstack-swift | 17:12 | |
| *** tdasilva has joined #openstack-swift | 17:21 | |
| smart_developer | Basically the auditor process forks more auditor processes. | 17:21 |
| smart_developer | Until all the system resources are used up, and can't spawn any more. | 17:21 |
| *** tongli has quit IRC | 17:22 | |
| *** geaaru has quit IRC | 17:25 | |
| *** vu has joined #openstack-swift | 17:31 | |
| *** openstackgerrit has quit IRC | 17:31 | |
| *** openstackgerrit has joined #openstack-swift | 17:33 | |
| *** vu has quit IRC | 17:33 | |
| *** vu has joined #openstack-swift | 17:43 | |
| smart_developer | I think there might be a bug in auditor.py for .../swift/obj/ | 17:43 |
| *** annegentle has joined #openstack-swift | 17:44 | |
| smart_developer | and it keeps looping where child processes get forked. | 17:44 |
| smart_developer | in that section. | 17:44 |
| notmyname | openstack board of directors meeting is in about an hour. the defcore recommendation which excludes swift from the openstack trademark will be discussed. I'll report back with a summary | 17:48 |
| *** annegentle has quit IRC | 17:48 | |
| *** Kbee has quit IRC | 17:52 | |
| *** marcusvrn has joined #openstack-swift | 17:57 | |
| *** NM has quit IRC | 17:59 | |
| *** nshaikh has joined #openstack-swift | 17:59 | |
| *** kevinc has quit IRC | 18:00 | |
| *** annegentle has joined #openstack-swift | 18:02 | |
| *** kevinc has joined #openstack-swift | 18:03 | |
| *** tsg has quit IRC | 18:14 | |
| openstackgerrit | Alan Erwin proposed a change to openstack/swift: Adding verify_object_partition https://review.openstack.org/122489 | 18:17 |
| *** annegentle has quit IRC | 18:21 | |
| *** annegentle has joined #openstack-swift | 18:22 | |
| *** tsg has joined #openstack-swift | 18:24 | |
| smart_developer | In the configuration files, why is middleware also known as a "filter"? | 18:25 |
| *** rmcall has joined #openstack-swift | 18:27 | |
| *** mahatic has quit IRC | 18:32 | |
| *** kevinc has quit IRC | 18:38 | |
| *** occup4nt is now known as occupant | 18:42 | |
| openstackgerrit | Alan Erwin proposed a change to openstack/swift: Adding object partition check https://review.openstack.org/122194 | 18:44 |
| *** aerwin has joined #openstack-swift | 18:47 | |
| *** mahatic has joined #openstack-swift | 18:50 | |
| openstackgerrit | A change was merged to openstack/python-swiftclient: Stop creating extraneous directories https://review.openstack.org/122301 | 18:55 |
| *** annegentle_ has joined #openstack-swift | 18:56 | |
| *** annegentle_ is now known as Guest62135 | 18:57 | |
| *** rmcall has quit IRC | 19:02 | |
| *** rmcall has joined #openstack-swift | 19:04 | |
| *** aix has quit IRC | 19:05 | |
| *** ZBhatti has quit IRC | 19:07 | |
| *** zaitcev has joined #openstack-swift | 19:13 | |
| *** ChanServ sets mode: +v zaitcev | 19:13 | |
| *** kevinc_ has joined #openstack-swift | 19:20 | |
| *** vu has quit IRC | 19:21 | |
| openstackgerrit | Michael Barton proposed a change to openstack/swift: container.merge_items bug https://review.openstack.org/122503 | 19:27 |
| *** annegentle has quit IRC | 19:33 | |
| *** vu has joined #openstack-swift | 19:33 | |
| smart_developer | How do you increase the # of processes allowed on your system by ulimit ? | 19:34 |
| *** ZBhatti has joined #openstack-swift | 19:36 | |
| *** nshaikh has left #openstack-swift | 19:36 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: container.merge_items bug https://review.openstack.org/122503 | 19:38 |
| *** annegentle has joined #openstack-swift | 19:38 | |
| *** openstackgerrit has quit IRC | 19:45 | |
| *** rmcall has quit IRC | 19:53 | |
| *** tdasilva has quit IRC | 19:54 | |
| chmouel | i like monty idea here: http://inaugust.com/post/108 "Rather than tagging things negatively like "beta" or "experimental", positively tag things as "Production Ready"." "today, I'd say we should start with Swift and Layer #1 and that's it." | 19:59 |
| *** annegent_ has joined #openstack-swift | 20:01 | |
| *** annegent_ has quit IRC | 20:02 | |
| *** annegentle has quit IRC | 20:05 | |
| *** rmcall has joined #openstack-swift | 20:20 | |
| *** serverascode has quit IRC | 20:23 | |
| *** tsg has quit IRC | 20:24 | |
| *** serverascode has joined #openstack-swift | 20:26 | |
| *** jergerber has joined #openstack-swift | 20:28 | |
| *** fifieldt__ has joined #openstack-swift | 20:32 | |
| *** hhuang has quit IRC | 20:35 | |
| *** fifieldt_ has quit IRC | 20:35 | |
| *** ZBhatti has quit IRC | 20:37 | |
| *** ZBhatti has joined #openstack-swift | 20:41 | |
| *** hhuang has joined #openstack-swift | 20:48 | |
| *** vu has quit IRC | 21:02 | |
| *** shri has quit IRC | 21:03 | |
| notmyname | board of directors meeting summary (on the defcore proposal): no consensus around the proposal to exclude swift and keystone, the defcore committee needs to wrk with the foundation to look at it from a perspective of multiple trademarks instead of just one "openstack" one | 21:07 |
| *** shri has joined #openstack-swift | 21:08 | |
| *** shri has quit IRC | 21:08 | |
| *** openstackgerrit has joined #openstack-swift | 21:09 | |
| notmyname | although not yet defined, it could eg be something like "openstack" (all of integrated) + "made with openstack parts" (some subset). or "openstack api" + "openstack code" or "openstack compute" + "openstack storage" + etc | 21:09 |
| notmyname | but the point is that there isn't (currently) something called "openstack" that is some subset of the integrated release, and there is guidance to ensure that doesn't happen | 21:10 |
| mattoliverau | notmyname: thanks for the update.. So the defcore saga continues and looks like it will for a long while yet. | 21:14 |
| *** annegentle has joined #openstack-swift | 21:15 | |
| *** rmcall has quit IRC | 21:26 | |
| *** annegentle has quit IRC | 21:35 | |
| *** tsg has joined #openstack-swift | 21:40 | |
| *** tgohad has joined #openstack-swift | 21:43 | |
| *** tsg has quit IRC | 21:46 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Merge master to feature/ec https://review.openstack.org/122541 | 22:02 |
| *** dmsimard is now known as dmsimard_away | 22:32 | |
| *** ZBhatti has quit IRC | 22:41 | |
| notmyname | redbo: do we need to get a unicode patch for containers into master? | 22:43 |
| notmyname | dfg: ^ | 22:43 |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Reject overly-taxing ranged-GET requests https://review.openstack.org/117579 | 22:44 |
| *** jergerber has quit IRC | 22:54 | |
| redbo | the patch I submitted? it needs to be fixed | 23:00 |
| notmyname | redbo: ah, now I see it. https://review.openstack.org/#/c/122503/ right? I had seen it land in the rackspace production branch | 23:01 |
| *** aerwin has quit IRC | 23:01 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Zero-copy object-server GET responses with splice() https://review.openstack.org/102609 | 23:02 |
| redbo | yes. I can't promise my solution is the best possible, but it's at least better :) | 23:06 |
| *** Guest62135 is now known as annegentle | 23:13 | |
| *** annegentle is now known as annegentleznc | 23:13 | |
| *** annegentleznc is now known as annegentle | 23:14 | |
| *** mkollaro has quit IRC | 23:38 | |
| *** mkollaro has joined #openstack-swift | 23:41 | |
| *** kevinc_ has quit IRC | 23:42 | |
| *** elambert has quit IRC | 23:47 | |
| *** kyles_ne has quit IRC | 23:58 | |
| *** kyles_ne has joined #openstack-swift | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!