| *** ^andrea^ has quit IRC | 00:06 | |
| *** caiobrentano has joined #openstack-swift | 00:06 | |
| *** frank_young has joined #openstack-swift | 00:15 | |
| *** frank_young has quit IRC | 00:20 | |
| *** lucasxu has quit IRC | 00:22 | |
| *** lucasxu has joined #openstack-swift | 00:35 | |
| *** xrb has quit IRC | 00:45 | |
| *** kiennt has joined #openstack-swift | 00:49 | |
| *** xrb has joined #openstack-swift | 00:52 | |
| *** tovin07_ has joined #openstack-swift | 00:52 | |
| *** lucasxu has quit IRC | 00:55 | |
| *** ukaynar_ has quit IRC | 01:09 | |
| *** JimCheung has quit IRC | 01:15 | |
| *** caiobrentano has quit IRC | 01:15 | |
| openstackgerrit | junbo.li proposed openstack/swift master: [Api-ref] fix response status https://review.openstack.org/484254 | 01:18 |
|---|---|---|
| *** frank_young has joined #openstack-swift | 01:23 | |
| *** JimCheung has joined #openstack-swift | 01:30 | |
| *** vikram has joined #openstack-swift | 01:30 | |
| *** JimCheung has quit IRC | 01:35 | |
| *** caiobrentano has joined #openstack-swift | 01:43 | |
| *** caiobrentano has quit IRC | 01:43 | |
| *** RayLei has joined #openstack-swift | 01:44 | |
| *** JimCheung has joined #openstack-swift | 01:47 | |
| *** JimCheung has quit IRC | 01:51 | |
| kota_ | good morning | 02:10 |
| kota_ | timburke: nice, go code!! | 02:14 |
| *** gyee has quit IRC | 02:17 | |
| kota_ | notmyname: for now, the script reached at k=6, m=20. To make it fast, I need more computation resource, more server or many core cpus. It doesn't waste the memory so just cpu cores are needed to run in parallel. | 02:22 |
| kota_ | i think, even using timburke's one, it requires a many core machine to run speedy | 02:23 |
| *** lucasxu has joined #openstack-swift | 02:30 | |
| *** caiobrentano has joined #openstack-swift | 02:49 | |
| *** lucasxu has quit IRC | 02:53 | |
| *** caiobrentano has quit IRC | 02:55 | |
| *** JimCheung has joined #openstack-swift | 02:57 | |
| *** JimCheung has quit IRC | 03:02 | |
| *** ukaynar has joined #openstack-swift | 03:03 | |
| *** SkyRocknRoll has joined #openstack-swift | 03:14 | |
| *** lucasxu has joined #openstack-swift | 03:17 | |
| *** gkadam has joined #openstack-swift | 03:53 | |
| *** gkadam has quit IRC | 03:53 | |
| *** gkadam has joined #openstack-swift | 03:53 | |
| *** hieulq has quit IRC | 03:57 | |
| *** hieulq has joined #openstack-swift | 03:57 | |
| *** links has joined #openstack-swift | 03:58 | |
| *** lucasxu has quit IRC | 04:02 | |
| *** Dinesh_Bhor has joined #openstack-swift | 04:11 | |
| *** Dinesh_Bhor has quit IRC | 04:14 | |
| *** Dinesh_Bhor has joined #openstack-swift | 04:15 | |
| *** psachin has joined #openstack-swift | 04:21 | |
| *** RayLei has quit IRC | 04:25 | |
| *** frank_young has quit IRC | 04:26 | |
| *** frank_young has joined #openstack-swift | 04:36 | |
| *** frank_young has quit IRC | 04:41 | |
| *** ukaynar_ has joined #openstack-swift | 05:14 | |
| *** Dinesh_Bhor has quit IRC | 05:17 | |
| *** ukaynar has quit IRC | 05:17 | |
| *** ukaynar_ has quit IRC | 05:29 | |
| *** ukaynar has joined #openstack-swift | 05:29 | |
| *** ukaynar has quit IRC | 05:34 | |
| *** tonanhngo has quit IRC | 05:34 | |
| *** kiennt has quit IRC | 05:42 | |
| *** frank_young has joined #openstack-swift | 05:48 | |
| *** rcernin has joined #openstack-swift | 05:50 | |
| *** cshastri has joined #openstack-swift | 05:58 | |
| *** kiennt has joined #openstack-swift | 06:02 | |
| openstackgerrit | junbo.li proposed openstack/swift master: Enabling off-by-default checks https://review.openstack.org/474562 | 06:05 |
| *** SkyRocknRoll has quit IRC | 06:15 | |
| *** kiennt has quit IRC | 06:21 | |
| *** Dinesh_Bhor has joined #openstack-swift | 06:50 | |
| *** Dinesh_Bhor has quit IRC | 06:50 | |
| *** tonanhngo has joined #openstack-swift | 07:01 | |
| *** tonanhngo has quit IRC | 07:04 | |
| *** cschwede_ has joined #openstack-swift | 07:11 | |
| *** itlinux has quit IRC | 07:11 | |
| *** tesseract has joined #openstack-swift | 07:17 | |
| *** klrmn has quit IRC | 07:22 | |
| *** oshritf has joined #openstack-swift | 07:26 | |
| ntt | Hi, I have a swift cluster with 3 zones and 3 different servers (with replica=3). One of my servers went down and now I have to restart after 10 days of inactivity. When I try to restart rsyncd or any other service (with swift-init) it simply hangs without error messages or kernel panic. It seems to me that is a "swap problem", but I'm not sure because when it hangs I completely lost the access to the machine even with the local physical console. | 07:29 |
| ntt | Please, someone can help me? Thank you | 07:29 |
| *** catintheroof has joined #openstack-swift | 07:31 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:35 | |
| *** catintheroof has quit IRC | 07:36 | |
| *** kiennt has joined #openstack-swift | 07:40 | |
| *** cbartz has joined #openstack-swift | 07:46 | |
| *** openstackgerrit has quit IRC | 08:49 | |
| *** tovin07_ has quit IRC | 08:50 | |
| *** mvk has joined #openstack-swift | 09:50 | |
| *** RayLei has joined #openstack-swift | 10:16 | |
| RayLei | If swift has two regions, swift how to timely detect a region is not available？ | 10:17 |
| *** Dinesh_Bhor has joined #openstack-swift | 10:27 | |
| *** RayLei has quit IRC | 10:33 | |
| *** kiennt has quit IRC | 10:45 | |
| *** tonanhngo has joined #openstack-swift | 11:03 | |
| *** frank_young has quit IRC | 11:04 | |
| *** tonanhngo has quit IRC | 11:05 | |
| *** frank_young has joined #openstack-swift | 11:09 | |
| *** chlong_ has quit IRC | 11:43 | |
| *** caiobrentano has joined #openstack-swift | 11:47 | |
| *** hoonetorg has joined #openstack-swift | 12:09 | |
| *** catintheroof has joined #openstack-swift | 12:21 | |
| *** frank_young has quit IRC | 12:24 | |
| *** gkadam has quit IRC | 12:24 | |
| *** caiobrentano has joined #openstack-swift | 12:34 | |
| *** caiobrentano has quit IRC | 12:34 | |
| *** MVenesio has joined #openstack-swift | 12:37 | |
| *** RayLei has joined #openstack-swift | 12:40 | |
| RayLei | Can the middleware get all the actions of the object to the proxy? | 12:40 |
| RayLei | I repeat it again ，Can the middleware get the response from all objects to the proxy? | 12:42 |
| *** caiobrentano has joined #openstack-swift | 12:46 | |
| *** RayLei has quit IRC | 12:53 | |
| *** frank_young has joined #openstack-swift | 12:54 | |
| acoles | RayLei: if you mean individual backend object server responses to the proxy, no they are not visible to middleware | 12:55 |
| *** frank_young has quit IRC | 12:59 | |
| *** lucasxu has joined #openstack-swift | 13:06 | |
| *** ukaynar has joined #openstack-swift | 13:35 | |
| *** vint_bra has joined #openstack-swift | 13:36 | |
| ntt | Hi, I partially solved the problem of restarting a server after a long downtime (10 days). Almost all is resynchronized but I have some errors: Jul 19 15:36:18 r1z1 object-replicator: ERROR __call__ error with REPLICATE /sde1/918 : #012Traceback (most recent call last):#012 File "/usr/lib/python2.7/site-packages/swift/obj/server.py", line 904, in __call__#012 res = method(req)#012 File "/usr/lib/python2.7/site-packages/swift/common/utils.py", line | 13:38 |
| ntt | 2648, in wrapped#012 return func(*a, **kw)#012 File "/usr/lib/python2.7/site-packages/swift/common/utils.py", line 1205, in _timing_stats#012 resp = func(ctrl, *args, **kwargs)#012 File "/usr/lib/python2.7/site-packages/swift/obj/server.py", line 873, in REPLICATE#012 device, partition, suffixes, policy)#012 File "/usr/lib/python2.7/site-packages/swift/obj/diskfile.py", line 768, in get_hashes#012 self._get_hashes, partition_path, | 13:38 |
| ntt | recalculate=suffixes)#012 File "/usr/lib/python2.7/site-packages/swift/common/utils.py", line 3068, in force_run_in_thread#012 return self._run_in_eventlet_tpool(func, *args, **kwargs)#012 File "/usr/lib/python2.7/site-packages/swift/common/utils.py", line 3048, in _run_in_eventlet_tpool#012 raise result#012OSError: [Errno 21] Is a directory | 13:38 |
| ntt | It seems that there is a problem with replication. Someone can help me about this? Thank you | 13:38 |
| ntt | One solution could be to delete the folder sde1/objects/918 on the node I've rebooted? | 13:44 |
| *** frank_young has joined #openstack-swift | 13:52 | |
| *** nathaniel has joined #openstack-swift | 13:53 | |
| *** nathaniel is now known as Guest18367 | 13:53 | |
| Guest18367 | Can I use swift with other file system instead of xfs? | 13:54 |
| Guest18367 | in Object node | 13:54 |
| *** frank_young has quit IRC | 13:57 | |
| *** tonanhngo has joined #openstack-swift | 14:01 | |
| *** tonanhngo has quit IRC | 14:03 | |
| tdasilva | Guest18367: technically you could use any filesystem that supports xattr | 14:07 |
| tdasilva | but the great majority of people (AFAIK) use xfs | 14:08 |
| acoles | ntt: that traceback is hard to diagnose because the source of the exception isn't logged, but (I'm taking a guess) this may be because you have a corrupt hashes.pkl that is a dir - so check on the node that is being replicated to that all */objects/*/hashes.pkl are files or non-existent but not dirs | 14:09 |
| ntt | acoles: thank you. Can I simply delete the folder sde1/objects/918 ? | 14:10 |
| ntt | swift should replicate again (I think) | 14:10 |
| acoles | ntt: that folder may have object data in it | 14:11 |
| ntt | I know | 14:11 |
| ntt | but I have 3 zones on 3 different servers. If I "accidentally" delete sde1/objects/918 from one zone I think swift should replicate again. Is this wrong? | 14:11 |
| acoles | if you suspect folder is the cause then check the sde1/objects/918/hashes.pkl is not a dir | 14:11 |
| ntt | hashes.pkl IS a dir inside the folder | 14:12 |
| tdasilva | ntt: so I think notmyname had suggested completely erasing the whole drive since it's coming from a long downtime, did you consider that? | 14:12 |
| tdasilva | ntt: if you are running with default config, you might run into a chance of introducing dark data in your system | 14:12 |
| ntt | tdasilva: yes... but the complete resync is a long task and I need to solve fast | 14:13 |
| Guest18367 | tdasilva: thank you | 14:13 |
| acoles | ntt: ok it should not be a dir, so first you may want to check what is in that dir, possibly create a copy of it, but ultimately you'll need to delete the hashes.pkl that is a dir. | 14:13 |
| *** frank_young has joined #openstack-swift | 14:14 | |
| ntt | acoles: I have only one dir inside hashes.pkl with only 1 file inside it | 14:14 |
| ntt | for a total of 51K of data | 14:14 |
| acoles | ntt: also, take not of tdasilva's comment | 14:14 |
| ntt | ok | 14:15 |
| ntt | so, after a backup of hashes.pkl, I have to delete the folder. Right? | 14:15 |
| *** links has quit IRC | 14:18 | |
| *** Guest18367 has quit IRC | 14:19 | |
| acoles | ntt: yes, make a backup, delete the bad hashes.pkl folder, then next run of object replicator should replace it with a file. | 14:20 |
| *** frank_young has quit IRC | 14:30 | |
| *** frank_young has joined #openstack-swift | 14:35 | |
| ntt | acoles: It seems to work!! All errors disappeared | 14:38 |
| ntt | thank you | 14:38 |
| *** cshastri has quit IRC | 14:39 | |
| *** gyee has joined #openstack-swift | 14:39 | |
| *** frank_young has quit IRC | 14:40 | |
| acoles | ntt: out of curiosity, if you are able to share publically, what was in the hashes.pkl folder? (please paste to http://paste.openstack.org/ rather than dumping in irc) | 14:40 |
| ntt | I have only another error. From zone2 and zone3 I see an rsync error: http://paste.openstack.org/show/615861/ When I check /srv/node/sdb1/objects/487/92e on 192.168.128.1, 92e is a file and NOT a folder. How can I solve the issue? | 14:41 |
| ntt | acoles: sure! I can share publically and write some doc if needed | 14:41 |
| acoles | ntt: I'm just curious how it got in that state | 14:42 |
| ntt | acoles: no problem, I can give you all details about the cluster. Anyway, I need to solve last problem. Can you give me some advice? | 14:43 |
| xrb | hi all, making progress with my setup. Middlewares are really a cool (and handy way) to extend Swift's functionality! | 14:55 |
| xrb | I am currently getting issues with two things (not related to middlewares). 1) when a user creates a bucket, is the creator tracked somewhere (not seeing it with 's3cmd info' or 'swift stat').. Does he get some special permission on his bucket? | 14:57 |
| *** rloo has joined #openstack-swift | 14:58 | |
| xrb | 2) if not through (1), within a project, how are permissions on a bucket granted to users? Is it only done using 'ACLs' or is there also some other mechanism? | 15:00 |
| *** caiobrentano has quit IRC | 15:00 | |
| *** caiobrentano has joined #openstack-swift | 15:01 | |
| *** caiobrentano has quit IRC | 15:01 | |
| rloo | hi swift'ers. we've (ironic) been seeing sporatic issues with devstack, setting up swift: http://logs.openstack.org/64/483464/8/check/gate-tempest-dsvm-ironic-lib-partition-agent_ipmitool-ubuntu-xenial/a3125e6/logs/devstacklog.txt.gz | 15:01 |
| rloo | does that look familiar to anyone here? is it a swift issue, devstack issue, ?? | 15:02 |
| *** hoonetorg has quit IRC | 15:02 | |
| *** psachin has quit IRC | 15:08 | |
| *** frank_young has joined #openstack-swift | 15:13 | |
| *** klrmn has joined #openstack-swift | 15:22 | |
| *** Sukhdev_ has joined #openstack-swift | 15:39 | |
| *** gyee has quit IRC | 15:40 | |
| *** frank_young has quit IRC | 15:41 | |
| *** gyee has joined #openstack-swift | 15:42 | |
| acoles | xrb: usual caveat from me - answers are for keystoneauth rather than s3 :) 1) the creator id is not tracked when a container is created 2) users with an operator role for the project (e.g. admin) can read/write containers; ACLs of form user:project can grant cross-project access (but nor restrict within project) | 15:45 |
| acoles | xrb: there is another form of ACL that grants access to a user based on role on project | 15:46 |
| acoles | so for example if you have a user with role 'foo' on project (but not role admin) and set container ACL to foo, then that user can access the container. sadly I cannot find docs for that form of ACL :/ | 15:47 |
| tdasilva | rloo: are you referring to this error: http://logs.openstack.org/64/483464/8/check/gate-tempest-dsvm-ironic-lib-partition-agent_ipmitool-ubuntu-xenial/a3125e6/logs/devstacklog.txt.gz#_2017-07-19_13_54_31_908 ? | 15:49 |
| rloo | tdasilva: yes! sorry i should have been more specific | 15:49 |
| tdasilva | rloo: so, that's very interesting. I have personally seen a similar error when running our probe tests. It happens sporadically and it seems to do with restarting a service and sending a request right away to that service | 15:50 |
| tdasilva | sometimes the service is not ready yet, so you get that Errno 111] Connection refused | 15:51 |
| *** frank_young has joined #openstack-swift | 15:51 | |
| rloo | tdasilva: the ironic jobs have been failing due to a neutron issue, so we haven't been looking too much, just assuming neutron. i don't know when this started, but i think in the last week. maybe. any idea what might have changed recently? | 15:51 |
| *** gyee has quit IRC | 15:52 | |
| rloo | tdasilva: maybe things are just 'slower'? lame excuse... | 15:52 |
| rloo | tdasilva: what service is it looking for that may not be ready yet? | 15:52 |
| tdasilva | rloo: for the swift probe tests I belive I saw that with the proxy service, but it could be any service really... | 15:53 |
| tdasilva | do you know where I can find lib/swift:swift_configure_tempurls | 15:53 |
| tdasilva | see what it is oding | 15:53 |
| rloo | tdasilva: so far, i've only seen it wrt starting swift. does devstack start swift first? | 15:53 |
| rloo | tdasilva: sec, let me find link for you | 15:53 |
| rloo | tdasilva: https://github.com/openstack-dev/devstack/blob/master/lib/swift | 15:54 |
| rloo | tdasilva: does swift use devstack for their own testing? | 15:54 |
| tdasilva | rloo: not for the probe tests | 15:55 |
| tdasilva | rloo: https://8b86aea46fb38e6450f2-0e5f4c086da474abc1df58826577db2f.ssl.cf1.rackcdn.com/427911/7022/probetests/console.txt | 15:55 |
| tdasilva | these are the probe tests failures, search for ECONNREFUSED | 15:55 |
| rloo | tdasilva: yeah, looks similar | 15:56 |
| *** frank_young has quit IRC | 15:56 | |
| rloo | tdasilva: it can't be just ironic & swift that has noticed this... | 15:56 |
| *** frank_young has joined #openstack-swift | 16:00 | |
| *** klrmn has quit IRC | 16:05 | |
| tdasilva | rloo: so that issue looks very similar to what i have been seeing with our probe tests, so it could be swift regression, but I need to investigate further | 16:09 |
| tdasilva | I will open a bug in swift to track this | 16:09 |
| *** rcernin has quit IRC | 16:09 | |
| notmyname | good morning | 16:10 |
| *** vinsh has quit IRC | 16:11 | |
| *** caiobrentano has joined #openstack-swift | 16:11 | |
| rloo | tdasilva: thx!!! | 16:12 |
| *** oshritf has quit IRC | 16:13 | |
| *** oshritf has joined #openstack-swift | 16:15 | |
| *** oshritf has quit IRC | 16:17 | |
| *** cbartz has quit IRC | 16:18 | |
| *** frank_young has quit IRC | 16:23 | |
| *** rcernin has joined #openstack-swift | 16:23 | |
| timburke | good morning | 16:24 |
| *** JimCheung has joined #openstack-swift | 16:25 | |
| notmyname | https://review.openstack.org/#/c/448480 and https://review.openstack.org/#/c/478416/ and https://review.openstack.org/#/c/475038/ need some reviews, please | 16:29 |
| patchbot | patch 448480 - swift - DB replicator cleanup | 16:29 |
| patchbot | patch 478416 - swift - Add multiple worker processes strategy to reconstr... | 16:29 |
| patchbot | patch 475038 - python-swiftclient - Allow for uploads from standard input. | 16:29 |
| *** frank_young has joined #openstack-swift | 16:33 | |
| *** caiobrentano_ has joined #openstack-swift | 16:36 | |
| *** frank_young has quit IRC | 16:37 | |
| timburke | rloo: tdasilva: yeah, looks like a race between the proxy starting up and devstack configuring tempurls. makes me wonder how many retries osc uses, and what their backoff strategy is | 16:38 |
| *** caiobrentano has quit IRC | 16:38 | |
| rloo | timburke: sigh, something changed recently. Can't it just work. like. magic? | 16:41 |
| timburke | i'm gonna guess that change was to enable tempurls? | 16:42 |
| timburke | i *think* if account POST was done with python-swiftclient this wouldn't be a problem? we default to like 5 retries with an exponential backoff; all told i think that'd give the proxy something like 30 seconds to start (as opposed to the not-quite 2 seconds osc has between initial call and reporting the error) | 16:43 |
| rloo | timburke: no idea, wish i knew when it started happening. i suppose i could do some digging. i don't know much about swift. (i know very little.) | 16:44 |
| *** mvk has quit IRC | 16:51 | |
| *** ukaynar has quit IRC | 16:52 | |
| timburke | rloo: fwiw, it looks like tempurls were enabled back in feb: https://github.com/openstack-infra/project-config/commit/17c5302 -- does that about line up with when you started seeing these failures? i wonder if we just need a patch to add a `sleep 5` or something at the start of swift_configure_tempurls in lib/swift | 16:52 |
| *** ukaynar has joined #openstack-swift | 16:52 | |
| rloo | timburke: i only noticed it last week. which doesn't mean it didn't happen before. it could have been a rare occurrence before, but seems to be more frequent in the last week. | 16:53 |
| rloo | timburke: a sleep would work for me | 16:53 |
| *** vinsh has joined #openstack-swift | 16:56 | |
| *** ukaynar has quit IRC | 16:57 | |
| *** klrmn has joined #openstack-swift | 17:07 | |
| timburke | rloo: submitted patch 485282 | 17:12 |
| patchbot | https://review.openstack.org/#/c/485282/ - openstack-dev/devstack - When configuring temp urls, give Swift time to sta... | 17:12 |
| rloo | Thanks timburke! | 17:14 |
| rloo | timburke: do we need to open a bug for that? | 17:14 |
| timburke | *shrug* i just try to fix things :-) | 17:14 |
| rloo | timburke: that's a good attitude :D let's see what they say. | 17:15 |
| *** ukaynar has joined #openstack-swift | 17:18 | |
| tdasilva | timburke: the problem is that we are seeing something very similar with our probe tests, at first I tried adding sleeps everywhere, but I'm not sure that's a good solution, since we did not have that problem before | 17:21 |
| tdasilva | timburke: I tried reverting this change https://github.com/openstack/swift/commit/537f9a3f64428d73bec6776c1d6ee519b63a7769 | 17:22 |
| *** tonanhngo has joined #openstack-swift | 17:22 | |
| tdasilva | and it seemed to solve the problem for me | 17:22 |
| *** gyee has joined #openstack-swift | 17:23 | |
| *** ukaynar has quit IRC | 17:29 | |
| rloo | tdasilva, that might be it then. and/or it exasperated (sp) the problem... | 17:30 |
| *** ukaynar has joined #openstack-swift | 17:32 | |
| *** ukaynar_ has joined #openstack-swift | 17:36 | |
| *** ukaynar has quit IRC | 17:39 | |
| *** saint_ has quit IRC | 17:40 | |
| *** itlinux has joined #openstack-swift | 17:42 | |
| *** gyee has quit IRC | 17:42 | |
| *** gyee has joined #openstack-swift | 17:44 | |
| *** gyee has quit IRC | 17:44 | |
| *** mvk has joined #openstack-swift | 17:46 | |
| *** tesseract has quit IRC | 17:46 | |
| *** ntt_ has joined #openstack-swift | 18:24 | |
| *** u_nuSLASHkm8 has joined #openstack-swift | 18:24 | |
| ntt_ | Hi, I have an error with rsync in my swift cluster: from zone2 and zone3 (one server per zone) I see an rsync error: http://paste.openstack.org/show/615861/ When I check /srv/node/sdb1/objects/487/92e on 192.168.128.1 (zone 1), 92e is a file and NOT a folder. How can I solve the issue? | 18:25 |
| ntt_ | acoles: have you some advice? | 18:26 |
| ntt_ | can I simply delete the file in zone1? | 18:26 |
| *** u_nuSLASHkm8 has left #openstack-swift | 18:27 | |
| notmyname | kota_: timburke and I got his cauchy checker running on a different box with 16 cores, so we should be able to finish the checking | 18:29 |
| *** lucasxu has quit IRC | 18:32 | |
| notmyname | the important question now is, now that timburke isn't running the script on his laptop, how will he stay warm? | 18:34 |
| timburke | so... cold... | 18:34 |
| *** lucasxu has joined #openstack-swift | 18:35 | |
| notmyname | interesting/good lesson for our community to learn from: http://www.nickmilton.com/2017/07/a-story-of-how-community-lost-trust.html?utm_source=dlvr.it&utm_medium=twitter | 18:35 |
| timburke | rloo: tdasilva: yeah, that makes sense... loading the app takes some non-trivial amount of time; if we bind ports *before* doing that, we wait on a connection timeout rather than immediately coming back with a refused connection | 18:40 |
| *** lucasxu has quit IRC | 18:41 | |
| *** lucasxu has joined #openstack-swift | 18:41 | |
| *** ChubYann has joined #openstack-swift | 18:47 | |
| *** SkyRocknRoll has quit IRC | 18:53 | |
| *** cschwede_ has quit IRC | 18:54 | |
| *** lucasxu has quit IRC | 18:55 | |
| *** ntt_ has quit IRC | 19:00 | |
| *** frank_young has joined #openstack-swift | 19:01 | |
| *** frank_young has quit IRC | 19:06 | |
| *** chlong_ has joined #openstack-swift | 19:15 | |
| *** gyee has joined #openstack-swift | 19:18 | |
| *** silor has joined #openstack-swift | 19:27 | |
| *** lucasxu has joined #openstack-swift | 19:30 | |
| *** openstackgerrit has joined #openstack-swift | 19:32 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Ring rebalance respects co-builders' last_part_moves https://review.openstack.org/477000 | 19:32 |
| *** tinyurl_comSLASH has joined #openstack-swift | 19:36 | |
| *** tinyurl_comSLASH has left #openstack-swift | 19:38 | |
| *** lucasxu has quit IRC | 19:38 | |
| *** lucasxu has joined #openstack-swift | 19:39 | |
| *** lucasxu has quit IRC | 19:43 | |
| *** lucasxu has joined #openstack-swift | 19:44 | |
| *** frank_young has joined #openstack-swift | 19:44 | |
| *** frank_young has quit IRC | 19:48 | |
| *** lucasxu has quit IRC | 19:53 | |
| *** lucasxu has joined #openstack-swift | 19:54 | |
| *** silor has quit IRC | 19:56 | |
| *** chlong_ has quit IRC | 20:03 | |
| *** ujjain has quit IRC | 20:06 | |
| *** Sukhdev_ has quit IRC | 20:19 | |
| *** Sukhdev has joined #openstack-swift | 20:20 | |
| *** ujjain has joined #openstack-swift | 20:20 | |
| *** ujjain has joined #openstack-swift | 20:20 | |
| *** MVenesio has quit IRC | 20:41 | |
| *** chlong_ has joined #openstack-swift | 20:48 | |
| *** stewie925 has quit IRC | 20:51 | |
| kota_ | morning | 20:55 |
| *** lucasxu has quit IRC | 20:57 | |
| kota_ | notmyname: nice, thx! | 20:57 |
| notmyname | kota_: I just ran it. timburke wrote the code :-) | 20:58 |
| notmyname | I hope to have it finish this week sometime. it's running at 90% of 16 cores, so it should be soon ;-) | 20:58 |
| timburke | thanks for the code, https://docs.python.org/2/library/itertools.html#itertools.combinations ! | 20:58 |
| notmyname | meeting time in #openstack-meeting | 20:59 |
| kota_ | parallel processing is huge win! :D | 20:59 |
| notmyname | also, I started with the 16 data chunks, since we've already done up to that | 21:00 |
| timburke | kota_: were you going straight for the matrix in your testing, or doing a full encode/decode? | 21:00 |
| notmyname | it's currently at the end of checking 17 data | 21:01 |
| *** caiobrentano_ has quit IRC | 21:01 | |
| *** frank_young has joined #openstack-swift | 21:03 | |
| *** chlong_ has quit IRC | 21:04 | |
| kota_ | timburke: one encode, and decode/reconstruct for all available conbinations for each k + m parameters | 21:05 |
| kota_ | the work has been getting hard around k+m ~= 20 | 21:05 |
| *** frank_young has quit IRC | 21:08 | |
| timburke | i think that might also be netting me some performance improvements -- i don't actually go through the full encode/decode/reconstruct, but instead build the underlying k-by-k+m matrix and check for invertibility of k-by-k matrixes made from combinations of k columns | 21:08 |
| *** caiobrentano has joined #openstack-swift | 21:23 | |
| *** caiobrentano has quit IRC | 21:35 | |
| *** rloo has quit IRC | 21:38 | |
| *** Sukhdev has quit IRC | 21:42 | |
| *** Sukhdev has joined #openstack-swift | 21:43 | |
| *** m_kazuhiro has joined #openstack-swift | 21:54 | |
| timburke | tdasilva: wait, so barbicanclient needs to change from hitting /v1/... to hitting /1/...? that doesn't seem to line up with their api-ref at https://docs.openstack.org/barbican/latest/api/reference/secrets.html | 21:56 |
| timburke | and that would kinda explain all the 404s (starting around http://logs.openstack.org/26/484926/2/check/gate-python-barbicanclient-devstack-dsvm-ubuntu-xenial/243a654/console.html#_2017-07-19_20_02_51_411769) in that tempest run... | 21:58 |
| tdasilva | timburke: yeah i noticed that | 22:02 |
| tdasilva | but changing the barbicanclient/osc_plugin.py doesn't seem to work either | 22:02 |
| tdasilva | so I think I will need their help | 22:02 |
| timburke | i wonder if there's some issue with how the barbican endpoint gets defined in keystone...? | 22:02 |
| tdasilva | timburke: maybe that's it | 22:03 |
| tdasilva | this is what i get when changing osc_plugin | 22:03 |
| tdasilva | http://paste.openstack.org/show/615916/ | 22:03 |
| tdasilva | i don't see barbican endpoint getting setup in mathiasb script, so i'm assuming that's done by devstack | 22:06 |
| timburke | yeah, and it seems to be versionless -- https://github.com/openstack/barbican/blob/master/devstack/lib/barbican#L357-L371 | 22:07 |
| timburke | so i guess somewhere else in the client, we need to change '/%s/..' to be '/v%d/...' ? | 22:08 |
| timburke | https://github.com/openstack/python-barbicanclient/blob/master/barbicanclient/client.py#L44 maybe? | 22:09 |
| tdasilva | timburke: not sure, because like you said, i think the right aproach is not to change https://github.com/openstack/python-barbicanclient/blob/master/barbicanclient/client.py#L29 | 22:11 |
| tdasilva | no? | 22:11 |
| *** frank_young has joined #openstack-swift | 22:12 | |
| timburke | hrm. and it seems like the discovery api should be exposing 'v1'... https://github.com/openstack/barbican/blob/master/barbican/api/controllers/versions.py#L80-L84 | 22:15 |
| timburke | tdasilva: what happens if you change https://github.com/openstack/python-barbicanclient/blob/master/barbicanclient/osc_plugin.py#L18-L23 to use 'v1'? | 22:16 |
| *** frank_young has quit IRC | 22:16 | |
| timburke | oh, right, the mismatch thing you were saying before... | 22:16 |
| tdasilva | yep :( | 22:17 |
| timburke | was that *with* the other change? or *instead of*? | 22:17 |
| tdasilva | instead of | 22:17 |
| timburke | ...and like in your current change, you fixed both the default and the map? i'm trying to figure out how the '1' got there... | 22:21 |
| timburke | maybe i should just spin it up and poke around | 22:22 |
| *** m_kazuhiro has quit IRC | 22:30 | |
| *** caiobrentano has joined #openstack-swift | 22:34 | |
| *** itlinux has quit IRC | 22:58 | |
| *** frank_young has joined #openstack-swift | 22:59 | |
| *** vint_bra has quit IRC | 23:01 | |
| *** catintheroof has quit IRC | 23:03 | |
| *** ukaynar_ has quit IRC | 23:17 | |
| *** frank_young has quit IRC | 23:27 | |
| *** rcernin has quit IRC | 23:34 | |
| *** Renich has joined #openstack-swift | 23:36 | |
| *** frank_young has joined #openstack-swift | 23:37 | |
| *** frank_young has quit IRC | 23:41 | |
| *** itlinux has joined #openstack-swift | 23:47 | |
| *** Renich has quit IRC | 23:54 | |
| *** tonanhngo has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!