| ho | Morning! | 00:02 |
|---|---|---|
| mattoliverau | ho: morning | 00:04 |
| notmyname | hi | 00:06 |
| ho | mattoliverau: morning! | 00:09 |
| torgomatic | alright, plunger time | 00:10 |
| torgomatic | wait | 00:10 |
| torgomatic | notmyname: that one hit the gate queue, so I'm gonna recheck the world now | 00:10 |
| notmyname | torgomatic: thanks | 00:11 |
| torgomatic | alright, gerrit spam bomb inbound ;) | 00:11 |
| *** kota_ has joined #openstack-swift | 00:11 | |
| *** ChanServ sets mode: +v kota_ | 00:11 | |
| ho | torgomatic: lol | 00:14 |
| *** remix_tj has quit IRC | 00:15 | |
| notmyname | I have 24 new emails! | 00:15 |
| kota_ | the infra issue fixed? | 00:16 |
| notmyname | kota_: yup | 00:16 |
| kota_ | notmyname: nice! | 00:16 |
| *** remix_tj has joined #openstack-swift | 00:16 | |
| *** chlong-zzz has joined #openstack-swift | 00:17 | |
| *** evanjfraser has quit IRC | 00:19 | |
| *** evanjfraser has joined #openstack-swift | 00:21 | |
| mattoliverau | notmyname: 24, your popular! | 00:22 |
| *** marcusvrn has quit IRC | 00:24 | |
| kota_ | ho: I wrote up Japanese subscription for SwiftStack Erasure Code Video, here. | 00:27 |
| kota_ | ho: if you find my bad for the translation, could you please let me know that? | 00:28 |
| kota_ | oh, I missed link, here, https://youtu.be/I0HytEfQzjw | 00:28 |
| notmyname | that's really cool!@ | 00:35 |
| kota_ | notmyname: Thanks.I hope it would be helpful for Japanese to educate Swift more ;) | 00:36 |
| *** asettle has joined #openstack-swift | 00:36 | |
| *** dmorita has joined #openstack-swift | 00:38 | |
| ho | kota_: cool! now I stay in yokohama for business trip so maybe tomorrow I will see it. | 00:47 |
| kota_ | ho: that's not urgent, please take your time :) | 00:48 |
| *** dstanek has joined #openstack-swift | 01:03 | |
| *** blmartin has joined #openstack-swift | 01:07 | |
| *** kota_ has quit IRC | 01:12 | |
| *** blmartin_ has joined #openstack-swift | 01:14 | |
| *** blmartin has quit IRC | 01:17 | |
| *** kota_ has joined #openstack-swift | 01:19 | |
| *** ChanServ sets mode: +v kota_ | 01:19 | |
| *** haigang has joined #openstack-swift | 01:37 | |
| *** haigang has quit IRC | 01:47 | |
| ho | kota_: thanks! :-) | 01:47 |
| *** haigang has joined #openstack-swift | 01:48 | |
| *** jrichli has joined #openstack-swift | 01:58 | |
| *** kota_ has quit IRC | 02:02 | |
| *** asettle has quit IRC | 02:04 | |
| *** janonymous_ has quit IRC | 02:10 | |
| *** dmorita has quit IRC | 02:14 | |
| *** haigang has quit IRC | 02:31 | |
| *** pagit has joined #openstack-swift | 02:32 | |
| pagit | hello.Today the first to know swift, it is written by python. i have to learn python | 02:35 |
| pagit | . | 02:35 |
| *** gyee is now known as operator99 | 02:42 | |
| *** asettle has joined #openstack-swift | 02:48 | |
| *** asettle has quit IRC | 02:49 | |
| *** pagit has left #openstack-swift | 03:07 | |
| *** km has joined #openstack-swift | 03:10 | |
| *** jrichli has quit IRC | 03:30 | |
| openstackgerrit | Michael Barton proposed openstack/swift: go: fix output of /recon/mounted https://review.openstack.org/190012 | 03:51 |
| *** km has quit IRC | 04:01 | |
| *** km has joined #openstack-swift | 04:02 | |
| *** blmartin_ has quit IRC | 04:04 | |
| openstackgerrit | Merged openstack/swift-specs: Updates to encryption spec https://review.openstack.org/154318 | 04:07 |
| openstackgerrit | Merged openstack/swift: Object are sorted and concatenated by swift in cardinal order. https://review.openstack.org/189015 | 04:08 |
| openstackgerrit | Merged openstack/swift: The hash_cleanup_listdir function should only be called when necessary. https://review.openstack.org/178317 | 04:15 |
| openstackgerrit | Merged openstack/swift: Change usage help and Attention messages to warnings https://review.openstack.org/187974 | 04:15 |
| openstackgerrit | Merged openstack/swift: Add metadata size tests on the border https://review.openstack.org/187489 | 04:15 |
| *** ppai has joined #openstack-swift | 04:17 | |
| *** km has quit IRC | 04:31 | |
| *** chlong-zzz is now known as chlong | 04:32 | |
| *** zaitcev has quit IRC | 05:08 | |
| *** silor has joined #openstack-swift | 05:50 | |
| *** haomaiwang has quit IRC | 05:54 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:29 | |
| *** mmcardle has joined #openstack-swift | 06:51 | |
| *** zhill has joined #openstack-swift | 07:07 | |
| *** zhill has quit IRC | 07:23 | |
| *** geaaru has joined #openstack-swift | 07:28 | |
| *** csmart has quit IRC | 07:32 | |
| *** jistr has joined #openstack-swift | 07:37 | |
| *** csmart has joined #openstack-swift | 07:39 | |
| *** ronenkat has joined #openstack-swift | 07:41 | |
| *** csmart has quit IRC | 07:43 | |
| *** csmart has joined #openstack-swift | 07:46 | |
| *** chlong has quit IRC | 07:51 | |
| *** haomaiwang has joined #openstack-swift | 07:56 | |
| *** jordanP has joined #openstack-swift | 07:59 | |
| *** leopoldj has joined #openstack-swift | 08:00 | |
| *** acoles_away is now known as acoles | 08:05 | |
| acoles | mattoliverau: thank you for landing that encryption spec update :D | 08:07 |
| mattoliverau | acoles: nps, I'm becoming a spec nazi, spec rule #1, land quickly :) | 08:08 |
| mattoliverau | acoles: and good morning! | 08:08 |
| acoles | mattoliverau: oh yeah, good morning! :P | 08:08 |
| acoles | someone unblocked that gate i see. whoosh. | 08:09 |
| mattoliverau | acoles: yup, turns out having the EC dependancies installed on the gate test nodes is important for swift patches.. who knew :P | 08:11 |
| acoles | heh | 08:11 |
| *** joeljwright has joined #openstack-swift | 08:15 | |
| openstackgerrit | Joel Wright proposed openstack/python-swiftclient: Log and report trace on service operation fails https://review.openstack.org/171692 | 08:19 |
| openstackgerrit | Matthew Oliver proposed openstack/swift-specs: Add Spec Lifecycle Rules to readme https://review.openstack.org/190066 | 08:22 |
| mattoliverau | acoles: spec rules I've written while wating in a car.. long story :P | 08:23 |
| mattoliverau | ^^ | 08:23 |
| mattoliverau | Better section name might be required, but hopefully this patch will start discussions. | 08:23 |
| mattoliverau | anyway, time to go, have a great night all. | 08:24 |
| *** haypo has joined #openstack-swift | 08:56 | |
| haypo | cschwede, morning. the swift CI was fixed yesterday. tests now pass on my six dependency patch https://review.openstack.org/#/c/189495/ and the depends-on trick works for my other py3 patches, cool | 08:56 |
| cschwede | haypo: :) we’re going to talk about the py3 patches this evening in the team meeting (https://wiki.openstack.org/wiki/Meetings/Swift) - if you have time | 08:59 |
| haypo | cschwede, i may join the meeting, but don't wait for me | 09:00 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Get StringIO and cStringIO from six.moves https://review.openstack.org/185457 | 09:07 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO in ssync https://review.openstack.org/186073 | 09:07 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO for file https://review.openstack.org/186072 | 09:07 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO for WSGI input https://review.openstack.org/186071 | 09:07 |
| janonymous | i am new in openstack-community, how can i join the meeting. | 09:07 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace StringIO with BytesIO in ssync https://review.openstack.org/186073 | 09:09 |
| haypo | janonymous, join #openstack-meeting at 21:00 UTC (tonight) | 09:10 |
| janonymous | Thanks. | 09:12 |
| acoles | janonymous: also see https://wiki.openstack.org/wiki/Meetings/Swift | 09:24 |
| *** aix has joined #openstack-swift | 09:34 | |
| *** ho has quit IRC | 09:41 | |
| *** fesp has joined #openstack-swift | 09:49 | |
| *** proteusguy has quit IRC | 09:52 | |
| *** proteusguy has joined #openstack-swift | 09:54 | |
| *** fesp has quit IRC | 09:55 | |
| janonymous | Thanks just watched the issues listed. thanks :) | 09:58 |
| *** fanyaohong has joined #openstack-swift | 10:01 | |
| *** kei_yama has quit IRC | 10:07 | |
| *** acoles is now known as acoles_away | 10:23 | |
| *** acoles_away is now known as acoles | 11:24 | |
| *** proteusguy has quit IRC | 11:30 | |
| *** proteusguy has joined #openstack-swift | 11:43 | |
| *** marcusvrn has joined #openstack-swift | 11:51 | |
| *** joeljwright has quit IRC | 11:55 | |
| *** joeljwright has joined #openstack-swift | 11:56 | |
| *** ppai has quit IRC | 12:06 | |
| *** mmcardle has quit IRC | 12:11 | |
| *** aix has quit IRC | 12:14 | |
| *** fanyaohong has quit IRC | 12:15 | |
| *** joeljwright has quit IRC | 12:18 | |
| *** ppai has joined #openstack-swift | 12:20 | |
| *** chlong has joined #openstack-swift | 12:29 | |
| *** aix has joined #openstack-swift | 12:50 | |
| *** ppai has quit IRC | 12:54 | |
| *** mmcardle has joined #openstack-swift | 12:54 | |
| *** thurloat_isgone is now known as thurloat | 12:59 | |
| openstackgerrit | Merged openstack/python-swiftclient: Fix inconsistent usage of "Positional argument" https://review.openstack.org/189390 | 13:06 |
| *** ppai has joined #openstack-swift | 13:06 | |
| *** NM has joined #openstack-swift | 13:08 | |
| *** wbhuber__ has joined #openstack-swift | 13:10 | |
| *** SkyRocknRoll has quit IRC | 13:16 | |
| *** jrichli has joined #openstack-swift | 13:19 | |
| *** petertr7_away is now known as petertr7 | 13:22 | |
| *** qwebirc99779 has joined #openstack-swift | 13:24 | |
| *** qwebirc99779 has quit IRC | 13:24 | |
| openstackgerrit | Charles Hsu proposed openstack/python-swiftclient: Add ability to download objects to particular folder. https://review.openstack.org/160283 | 13:32 |
| *** acampbell has joined #openstack-swift | 13:36 | |
| *** thurloat has quit IRC | 13:42 | |
| *** joeljwright has joined #openstack-swift | 13:43 | |
| *** swat30 has quit IRC | 13:43 | |
| *** acampbell has quit IRC | 13:46 | |
| *** acampbell has joined #openstack-swift | 13:46 | |
| *** thurloat has joined #openstack-swift | 13:47 | |
| *** swat30 has joined #openstack-swift | 13:48 | |
| *** zaitcev has joined #openstack-swift | 13:55 | |
| *** ChanServ sets mode: +v zaitcev | 13:55 | |
| *** lastops has joined #openstack-swift | 13:59 | |
| *** breitz has quit IRC | 14:02 | |
| *** breitz has joined #openstack-swift | 14:03 | |
| *** lcurtis has joined #openstack-swift | 14:07 | |
| lcurtis | Does swift provide out of box object encryption? | 14:14 |
| ppai | lcurtis, not yet but soon | 14:14 |
| lcurtis | ppai thanks! | 14:14 |
| ppai | lcurtis, http://specs.openstack.org/openstack/swift-specs/specs/in_progress/at_rest_encryption.html | 14:14 |
| lcurtis | awesome..thnk u | 14:15 |
| *** adutta has joined #openstack-swift | 14:21 | |
| *** ptb has joined #openstack-swift | 14:27 | |
| adutta | We are currently working on upgrading our clusters from swift 1.12 to 2.2.2. We are upgrading the object, container and the account servers first before upgrading the proxy. In the datacenters where we have done the upgrade, the account auditor is reporting an error something like this "The total object_count for the container (5) does not match the sum of object_count across policies (0)" for some of the accounts. | 14:29 |
| adutta | After investigating more I found out that the policy_stat table is missing from the account db for those accounts and as such the object_count across policies is 0. I checked the account db for an account which has no error and the policy_stat table is there. Does this depend on the proxy upgrade. My understanding is that it shouln't. If it doesn't what is the best remedy for this issue? I can give more details from my investi | 14:29 |
| *** lpabon has joined #openstack-swift | 14:31 | |
| *** ptb has quit IRC | 14:35 | |
| *** ptb has joined #openstack-swift | 14:39 | |
| *** openstackgerrit has quit IRC | 14:41 | |
| *** openstackgerrit has joined #openstack-swift | 14:41 | |
| *** ronenkat has quit IRC | 14:41 | |
| *** minwoob has joined #openstack-swift | 14:45 | |
| *** bill_az has joined #openstack-swift | 15:09 | |
| bill_az | general question: we have a customer requesting support user quotas; has this come up as a common request? Anyone thought about providing this kind of feature? | 15:13 |
| cschwede | bill_az: user quotas? there are already account quotas, is that not sufficient? | 15:15 |
| bill_az | cschwede: account and container quotas are not enough :-) Before I reply to customer, I wanted to find out if this kind of request had come up before, is common. | 15:16 |
| bill_az | cschwede: they would like to specify an amount of storage for each user, and it can be used across containers within an account | 15:18 |
| cschwede | bill_az: it’s the first time i hear about this kind of request - i think many deployments simply assign an account to a „real“ user, and that user can do whatever is needed in it’s own account | 15:19 |
| *** ppai has quit IRC | 15:22 | |
| *** janonymous_ has joined #openstack-swift | 15:25 | |
| cschwede | adutta: it doesn’t depend on the proxy update; the missing tables should be created by the replicators. so it should fix this issue itself | 15:25 |
| *** fthiagogv has joined #openstack-swift | 15:25 | |
| *** btorch has left #openstack-swift | 15:25 | |
| *** leopoldj has quit IRC | 15:26 | |
| *** SkyRocknRoll has joined #openstack-swift | 15:26 | |
| *** acampbell has quit IRC | 15:30 | |
| *** acampbell has joined #openstack-swift | 15:30 | |
| adutta | cschwede: That is what I was expecting after looking at the replicator code but it doesn't seem to be happening for some reason. The errors have persisted for a long time. | 15:30 |
| *** sileht has quit IRC | 15:33 | |
| *** sileht has joined #openstack-swift | 15:36 | |
| si1v3r | zaitcev: This swift cluster was built by someone not here anymore, so I don't know if a device was named 'account'. How can I get a list of devices from the container updater to see? | 15:37 |
| si1v3r | Thanks for finding that line in the source. | 15:37 |
| cschwede | adutta: hmm, strange. can you open a new bugreport (https://bugs.launchpad.net/openstack/swift) and share a few more details? | 15:38 |
| *** sileht has quit IRC | 15:39 | |
| *** sileht has joined #openstack-swift | 15:40 | |
| adutta | cschwede: No problem. I will do that. I did check that the account-replicator is running and it has completed multiple runs successfully. I don't see any errors from the account-replicator in the logs. For reference, here is a snippet from the log related to replication http://paste.openstack.org/show/281592/ and here is related to auditor http://paste.openstack.org/show/281595/ | 15:42 |
| *** wbhuber__ has quit IRC | 15:42 | |
| *** ptb has quit IRC | 15:42 | |
| *** petertr7 is now known as petertr7_away | 15:44 | |
| cschwede | adutta: thx, that helps. I’ll will do a few tests later on my dev env, trying to replicate this error | 16:01 |
| cschwede | adutta: i did a quick test locally: dropped the table policy_stat in all account db replicas, run the auditor, saw your error, run the replicators, run the auditors again and the error disappeared | 16:07 |
| cschwede | adutta: which swift version are you running? | 16:07 |
| adutta | cschwede: 2.2.2. We upgraded from 1.12 to 2.2.2 | 16:07 |
| cschwede | oh yeah, you wrote that above. hmm | 16:08 |
| adutta | cschwede: Is it something related to https://bugs.launchpad.net/swift/+bug/1424108 ? | 16:09 |
| openstack | Launchpad bug 1424108 in OpenStack Object Storage (swift) "KeyError: 'storage_policy_index' in _really_merge_items swift/account/backend.py" [Critical,Fix committed] - Assigned to clayg (clay-gerrard) | 16:09 |
| cschwede | adutta: oh yes, good catch! and there is also the reason in the commit msg why you see this on some accounts: | 16:10 |
| cschwede | „but if the account databases have rows out of sync and there's no activity in | 16:10 |
| cschwede | the account otherwise, there's nothing to force the old schemas to be | 16:10 |
| cschwede | upgraded.“ | 16:10 |
| *** annegentle has joined #openstack-swift | 16:11 | |
| *** jistr has quit IRC | 16:14 | |
| adutta | cschwede: Thanks for the help. So errors should go away for the accounts as we see some activity for them, if i get it correctly. | 16:17 |
| *** proteusguy has quit IRC | 16:17 | |
| *** proteusguy has joined #openstack-swift | 16:20 | |
| cschwede | adutta: you’re welcome! and yes, any modification on the account level should fix this (or updating swift, if that is an option for you) | 16:20 |
| notmyname | good morning | 16:39 |
| *** jordanP has quit IRC | 16:39 | |
| *** petertr7_away is now known as petertr7 | 16:39 | |
| *** acoles is now known as acoles_away | 16:42 | |
| tdasilva | notmyname: good morning | 16:45 |
| notmyname | the thing is, now that we have so many people working on swift that aren't in _my_ timezone, I have to start every day with dozens and dozens of emails from gerrit ;-) | 16:46 |
| tdasilva | i guess that's a good problem to have??? | 16:46 |
| notmyname | definitely! | 16:46 |
| tdasilva | :) | 16:46 |
| tdasilva | notmyname, cschwede, acoles_away: wanted to get your opinion on patch 182564 | 16:46 |
| patchbot | tdasilva: https://review.openstack.org/#/c/182564/ | 16:46 |
| tdasilva | i had a follow-on patch to fix DLO, but thinking more about it I'm not sure it is the same case | 16:47 |
| tdasilva | in the case of SLO, the header is set by the system, but in the case of DLO it is actually set by the user. so I guess it would be ok to require the user to always set the flag on subsequent POST requests | 16:48 |
| *** haypo has quit IRC | 16:48 | |
| tdasilva | s/set the flag/set the header | 16:48 |
| *** zhill_desktop has joined #openstack-swift | 16:54 | |
| *** esmute has joined #openstack-swift | 16:55 | |
| *** SkyRocknRoll has quit IRC | 16:55 | |
| *** petertr7 is now known as petertr7_away | 16:56 | |
| *** annegentle has quit IRC | 16:56 | |
| *** esmute has quit IRC | 17:00 | |
| *** zhill_desktop has quit IRC | 17:01 | |
| *** janonymous_ has quit IRC | 17:08 | |
| *** petertr7_away is now known as petertr7 | 17:15 | |
| *** esmute has joined #openstack-swift | 17:22 | |
| *** esmute_ has joined #openstack-swift | 17:30 | |
| *** esmute has quit IRC | 17:30 | |
| *** esmute_ has quit IRC | 17:30 | |
| *** esmute has joined #openstack-swift | 17:31 | |
| *** B4rker has joined #openstack-swift | 17:31 | |
| *** aix has quit IRC | 17:33 | |
| *** petertr7 is now known as petertr7_away | 17:37 | |
| *** zul has quit IRC | 17:38 | |
| *** zul has joined #openstack-swift | 17:40 | |
| *** petertr7_away is now known as petertr7 | 17:42 | |
| *** ronenkat has joined #openstack-swift | 17:46 | |
| *** lastops has quit IRC | 17:49 | |
| *** zhill_desktop has joined #openstack-swift | 17:53 | |
| *** geaaru has quit IRC | 17:53 | |
| *** annegentle has joined #openstack-swift | 17:56 | |
| *** harlowja has quit IRC | 17:58 | |
| *** blmartin has joined #openstack-swift | 18:01 | |
| *** harlowja has joined #openstack-swift | 18:02 | |
| *** lastops has joined #openstack-swift | 18:02 | |
| *** annegentle has quit IRC | 18:02 | |
| *** annegentle has joined #openstack-swift | 18:12 | |
| *** B4rker has quit IRC | 18:14 | |
| *** ronenkat_ has joined #openstack-swift | 18:17 | |
| *** ronenkat has quit IRC | 18:20 | |
| *** potato_farmer has joined #openstack-swift | 18:21 | |
| peluse | clayg? ping me when you're around. I'll be in an out, check email if I'm not here (question on a patch you sent over) | 18:22 |
| potato_farmer | Hi. I set up swift with keystone (non-SSL). I was hoping to use Apache with SSL to proxy to the non-SSL proxy-server. Using a simply proxy does not seem to work. Do I need to configure the proxy-server and keystone to use SSL and skip Apache? Or is there a way to use Apache with SSL to proxy to a non-SSL proxy-server setup? | 18:23 |
| notmyname | potato_farmer: you've got several questions in there | 18:25 |
| clayg | peluse: it's not the one i responded to yesterday? | 18:25 |
| notmyname | let me try to separate them | 18:25 |
| notmyname | potato_farmer: first SSL vs non-SSL | 18:25 |
| potato_farmer | I suppose the questions are (apache ssl -> non-ssl swift) or (ssl swift and keystone - no apache needed) | 18:26 |
| notmyname | potato_farmer: clients should always talk to swift via TLS. you can terminate it on the proxy server box(es) or on a box in front of them (eg a load balancer). and so keystone should be handing back to the user an https URL | 18:26 |
| notmyname | most deployments I know of either terminate TLS in the load balancer or in a process on the proxy server using something like stud | 18:27 |
| potato_farmer | I see. So the client gets additional URLs during the transaction, which would explain why simply proxying the auth_URL doesn't work. | 18:27 |
| notmyname | don't use the actual swift proxy server process to terminate ssl | 18:27 |
| peluse | clayg, checking... | 18:28 |
| notmyname | potato_farmer: now, you could use apache to proxy swift. a while back there was some work in swift to do that. swift is a python wsgi server, so you can integrate that with apache | 18:29 |
| peluse | clayg, OK, see it now. Too man f'ing emails lately :) | 18:29 |
| *** gyee has joined #openstack-swift | 18:29 | |
| notmyname | potato_farmer: however, I don't know of any prod (or dev) deployments doing that | 18:29 |
| peluse | clayg, after we are done getting things cleaned off and udpated, if I run into issues there I'll cleanup/root cause remaining things and propose a patch based on what you sent | 18:29 |
| notmyname | potato_farmer: normally deployers use haproxy or some other load balancer to talk directly to the proxy server processes (eg client -> LB -> swift-proxy-server) | 18:30 |
| *** si1v3r has quit IRC | 18:30 | |
| potato_farmer | This setup is for testing only. We're running tests comparing different object storage solutions. I'll look into haproxy and possibly enabling SSL for keystone and see if any of that helps. | 18:31 |
| potato_farmer | Thanks for your help. | 18:31 |
| notmyname | potato_farmer: functional comparison or performance comparison? | 18:31 |
| notmyname | potato_farmer: you can certainly use swift without TLS. the reason to use it is to protect the auth tokens. but if you're just doing testing, not having that complexity in your deployment may make things simpler | 18:32 |
| *** B4rker has joined #openstack-swift | 18:32 | |
| potato_farmer | I think it's performance testing against things like ceph, etc... | 18:36 |
| potato_farmer | I'm just supporting the users, so I don't have all the details | 18:36 |
| notmyname | ah. ok | 18:36 |
| notmyname | definitely feel free to come in here and ask questions. there's a lot of people in here who have deployed swift at all different scales | 18:37 |
| potato_farmer | thanks, will do :) | 18:37 |
| notmyname | potato_farmer: what are you using object storage for? | 18:37 |
| * notmyname loves to hear new use cases and new users | 18:38 | |
| potato_farmer | My assumption is that at the moment we're testing different solutions to determine which is the optimal one to use for storage of big data (things like climate data, etc). I don't have all the details so this is my best guestimate. | 18:42 |
| notmyname | cool | 18:43 |
| notmyname | swift is really good at storing large amounts of data. I know some places where it's used today for genomics and cancer research (ie both large-scale scientific data sets that can be individually very large) | 18:43 |
| notmyname | it's also used quite a bit for storing large numbers of documents, videos, web contenct, and backups in various other deployments | 18:44 |
| notmyname | hmm...just last week I was talking to some people considering swift for climate data storage | 18:45 |
| notmyname | (from a large organization that is known for going to space) | 18:45 |
| openstackgerrit | Merged openstack/swift: go: ring self-reloading https://review.openstack.org/187030 | 18:46 |
| potato_farmer | Yes I am quite familiar with that organization. ;) | 18:46 |
| notmyname | heh | 18:46 |
| notmyname | potato_farmer: were you in santa clara last monday at a swiftstack workshop? | 18:46 |
| notmyname | at the MSST conference | 18:47 |
| potato_farmer | No I did not attend that. | 18:47 |
| notmyname | ah ok | 18:47 |
| *** petertr7 is now known as petertr7_away | 18:48 | |
| *** gvernik has joined #openstack-swift | 18:49 | |
| gvernik | i am still confused with hours...Swift IRC meeting in one hour, right? | 18:49 |
| notmyname | gvernik: nope | 18:49 |
| notmyname | gvernik: 2100UTC | 18:49 |
| notmyname | 2 hours | 18:50 |
| gvernik | aha....12:00 am :) for me ... | 18:51 |
| *** potato_farmer has quit IRC | 18:51 | |
| *** petertr7_away is now known as petertr7 | 18:52 | |
| *** silor1 has joined #openstack-swift | 18:52 | |
| *** silor has quit IRC | 18:52 | |
| notmyname | gvernik: ouch | 18:52 |
| gvernik | it's ok :) well...Mathew woke up at 04:00 for a year...now it's my turn for the extreeme hours | 18:54 |
| *** B4rker has quit IRC | 18:59 | |
| openstackgerrit | Merged openstack/swift: go: fix output of /recon/mounted https://review.openstack.org/190012 | 19:00 |
| *** cutforth has joined #openstack-swift | 19:00 | |
| *** lastops has quit IRC | 19:01 | |
| notmyname | it's very nice when you think you need to write something and then you realize that a library function does exactly what you were wanting to do | 19:02 |
| notmyname | in my case right now, eventlet.greenthread.spawn_after() | 19:02 |
| *** B4rker has joined #openstack-swift | 19:14 | |
| openstackgerrit | Charles Hsu proposed openstack/python-swiftclient: Add ability to download objects to particular folder. https://review.openstack.org/160283 | 19:15 |
| *** lastops has joined #openstack-swift | 19:17 | |
| *** silor1 has quit IRC | 19:19 | |
| *** NM has quit IRC | 19:36 | |
| *** B4rker has quit IRC | 19:37 | |
| *** B4rker has joined #openstack-swift | 19:42 | |
| *** NM has joined #openstack-swift | 19:43 | |
| *** RackerShagz has joined #openstack-swift | 19:49 | |
| *** B4rker has quit IRC | 19:51 | |
| *** lpabon has quit IRC | 19:53 | |
| *** aix has joined #openstack-swift | 19:53 | |
| *** B4rker has joined #openstack-swift | 19:54 | |
| *** thurloat is now known as thurloat_isgone | 20:07 | |
| *** fthiagogv has quit IRC | 20:16 | |
| *** lastops has quit IRC | 20:18 | |
| *** haypo has joined #openstack-swift | 20:20 | |
| *** annegentle has quit IRC | 20:26 | |
| *** dstanek has left #openstack-swift | 20:36 | |
| notmyname | reminder: swift team meeting is in 20 minutes in #openstack-meeting | 20:40 |
| openstackgerrit | Michael Barton proposed openstack/swift: go: limit replication per disk https://review.openstack.org/190340 | 20:40 |
| *** ho has joined #openstack-swift | 20:49 | |
| ho | good morning | 20:50 |
| notmyname | hello ho | 20:50 |
| ho | notmyname: hello | 20:52 |
| *** kota_ has joined #openstack-swift | 20:53 | |
| *** ChanServ sets mode: +v kota_ | 20:53 | |
| kota_ | morning | 20:53 |
| ho | kota_: morning! | 20:54 |
| *** dmorita has joined #openstack-swift | 20:56 | |
| *** acoles_away is now known as acoles | 20:56 | |
| *** B4rker has quit IRC | 20:57 | |
| *** B4rker_ has joined #openstack-swift | 20:57 | |
| mattoliverau | morning | 20:58 |
| kota_ | mattoliverau: hi :) | 20:59 |
| dmorita | morning | 20:59 |
| notmyname | look at all the people :-) | 21:00 |
| notmyname | ok, meeting time in #openstack-meeting | 21:00 |
| *** Barker has joined #openstack-swift | 21:01 | |
| *** B4rker_ has quit IRC | 21:02 | |
| *** petertr7 is now known as petertr7_away | 21:02 | |
| acoles | tdasilva: i agree re patch 182564 - DLO is different since client sets header | 21:03 |
| patchbot | acoles: https://review.openstack.org/#/c/182564/ | 21:03 |
| tdasilva | acoles: thanks! | 21:03 |
| tdasilva | kota_: ^^^ | 21:03 |
| *** wbhuber has joined #openstack-swift | 21:03 | |
| *** cutforth has quit IRC | 21:05 | |
| *** gvernik has quit IRC | 21:13 | |
| *** RackerShagz has quit IRC | 21:15 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 21:27 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 21:28 |
| openstackgerrit | Roxana Gherle proposed openstack/swift: Keystonemiddleware needs a "project" config option https://review.openstack.org/190370 | 21:53 |
| *** marcusvrn has quit IRC | 21:54 | |
| openstackgerrit | Roxana Gherle proposed openstack/swift: Keystonemiddleware needs a "project" config option https://review.openstack.org/190370 | 21:55 |
| *** haypo has left #openstack-swift | 21:57 | |
| notmyname | whew. that was a jam-packed meeting | 21:59 |
| kota_ | exactly | 22:00 |
| *** dmorita has quit IRC | 22:00 | |
| *** esmute has quit IRC | 22:00 | |
| MooingLemur | I'm happy about changing policies being in the pipeline :D | 22:00 |
| kragniz | I was waiting for open discussion! | 22:00 |
| kragniz | someone said to bring this to the meeting: https://review.openstack.org/#/c/189815/ | 22:00 |
| notmyname | kragniz: I'm sorry we didn't get to it this week | 22:01 |
| torgomatic | MooingLemur: I don't know how I feel about that one. It seems like a really dangerous thing to let users go do | 22:01 |
| kragniz | notmyname: that's totally okay | 22:01 |
| MooingLemur | torgomatic: I feel it should be disabled by default, but perhaps open to reseller admins | 22:01 |
| torgomatic | operators, sure, they care about the health of the cluster... end users just want stuff done now, consequences be damned | 22:01 |
| torgomatic | MooingLemur: that seems like a good starting point if this is something we're going to do | 22:02 |
| notmyname | now I've got to run to another meeting. be back in about an hour ish | 22:02 |
| MooingLemur | I really want to move swauth's containers to its own storage policy, but that seems difficult on a live prod cluster. | 22:02 |
| torgomatic | swauth rotates its containers over time, yes? | 22:03 |
| MooingLemur | Well, there's token_[0-9a-f] which would be easier to move manually than the places where the authentication stores are. | 22:04 |
| *** dmorita has joined #openstack-swift | 22:04 | |
| MooingLemur | I forget what those container names are | 22:04 |
| kota_ | I pinged dmorit | 22:04 |
| torgomatic | oh, okay, so not rotated over time... never mind then | 22:04 |
| dmorita | Sorry, I have logouted from IRC so i cannot see previous discussion | 22:05 |
| openstackgerrit | Minwoo Bae proposed openstack/swift: After the .durable has been written, fsync the directory. https://review.openstack.org/184113 | 22:06 |
| tdasilva | dmorita: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/latest.log.html | 22:06 |
| *** NM has quit IRC | 22:06 | |
| * mattoliverau is going to find some breakfast bbl | 22:06 | |
| dmorita | tdasilve: Thanks | 22:07 |
| *** kota_ has quit IRC | 22:07 | |
| MooingLemur | torgomatic: I've been trying to pin down clayg after he was helping me a couple weeks ago try to pin down what was going on in a scenario where most of the objects in an EC container were lost after reshuffling the ring. I haven't been able to get some free cycles from him since then :) I was wondering if you have a few minutes to take a look at what I have at the moment. I've pretty much left everything in that cluster alone ... | 22:09 |
| MooingLemur | ... since the incident. | 22:09 |
| torgomatic | MooingLemur: unfortunately I'm about to get sucked into a meeting here, so not at the moment | 22:09 |
| MooingLemur | so, I have two EC containers (on the same EC policy). I think the one that suffered data loss is one where I uploaded the data while the ring was unbalanced and some of the partitions had only 11 replicas instead of 12 | 22:10 |
| MooingLemur | gotcha. | 22:10 |
| MooingLemur | In case you or someone can think of the right questions to ask me later: to summarize, I see logs for some objects where the reconstructor has enough fragments for reconstruction, but they never get reconstructed. And it's been like that for a couple weeks. It's a 9+3 policy, and there are 10 surviving fragments for this one object, but it never becomes whole again. | 22:12 |
| MooingLemur | if the answer is "the fragments get hosed when your ring is broken", I can accept that. If an EC ring has the wrong number of replicas for a partition (smaller number than required), maybe some of the logic is destructive. That's my current hypothesis. | 22:15 |
| *** acoles is now known as acoles_away | 22:16 | |
| MooingLemur | and if that's the case, maybe the fragments that are in this object that appears to be sufficiently available are in reality not valid, and when passed into the EC library, something is failing. (no exceptions in the logs that I can find though) | 22:17 |
| torgomatic | MooingLemur: (briefly) if you can HEAD each of the fragments individually and get that output somewhere, that'd be great... it'd let us know if there's a dupe fragment or something | 22:18 |
| MooingLemur | sure | 22:18 |
| MooingLemur | thanks :) | 22:19 |
| *** ptb has joined #openstack-swift | 22:20 | |
| *** ptb has quit IRC | 22:20 | |
| *** ptb has joined #openstack-swift | 22:21 | |
| *** jrichli has quit IRC | 22:24 | |
| *** harlowja has quit IRC | 22:27 | |
| *** asettle has joined #openstack-swift | 22:29 | |
| *** harlowja has joined #openstack-swift | 22:32 | |
| *** blmartin has quit IRC | 22:36 | |
| *** dmorita has quit IRC | 22:40 | |
| *** Barker has quit IRC | 22:45 | |
| *** ptb has quit IRC | 22:47 | |
| MooingLemur | torgomatic: https://bpaste.net/raw/f23b4ada31dc | 22:50 |
| *** km has joined #openstack-swift | 23:07 | |
| *** joeljwright has quit IRC | 23:19 | |
| *** kei_yama has joined #openstack-swift | 23:21 | |
| openstackgerrit | Roxana Gherle proposed openstack/swift: Keystonemiddleware needs a "project" config option https://review.openstack.org/190370 | 23:22 |
| *** asettle has quit IRC | 23:28 | |
| *** david-lyle has quit IRC | 23:40 | |
| *** ptb has joined #openstack-swift | 23:51 | |
| *** ptb has quit IRC | 23:52 | |
| *** petertr7_away is now known as petertr7 | 23:54 | |
| *** harlowja has quit IRC | 23:54 | |
| *** harlowja has joined #openstack-swift | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!