| *** asettle has quit IRC | 00:21 | |
| mattoliverau | ho: morning | 00:22 |
|---|---|---|
| ho | mattoliverau: morning! | 00:24 |
| *** dmorita has joined #openstack-swift | 00:25 | |
| mattoliverau | FYI it's a public holiday in Oz, so am not working.. Well not much.. And not at all if the wife asks :p | 00:26 |
| *** kutija_ is now known as kutija|away | 00:28 | |
| ho | mattoliverau: lol | 00:29 |
| *** asettle has joined #openstack-swift | 00:33 | |
| *** mitz has joined #openstack-swift | 00:47 | |
| *** sc has quit IRC | 00:58 | |
| *** jamielennox|away is now known as jamielennox | 01:07 | |
| *** kota_ has joined #openstack-swift | 01:17 | |
| *** ChanServ sets mode: +v kota_ | 01:17 | |
| kota_ | morning | 01:18 |
| ho | kota_: morning! | 01:20 |
| kota_ | ho: morning :) | 01:20 |
| *** aix has quit IRC | 01:26 | |
| mattoliverau | kota_: morning | 01:46 |
| kota_ | mattoliverau: morning :) | 01:49 |
| charz | kota_: morning | 02:03 |
| kota_ | charz: morning! and leaving a while for meeting... | 02:04 |
| *** kota_ has quit IRC | 02:08 | |
| *** thurloat_isgone has quit IRC | 03:27 | |
| *** patchbot has quit IRC | 03:27 | |
| *** tanee has quit IRC | 03:27 | |
| *** klrmn has quit IRC | 03:27 | |
| *** mwheckmann has quit IRC | 03:27 | |
| *** mlanner has quit IRC | 03:27 | |
| *** cschwede has quit IRC | 03:27 | |
| *** aluria has quit IRC | 03:27 | |
| *** saltsa has quit IRC | 03:27 | |
| *** omame has quit IRC | 03:27 | |
| *** ondergetekende has quit IRC | 03:27 | |
| *** saltsa has joined #openstack-swift | 03:27 | |
| *** tanee has joined #openstack-swift | 03:28 | |
| *** cschwede has joined #openstack-swift | 03:28 | |
| *** ChanServ sets mode: +v cschwede | 03:28 | |
| *** aluria has joined #openstack-swift | 03:28 | |
| *** ondergetekende has joined #openstack-swift | 03:28 | |
| *** thurloat_isgone has joined #openstack-swift | 03:28 | |
| *** thurloat_isgone is now known as thurloat | 03:28 | |
| *** mwheckmann has joined #openstack-swift | 03:28 | |
| *** klrmn has joined #openstack-swift | 03:28 | |
| *** mlanner has joined #openstack-swift | 03:30 | |
| *** patchbot has joined #openstack-swift | 03:31 | |
| *** janonymous has joined #openstack-swift | 03:31 | |
| *** mwheckmann has quit IRC | 03:50 | |
| *** silor has joined #openstack-swift | 03:52 | |
| *** ppai has joined #openstack-swift | 04:22 | |
| *** proteusguy has joined #openstack-swift | 04:39 | |
| *** Triveni has joined #openstack-swift | 05:22 | |
| *** dimasot has joined #openstack-swift | 05:29 | |
| openstackgerrit | Daisuke Morita proposed openstack/swift: Improvement of logging for invalid policies https://review.openstack.org/141238 | 06:01 |
| *** proteusguy has quit IRC | 06:09 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:12 | |
| *** SkyRocknRoll has quit IRC | 06:38 | |
| *** silor has quit IRC | 06:47 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 06:52 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:52 | |
| *** mmcardle has joined #openstack-swift | 06:57 | |
| *** kota_ has joined #openstack-swift | 07:01 | |
| *** ChanServ sets mode: +v kota_ | 07:01 | |
| *** remix_tj has quit IRC | 07:04 | |
| *** remix_tj has joined #openstack-swift | 07:06 | |
| *** proteusguy has joined #openstack-swift | 07:21 | |
| *** gvernik has joined #openstack-swift | 07:23 | |
| *** [1]dimasot has joined #openstack-swift | 07:28 | |
| *** gvernik has quit IRC | 07:28 | |
| *** dimasot has quit IRC | 07:30 | |
| *** [1]dimasot is now known as dimasot | 07:30 | |
| *** geaaru has joined #openstack-swift | 07:48 | |
| *** jistr has joined #openstack-swift | 07:52 | |
| *** proteusguy has quit IRC | 07:56 | |
| *** proteusguy has joined #openstack-swift | 07:58 | |
| hugokuo | ho: hihi, would you mind to have me the link of your commit for account ACL support in keystoneauth.py ? Thanks ! | 08:12 |
| *** acoles_away is now known as acoles | 08:14 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Enable Object Replicator's failure count in recon https://review.openstack.org/138342 | 08:17 |
| ho | hugokuo: hello, it's in WIP state. https://review.openstack.org/#/c/149930/ | 08:19 |
| ho | hugokuo: any comments on it are welcome! :-) | 08:20 |
| hugokuo | ho: thanks ! | 08:20 |
| hugokuo | an gonna to read it later. | 08:21 |
| ho | hugokuo: if you want to have one without merge conflict, I will remove it for you. | 08:21 |
| hugokuo | ho: not now. I'll check the entire logic before test it. I'll let you know it when I need more assist from you. | 08:22 |
| ho | hugokuo: thanks for having interests on it :-) | 08:23 |
| kota_ | ho: btw, do you have any state of gerrit failures? | 08:24 |
| ho | kota_: as for the patch #187489, I didn't get any info from infra. | 08:27 |
| acoles | morning | 08:27 |
| ho | acoles: Morning! | 08:27 |
| kota_ | ho: ok, thanks | 08:28 |
| kota_ | acoles: morning! | 08:28 |
| ho | kota_: np! | 08:29 |
| acoles | kota_: ho good evening! | 08:30 |
| kota_ | acoles: just FYI, i'm now planning this evening dinner with tsg- he is now in Japan :) | 08:32 |
| acoles | kota_: nice | 08:33 |
| acoles | swifterdarrell: thanks for sharing those measurements, v interesting! | 08:37 |
| *** aix has joined #openstack-swift | 08:41 | |
| *** kota_ has quit IRC | 08:44 | |
| *** proteusguy_ has joined #openstack-swift | 08:52 | |
| *** leopoldj has joined #openstack-swift | 08:54 | |
| *** proteusguy has quit IRC | 08:55 | |
| *** jordanP has joined #openstack-swift | 09:08 | |
| *** jd__ has left #openstack-swift | 09:09 | |
| *** proteusguy_ has quit IRC | 09:13 | |
| *** SkyRocknRoll_ has quit IRC | 09:27 | |
| *** haigang has joined #openstack-swift | 09:40 | |
| *** kei_yama has quit IRC | 09:45 | |
| *** ho has quit IRC | 09:54 | |
| *** odsail has joined #openstack-swift | 10:04 | |
| *** dimasot has quit IRC | 10:05 | |
| *** openstackgerrit has quit IRC | 10:09 | |
| *** openstackgerrit has joined #openstack-swift | 10:09 | |
| *** km has quit IRC | 10:11 | |
| *** km has joined #openstack-swift | 10:13 | |
| *** odsail has quit IRC | 10:19 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 10:25 | |
| openstackgerrit | Jiri Suchomel proposed openstack/swift: Let swift-object-info print information about files in given directory. https://review.openstack.org/189258 | 10:30 |
| *** dmorita has quit IRC | 10:43 | |
| *** dimasot has joined #openstack-swift | 10:47 | |
| *** km has quit IRC | 10:57 | |
| *** Triveni has quit IRC | 11:00 | |
| *** dimasot has quit IRC | 11:04 | |
| *** fesp has joined #openstack-swift | 11:20 | |
| openstackgerrit | Jiri Suchomel proposed openstack/swift: Let swift-object-info print information about files in given directory. https://review.openstack.org/189258 | 11:25 |
| *** haigang has quit IRC | 11:41 | |
| *** haigang has joined #openstack-swift | 11:42 | |
| *** mmcardle has quit IRC | 11:46 | |
| cschwede | swifterdarrell: awesome work, thank you very much! the current impact with multiple slow disks is freightening, and the improvement with your patch impressive | 11:57 |
| *** ppai has quit IRC | 12:04 | |
| *** ppai has joined #openstack-swift | 12:06 | |
| *** annegentle has joined #openstack-swift | 12:09 | |
| *** SkyRocknRoll has quit IRC | 12:13 | |
| *** haigang has quit IRC | 12:16 | |
| *** haigang has joined #openstack-swift | 12:16 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:16 | |
| *** haigang has quit IRC | 12:18 | |
| *** mmcardle has joined #openstack-swift | 12:20 | |
| *** fesp is now known as flaper87 | 12:25 | |
| *** flaper87 has quit IRC | 12:25 | |
| *** flaper87 has joined #openstack-swift | 12:25 | |
| *** acoles is now known as acoles_away | 12:38 | |
| *** ppai has quit IRC | 12:42 | |
| *** SkyRocknRoll_ has quit IRC | 12:47 | |
| *** petertr7_away is now known as petertr7 | 12:47 | |
| *** SkyRocknRoll has quit IRC | 12:49 | |
| *** tellesnobrega has quit IRC | 12:50 | |
| *** bi_fa_fu has joined #openstack-swift | 12:51 | |
| *** SkyRocknRoll has joined #openstack-swift | 13:04 | |
| *** SkyRocknRoll has joined #openstack-swift | 13:04 | |
| *** tellesnobrega has joined #openstack-swift | 13:06 | |
| *** acampbell has joined #openstack-swift | 13:09 | |
| *** proteusguy has joined #openstack-swift | 13:10 | |
| *** jkugel has quit IRC | 13:12 | |
| *** acampbell has quit IRC | 13:15 | |
| *** mwheckmann has joined #openstack-swift | 13:16 | |
| *** acampbell has joined #openstack-swift | 13:26 | |
| *** mwheckmann has quit IRC | 13:31 | |
| *** acampbell has quit IRC | 13:32 | |
| *** acampbell has joined #openstack-swift | 13:33 | |
| *** acampbel11 has joined #openstack-swift | 13:33 | |
| *** mwheckmann has joined #openstack-swift | 13:35 | |
| *** jkugel has joined #openstack-swift | 13:35 | |
| *** ptb has joined #openstack-swift | 13:52 | |
| *** ptb has quit IRC | 13:54 | |
| *** wbhuber has joined #openstack-swift | 13:55 | |
| *** ptb has joined #openstack-swift | 13:57 | |
| *** ptb has quit IRC | 13:59 | |
| *** blmartin has joined #openstack-swift | 14:03 | |
| *** B4rker has joined #openstack-swift | 14:04 | |
| *** ahale has quit IRC | 14:08 | |
| hurricanerix_ | good morning | 14:08 |
| *** ahale has joined #openstack-swift | 14:08 | |
| *** wbhuber_ has joined #openstack-swift | 14:10 | |
| *** wbhuber_ has quit IRC | 14:11 | |
| *** logan2 has quit IRC | 14:12 | |
| *** jrichli has joined #openstack-swift | 14:13 | |
| *** logan2 has joined #openstack-swift | 14:13 | |
| *** wbhuber has quit IRC | 14:14 | |
| *** lpabon has joined #openstack-swift | 14:19 | |
| *** lastops has joined #openstack-swift | 14:20 | |
| *** ptb has joined #openstack-swift | 14:26 | |
| *** proteusguy has quit IRC | 14:26 | |
| *** annegentle has quit IRC | 14:29 | |
| *** dstanek has joined #openstack-swift | 14:34 | |
| *** proteusguy has joined #openstack-swift | 14:36 | |
| *** mahatic has joined #openstack-swift | 14:36 | |
| *** jamielennox is now known as jamielennox|away | 14:38 | |
| swifterdarrell | acoles_away: cschwede: thanks! I'm excited about the I/O isolation offered at little-or-no overhead; now all we need is to get the review banged out. :) | 14:43 |
| swifterdarrell | acoles_away: cschwede: mattoliverau: tdasilva: notmyname: clayg: torgomatic: when reviewing the servers_per_port patch, this PR code branch for vagrant-swift-all-in-one should prove useful: https://github.com/swiftstack/vagrant-swift-all-in-one/pull/16/ | 14:45 |
| *** minwoob has joined #openstack-swift | 14:47 | |
| *** annegentle has joined #openstack-swift | 14:52 | |
| *** petertr7 is now known as petertr7_away | 15:00 | |
| *** petertr7_away is now known as petertr7 | 15:02 | |
| swifterdarrell | acoles_away: cschwede: mattoliverau: tdasilva: notmyname: clayg: torgomatic: I'll be rebasing shortly to make sure no probe test regressions have snuck in on master | 15:05 |
| *** vinsh has joined #openstack-swift | 15:20 | |
| *** breitz has joined #openstack-swift | 15:28 | |
| *** wbhuber has joined #openstack-swift | 15:28 | |
| *** leopoldj has quit IRC | 15:31 | |
| *** aix has quit IRC | 15:36 | |
| *** gyee_ has joined #openstack-swift | 15:36 | |
| openstackgerrit | Darrell Bishop proposed openstack/swift: Allow 1+ object-servers-per-disk deployment https://review.openstack.org/184189 | 15:50 |
| *** petertr7 is now known as petertr7_away | 15:51 | |
| swifterdarrell | ^^^^^^^^^^^ that rebase | 15:58 |
| *** jrichli has quit IRC | 16:00 | |
| *** lastops has quit IRC | 16:02 | |
| *** annegentle has quit IRC | 16:06 | |
| *** B4rker has quit IRC | 16:06 | |
| *** jistr has quit IRC | 16:10 | |
| *** vinsh has quit IRC | 16:12 | |
| *** ptb has quit IRC | 16:13 | |
| *** lcurtis has joined #openstack-swift | 16:14 | |
| *** vinsh has joined #openstack-swift | 16:16 | |
| *** B4rker has joined #openstack-swift | 16:17 | |
| *** nadeem has joined #openstack-swift | 16:21 | |
| *** acampbel11 has quit IRC | 16:22 | |
| *** lastops has joined #openstack-swift | 16:25 | |
| swifterdarrell | acoles_away: cschwede: mattoliverau: tdasilva: notmyname: clayg: torgomatic: I've confirmed that the rebased patch-set 13 passes probe tests for both SERVERS_PER_PORT=0 and SERVERS_PER_PORT=3 (in localrc with vagrant-swift-all-in-one) | 16:34 |
| *** jordanP has quit IRC | 16:36 | |
| *** petertr7_away is now known as petertr7 | 16:46 | |
| *** mmcardle has quit IRC | 16:48 | |
| notmyname | good morning | 16:56 |
| notmyname | swifterdarrell: thanks | 16:56 |
| notmyname | I added a "needs-new-api" tag to a LP bug that can't be fixed unless we bump the API version. if you see other bugs like that, please use that tag | 16:59 |
| *** annegentle has joined #openstack-swift | 17:06 | |
| *** annegentle has quit IRC | 17:12 | |
| *** annegentle has joined #openstack-swift | 17:20 | |
| *** geaaru has quit IRC | 17:25 | |
| lcurtis | hello all...is there a way i can see total object count via swift-recon or something similar? | 17:27 |
| *** eranrom has joined #openstack-swift | 17:28 | |
| notmyname | lcurtis: no. there is no total object count tracked. instead, each account has a number. the total is the sum of what's in each account | 17:29 |
| lcurtis | thank you notmyname. what would be best method to get that number from each account? | 17:29 |
| notmyname | there are 2 methonds | 17:30 |
| *** jodah has joined #openstack-swift | 17:30 | |
| notmyname | one is to use your identity (auth) system to give you a list of all the accounts and then do a HEAD to each of them | 17:30 |
| notmyname | but I think the better way is to have a process that walks the drives and aggregates the numbers | 17:30 |
| lcurtis | notmyname excellent..will start there | 17:31 |
| notmyname | https://github.com/notmyname/slogging is one tool that can do that | 17:31 |
| lcurtis | wow..great | 17:32 |
| lcurtis | thank you as usual for your very kind assistance | 17:32 |
| *** zhill has joined #openstack-swift | 17:33 | |
| *** proteusguy has quit IRC | 17:37 | |
| *** proteusguy has joined #openstack-swift | 17:42 | |
| *** annegentle has quit IRC | 17:55 | |
| *** annegentle has joined #openstack-swift | 17:56 | |
| *** nadeem has quit IRC | 17:57 | |
| *** gyee_ has quit IRC | 17:57 | |
| *** zhill has quit IRC | 18:00 | |
| *** dontalton has joined #openstack-swift | 18:00 | |
| *** mahatic has quit IRC | 18:08 | |
| *** openstackgerrit has quit IRC | 18:09 | |
| *** openstackgerrit has joined #openstack-swift | 18:09 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Prompt for missing username and/or password https://review.openstack.org/189387 | 18:09 |
| *** bi_fa_fu has quit IRC | 18:12 | |
| *** B4rker has quit IRC | 18:17 | |
| openstackgerrit | Andreas Jaeger proposed openstack/python-swiftclient: Fix inconsistent usage of "Positional argument" https://review.openstack.org/189390 | 18:22 |
| *** proteusguy has quit IRC | 18:22 | |
| *** bi_fa_fu has joined #openstack-swift | 18:23 | |
| *** proteusguy has joined #openstack-swift | 18:25 | |
| *** bi_fa_fu has quit IRC | 18:25 | |
| *** bi_fa_fu has joined #openstack-swift | 18:26 | |
| *** mahatic has joined #openstack-swift | 18:28 | |
| *** acampbell has quit IRC | 18:30 | |
| *** B4rker has joined #openstack-swift | 18:32 | |
| *** thurloat is now known as thurloat_isgone | 18:37 | |
| notmyname | ok, I gotta find out (and fix?) whatever's going on with the gate for swift | 18:37 |
| *** B4rker has quit IRC | 18:37 | |
| *** acampbell has joined #openstack-swift | 18:46 | |
| *** acampbel11 has joined #openstack-swift | 18:46 | |
| *** acampbel11 has left #openstack-swift | 18:48 | |
| notmyname | hmm | 18:51 |
| notmyname | so a virtualenv running the unit tests didn't work | 18:51 |
| notmyname | or, rather, it worked. didn't repro the issue | 18:51 |
| *** mahatic has quit IRC | 18:51 | |
| notmyname | issue being http://logs.openstack.org/89/184189/13/check/gate-swift-python27/95a00d9/console.html#_2015-06-08_15_57_18_049 | 18:51 |
| notmyname | swifterdarrell: didn't you say it was something about the differences between pypi and source versions of pyeclib? | 18:52 |
| *** chrisnelson_ is now known as chrisnelson | 18:56 | |
| *** proteusguy has quit IRC | 18:57 | |
| *** eranrom has quit IRC | 18:57 | |
| *** eranrom has joined #openstack-swift | 18:58 | |
| *** annegentle has quit IRC | 19:06 | |
| *** annegentle has joined #openstack-swift | 19:06 | |
| *** acampbell has quit IRC | 19:08 | |
| *** BAKfr_ has joined #openstack-swift | 19:15 | |
| *** bi_fa_fu has quit IRC | 19:15 | |
| *** BAKfr has quit IRC | 19:15 | |
| *** BAKfr_ is now known as BAKfr | 19:15 | |
| *** acampbell has joined #openstack-swift | 19:23 | |
| *** eranrom has left #openstack-swift | 19:25 | |
| *** mahatic has joined #openstack-swift | 19:33 | |
| *** charlesw has joined #openstack-swift | 19:35 | |
| *** acampbell has quit IRC | 19:38 | |
| *** mahatic has quit IRC | 19:52 | |
| *** thurloat_isgone is now known as thurloat | 19:53 | |
| *** lpabon has quit IRC | 20:20 | |
| notmyname | doesn't look like jerasure is installed on the test nodes | 20:23 |
| notmyname | fungi is working on that now | 20:23 |
| notmyname | but, it's still not clear why that issue only seems to have started happening on june 3. why didn't it happen before? | 20:23 |
| notmyname | torgomatic: once the jenkins workers get their dependencies sorted, assuming that fixes it, we could use your CI-plunger :-) | 20:24 |
| *** thurloat is now known as thurloat_isgone | 20:28 | |
| *** SkyRocknRoll has quit IRC | 20:28 | |
| torgomatic | notmyname: you've seen my EventletPlungerString, yes? :D | 20:35 |
| *** annegentle has quit IRC | 20:44 | |
| *** breitz has quit IRC | 20:49 | |
| *** breitz has joined #openstack-swift | 20:52 | |
| *** bi_fa_fu has joined #openstack-swift | 20:55 | |
| *** vinsh has quit IRC | 20:55 | |
| *** bi_fa_fu has quit IRC | 21:01 | |
| *** petertr7 is now known as petertr7_away | 21:11 | |
| *** annegentle has joined #openstack-swift | 21:11 | |
| *** ptb has joined #openstack-swift | 21:18 | |
| *** blmartin has quit IRC | 21:18 | |
| *** gyee_ has joined #openstack-swift | 21:22 | |
| *** ptb has quit IRC | 21:26 | |
| notmyname | py27 tests that are failing should be fixed by https://review.openstack.org/#/c/189472/ | 21:30 |
| notmyname | once that lands, the worker images will need to be rebuilt and we should be good | 21:30 |
| notmyname | I'm told images are rebuilt at 1414UTC every day | 21:31 |
| swifterdarrell | notmyname: I hope all the gate servers run Trusty... | 21:32 |
| notmyname | yeah, I think they do | 21:32 |
| * tdasilva is wondering if someone is going to try writing a swift object server in swift | 21:34 | |
| swifterdarrell | tdasilva: haha! ya, that joke went around the office here earlier, too ;) | 21:35 |
| wbhuber | notmyname peluse: implementing and configuring the cluster here is going well. the specs insofar are 3 proxies, 3 meta, and 5 storage nodes (34 HDD's each). | 21:35 |
| notmyname | wbhuber: nice | 21:35 |
| swifterdarrell | wbhuber: 10G networking? | 21:35 |
| wbhuber | we are planning to feed the data points into COSBench VMs sometime this week to test the performance benchmarking. | 21:35 |
| wbhuber | yes, 10G. | 21:35 |
| wbhuber | we are also going to try the differnt config from yours that you sent to us a couple of weeks ago. | 21:36 |
| swifterdarrell | wbhuber: in my recent testing, I had trouble w/the network saturating before I could get full disk utilization with medium-to-large obj sizes, and I ended up bumping my proxy count == storage node count to at least get parity | 21:36 |
| wbhuber | 3 DFs + 2 parity using jerasure_rs_vand | 21:36 |
| *** mwheckmann has quit IRC | 21:37 | |
| swifterdarrell | wbhuber: (in my case I had 29 and 32 disks in my 2 storage nodes and a single 10G NIC--I considered bonding 2 NICs but never got around to it) | 21:37 |
| wbhuber | notmyname: what's the benefit of bonding 2 NICs? load balancing? | 21:38 |
| swifterdarrell | wbhuber: in any case, if you're going to do any large-object loads and want to test anything other than just the line-rate of your NICs, you may want bonded 10G on the storage nodes and more proxie | 21:38 |
| swifterdarrell | *proxies | 21:38 |
| swifterdarrell | wbhuber: I think I've seen bonding set up for load sharing (i.e. kind of like a 20G NIC), not just failover | 21:38 |
| wbhuber | swifterdarrel: ok. | 21:38 |
| swifterdarrell | wbhuber: (not sure how well that works, but it seems like a thing networky people do) | 21:39 |
| swifterdarrell | wbhuber: probably requires some cooperation from the switch? | 21:39 |
| wbhuber | swifterdarrell: by large objects, did you mean ec_large_test.scenario from this link: https://github.com/swiftstack/ssbench/pull/107/files | 21:39 |
| wbhuber | swifterdarrell: i probably have to look more into the interaction with the switch. | 21:40 |
| swifterdarrell | wbhuber: I was testing servers_per_port stuff with just normal 3-replica ring; I think the network saturated somewhere between 5 MB and 100 MB obj size | 21:40 |
| swifterdarrell | wbhuber: I think 5 MB obj didn't saturate it, but 100 MB did? | 21:41 |
| *** EmilienM is now known as EmilienM|afk | 21:41 | |
| swifterdarrell | wbhuber: I was specifically trying to get at how object-server behaves when its disks are heavily loaded, and if the network gives out before the disks do, that's hard.... so that was my problem | 21:42 |
| wbhuber | swifterdarrell: i'd have to see if that affects our end as well. | 21:42 |
| swifterdarrell | wbhuber: with EC you might get CPU to give out before network or disks? exciting! | 21:42 |
| swifterdarrell | wbhuber: anyway, just an FYI from my recent benchmarking experiments | 21:42 |
| swifterdarrell | wbhuber: watch your network utilization and take it into account, if it saturates, when you interpret the numbers | 21:43 |
| wbhuber | swifterdarrell: most definitely. i believe COSBench has the capability to store the network utilization data while the objects of different magnitudes are being stored. | 21:44 |
| swifterdarrell | wbhuber: sweet! | 21:44 |
| *** charlesw has quit IRC | 21:46 | |
| wbhuber | swifterdarrell: sorry if i wasn't clear before, what did your storage policy look like? have you tried different combos (i.e. different ec_type and k + m's)? | 21:47 |
| *** ptb has joined #openstack-swift | 21:47 | |
| swifterdarrell | wbhuber: I was benchmarking "normal" 3-replica storage policy | 21:47 |
| swifterdarrell | wbhuber: I have not been doing any EC benchmarking | 21:48 |
| wbhuber | swifterdarrell: roger. | 21:48 |
| notmyname | wbhuber: the only ec benchmarking I've done was a 4+6 scheme with 20 drives across 5 servers (the results you've seen | 21:50 |
| notmyname | I haven't varied any of the parameters | 21:50 |
| notmyname | also, I was using the ISA-L library instead of jerasure | 21:51 |
| torgomatic | notmyname: so I tried that experiment, and I think I don't understand flat XOR erasure codes enough, because when I tried feeding it flat_xor_hd_3 in lieu of jerasure_rs_vand, I got that same error | 21:53 |
| notmyname | the missing jerasure error? | 21:53 |
| torgomatic | notmyname: no, the invalid-argument error | 21:54 |
| torgomatic | notably, "flat_xor_hd3" produced an error telling me I had an invalid coding scheme, and listing the valid ones | 21:54 |
| notmyname | hmm | 21:54 |
| notmyname | torgomatic: do you see http://logs.openstack.org/89/184189/13/check/gate-swift-python27/95a00d9/console.html#_2015-06-08_15_56_06_814 ? | 21:55 |
| torgomatic | notmyname: no | 21:55 |
| notmyname | only http://logs.openstack.org/89/184189/13/check/gate-swift-python27/95a00d9/console.html#_2015-06-08_15_57_18_049 ? | 21:55 |
| *** zhill_desktop has joined #openstack-swift | 21:58 | |
| torgomatic | notmyname: the experiment I was referencing was the one where I swap out jerasure_rs_vand for a flat XOR code, and all I get is that second error there | 21:59 |
| notmyname | yeah, we're on the same page :-) | 21:59 |
| notmyname | so IOW flat xor isn't a drop-in replacement for rs? at least from a param perspective | 22:00 |
| * notmyname in meeting | 22:00 | |
| *** ptb has quit IRC | 22:02 | |
| wbhuber | torgomatic: guess we need to open a bug to handle the error when swapping out jerasure_rs_vand for flat XOR | 22:02 |
| torgomatic | wbhuber: not necessarily... it could be that flat XOR codes only handle certain numbers of data/parity fragments, and I fed in something that was outside what the algorithm could do | 22:03 |
| * torgomatic knows nothing of flat XOR codes | 22:03 | |
| *** mwheckmann has joined #openstack-swift | 22:06 | |
| wbhuber | torgomatic: should be good to know the certain numbers flat XOR codes only accept. should be recorded somewhere. | 22:09 |
| *** annegentle has quit IRC | 22:10 | |
| mattoliverau | Morning, long weekend over, so am back, sarcasm and all | 22:11 |
| *** annegentle has joined #openstack-swift | 22:15 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: Add six requirement https://review.openstack.org/189495 | 22:23 |
| *** minwoob has quit IRC | 22:35 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace xrange() with six.moves.xrange() https://review.openstack.org/185460 | 22:41 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Replace dict.iteritems() with six.iteritems(dict) https://review.openstack.org/185451 | 22:46 |
| *** wbhuber has quit IRC | 22:50 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: Fix Python 3 issues https://review.openstack.org/185450 | 23:01 |
| *** annegentle has quit IRC | 23:02 | |
| *** km has joined #openstack-swift | 23:03 | |
| *** jkugel has quit IRC | 23:15 | |
| notmyname | mattoliverau: yay, matt's back | 23:18 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Use six to fix imports on Python 3 https://review.openstack.org/185453 | 23:19 |
| *** peluse has joined #openstack-swift | 23:19 | |
| *** ChanServ sets mode: +v peluse | 23:19 | |
| mattoliverau | notmyname: yup, looks like I have a bazillion bug emails, youve done a great auditing job! | 23:19 |
| *** kei_yama has joined #openstack-swift | 23:20 | |
| *** jkugel has joined #openstack-swift | 23:20 | |
| *** EmilienM|afk is now known as EmilienM | 23:20 | |
| notmyname | not done yet ;-) | 23:21 |
| notmyname | first pass I've done is to simple get things out of a "new" state | 23:21 |
| notmyname | after that, i'll work on better tagging and actual prioritization | 23:21 |
| notmyname | but I hope that the fasted win for the community can be "bug in 'new'? let's look at it" | 23:21 |
| *** annegentle has joined #openstack-swift | 23:22 | |
| *** mwheckmann has quit IRC | 23:27 | |
| notmyname | acoles_away: clayg: the other day i got off the bus, looked down at my backback, and I realized that the pin from vancouver had fallen off. might have gotten pulled of in a crowd or on the train or bus. don't know | 23:29 |
| *** asettle is now known as asettle-afk | 23:33 | |
| *** lcurtis has quit IRC | 23:35 | |
| *** chlong has joined #openstack-swift | 23:36 | |
| *** annegentle has quit IRC | 23:37 | |
| *** dontalton has quit IRC | 23:45 | |
| *** annegentle has joined #openstack-swift | 23:51 | |
| *** annegentle has quit IRC | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!