| notmyname | I just registered for the Tokyo summit | 00:04 |
|---|---|---|
| *** sorrison_laptop has joined #openstack-swift | 00:06 | |
| *** asettle has quit IRC | 00:07 | |
| *** kevinc_ has joined #openstack-swift | 00:13 | |
| notmyname | UPDATE status SET random("Invalid", "Won't Fix) WHERE id = (SELECT * from launchpad_bugs WHERE date < 2014 AND status = "New"); | 00:19 |
| portante | notmyname: cute ... eh ... serious? | 00:34 |
| portante | missing a double quote after "Fix" | 00:34 |
| *** lyrrad has quit IRC | 00:39 | |
| notmyname | portante: ehh...just being cute. and currently going through old bugs. mostly pseudo-SQL, not meant to be real :-) | 00:41 |
| portante | ;) | 00:42 |
| notmyname | but most of them end up being either invalid (now) or won't fix | 00:42 |
| portante | so fedora closes most open bugs, slam door shut closed, after two releases (all bugs opened against F20 get closed once F22 goes), perhaps a good policy for swift? | 00:43 |
| notmyname | portante: like this one https://bugs.launchpad.net/swift/+bug/1273961 | 00:44 |
| openstack | Launchpad bug 1273961 in OpenStack Object Storage (swift) "Unhandled exception in request" [Undecided,Invalid] | 00:44 |
| notmyname | so sometime well over a year ago there was an unhandled exception. not much I can do on it now :-) | 00:44 |
| notmyname | except change the status in order to beat the list of bugs down and make the list somewhat useful (I hope!) | 00:45 |
| torgomatic | especially since the guy says he added some code to Swift, and then it blew the stack; while I sympathize, perhaps that's not worthy of a bug report, you know? | 00:45 |
| torgomatic | "Swift plus some stuff I typed in doesn't work" | 00:46 |
| portante | makes total sense | 00:46 |
| torgomatic | me neither, but you don't see bug reports about it ;) | 00:46 |
| notmyname | torgomatic: to be fair, most of the stuff I type in for Swift doesn't work either ;-) | 00:46 |
| portante | is there some bean counter that will cry foul? | 00:46 |
| notmyname | portante: I'm not too worried about the bean counters. only that we have a tool as a community that is useful for something | 00:47 |
| portante | yes! | 00:47 |
| portante | preach it! | 00:47 |
| notmyname | (but my launchpad karma is awesome!) | 00:47 |
| notmyname | (so I'm winning) | 00:47 |
| portante | ;) | 00:47 |
| portante | its the small battles that count | 00:47 |
| notmyname | I really don't like the categories LP gives | 00:48 |
| notmyname | I'd like something like "this is not a bug, but also not invalid" | 00:49 |
| notmyname | but it's not a bug so I can't "won't fix" it because there's nothing to fix | 00:49 |
| notmyname | nor is it in-progress | 00:50 |
| notmyname | the description on the "opinion" status is "doesn't fit with the project but can be discussed" | 00:51 |
| notmyname | but why would you discuss it if it doesn't fit with the project? | 00:53 |
| *** dmorita has joined #openstack-swift | 00:53 | |
| notmyname | ok, I'm out for now | 00:53 |
| portante | notmyname: it seems lots of folks want to create a tool that simplifies the world so that they feel better about the world | 00:53 |
| *** marzif_ has joined #openstack-swift | 01:01 | |
| *** gyee has quit IRC | 01:02 | |
| *** nexusz99 has joined #openstack-swift | 01:12 | |
| *** kota_ has joined #openstack-swift | 01:13 | |
| *** ChanServ sets mode: +v kota_ | 01:13 | |
| *** marzif_ has quit IRC | 01:14 | |
| *** chlong-afk has joined #openstack-swift | 01:15 | |
| openstackgerrit | Merged openstack/swift: Cleanup error messages in ECDiskFileWriter.commit https://review.openstack.org/199362 | 01:15 |
| kota_ | good morning | 01:15 |
| *** ho has joined #openstack-swift | 01:16 | |
| *** sorrison_laptop has quit IRC | 01:22 | |
| *** kevinc_ has quit IRC | 01:28 | |
| mattoliverau | kota_: morning | 02:04 |
| kota_ | mattoliverau: good morning and I found you pushed nice patch, https://review.openstack.org/#/c/199863 a few days ago. I like this stuff :) | 02:05 |
| kota_ | nice work | 02:06 |
| mattoliverau | kota_: thanks, that was because I needed a break from sharding and wanted to knock something off the OPs feedback list :) | 02:07 |
| kota_ | i see | 02:08 |
| hugokuo | handoff zone affinity support ! | 02:08 |
| hugokuo | mornign | 02:09 |
| kota_ | hugokuo: morning :) Are you interested in the patch too? | 02:11 |
| hugokuo | kota_: Indeed. That's make sense for multiple region setup. | 02:12 |
| kota_ | hugokuo: nice ;) | 02:12 |
| *** mutoulbj has joined #openstack-swift | 02:17 | |
| *** bkopilov has quit IRC | 02:27 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift-specs: Update sharding spec, notably regarding CountingTrie https://review.openstack.org/201000 | 02:44 |
| *** mgriffin has joined #openstack-swift | 02:45 | |
| *** DericHorn-HP has joined #openstack-swift | 03:00 | |
| *** DericHorn-HP has quit IRC | 03:01 | |
| *** minwoob has joined #openstack-swift | 03:04 | |
| *** nexusz99 has quit IRC | 03:30 | |
| *** lyrrad has joined #openstack-swift | 03:41 | |
| *** hrou has quit IRC | 03:46 | |
| *** breitz has quit IRC | 03:53 | |
| *** breitz has joined #openstack-swift | 03:56 | |
| *** sanchitmalhotra has joined #openstack-swift | 04:00 | |
| *** minwoob has quit IRC | 04:05 | |
| *** bkopilov has joined #openstack-swift | 04:07 | |
| *** nexusz99 has joined #openstack-swift | 04:20 | |
| *** ppai has joined #openstack-swift | 04:30 | |
| openstackgerrit | Pradeep Kumar Singh proposed openstack/swift: Emit warning log in object replicator https://review.openstack.org/185061 | 04:34 |
| *** proteusguy_ has quit IRC | 04:44 | |
| *** kota_ has quit IRC | 05:02 | |
| *** nexusz99 has quit IRC | 05:07 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 05:21 | |
| *** sanchitmalhotra has quit IRC | 05:23 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:33 | |
| *** zaitcev has quit IRC | 05:53 | |
| *** SkyRocknRoll has quit IRC | 05:53 | |
| *** kota_ has joined #openstack-swift | 05:54 | |
| *** ChanServ sets mode: +v kota_ | 05:54 | |
| *** chlong-afk has quit IRC | 06:09 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:11 | |
| *** nexusz99 has joined #openstack-swift | 06:33 | |
| *** sanchitmalhotra has joined #openstack-swift | 06:42 | |
| *** psanchez_ has joined #openstack-swift | 06:44 | |
| *** sanchitmalhotra1 has quit IRC | 06:45 | |
| *** _fortis has joined #openstack-swift | 07:02 | |
| *** mutoulbj has quit IRC | 07:14 | |
| *** mutoulbj has joined #openstack-swift | 07:14 | |
| *** _fortis has quit IRC | 07:17 | |
| *** hseipp has joined #openstack-swift | 07:19 | |
| *** nexusz99 has quit IRC | 07:21 | |
| *** nexusz99 has joined #openstack-swift | 07:25 | |
| *** _fortis has joined #openstack-swift | 07:30 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Make logic of unit tests responsive to the method names https://review.openstack.org/201442 | 07:32 |
| *** chlong-afk has joined #openstack-swift | 07:37 | |
| *** jistr has joined #openstack-swift | 07:42 | |
| hugokuo | 40 ffff8808715b4540 ffff8804819c58c0 ffff8802381f8d88 DIR /srv/node/d224/quarantined/objects/b146865bf8034bfc42570b747c341b32 | 07:55 |
| hugokuo | anyone knows why object-server try to open file in quarantined folder ? | 07:56 |
| *** joeljwright has joined #openstack-swift | 08:37 | |
| *** ChanServ sets mode: +v joeljwright | 08:37 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Improve Keystone v3 token support https://review.openstack.org/201461 | 08:44 |
| *** aix has joined #openstack-swift | 08:51 | |
| *** aix has quit IRC | 08:58 | |
| *** aix has joined #openstack-swift | 09:03 | |
| *** dmorita has quit IRC | 09:04 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 09:15 | |
| *** sanchitmalhotra has quit IRC | 09:17 | |
| *** sanchitmalhotra has joined #openstack-swift | 09:26 | |
| *** sanchitmalhotra1 has quit IRC | 09:28 | |
| *** marzif_ has joined #openstack-swift | 09:33 | |
| *** mahatic has joined #openstack-swift | 09:41 | |
| *** nexusz99_ has joined #openstack-swift | 09:50 | |
| *** nexusz99 has quit IRC | 09:53 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 10:10 | |
| *** sanchitmalhotra has quit IRC | 10:12 | |
| *** aix has quit IRC | 10:21 | |
| *** kota_ has quit IRC | 10:34 | |
| *** mutoulbj has quit IRC | 10:34 | |
| *** aix has joined #openstack-swift | 10:34 | |
| *** sanchitmalhotra has joined #openstack-swift | 10:40 | |
| *** sanchitmalhotra1 has quit IRC | 10:41 | |
| *** psanchez_ has quit IRC | 11:01 | |
| *** openstackgerrit has quit IRC | 11:09 | |
| *** openstackgerrit has joined #openstack-swift | 11:09 | |
| *** mutoulbj has joined #openstack-swift | 11:15 | |
| *** psanchez_ has joined #openstack-swift | 11:17 | |
| *** ho has quit IRC | 11:21 | |
| *** mahatic has quit IRC | 11:25 | |
| *** kei_yama has quit IRC | 11:28 | |
| *** mahatic has joined #openstack-swift | 11:53 | |
| *** openstack has joined #openstack-swift | 12:13 | |
| *** ig0r_ has joined #openstack-swift | 12:23 | |
| *** annegent_ has quit IRC | 12:24 | |
| *** SkyRocknRoll has quit IRC | 12:24 | |
| *** annegentle has joined #openstack-swift | 12:25 | |
| *** nexusz99_ has quit IRC | 12:30 | |
| *** bkopilov has quit IRC | 12:32 | |
| *** proteusguy has joined #openstack-swift | 12:42 | |
| *** robefran has joined #openstack-swift | 12:49 | |
| *** hrou has joined #openstack-swift | 13:00 | |
| *** petertr7_away is now known as petertr7 | 13:10 | |
| *** ppai has quit IRC | 13:11 | |
| *** km has quit IRC | 13:15 | |
| *** joeljwright has quit IRC | 13:17 | |
| *** thurloat_isgone is now known as thurloat | 13:22 | |
| *** petertr7 is now known as petertr7_away | 13:25 | |
| *** sweeper has joined #openstack-swift | 13:26 | |
| sweeper | morning folks | 13:26 |
| sweeper | I've got a swift cluster that lost 2 out of 40 drives yesterday, and it's being very very strange | 13:27 |
| sweeper | load average has shot up across the board | 13:27 |
| sweeper | we haven't replaced the drives yet | 13:27 |
| sweeper | but we haven't kicked off a rebalance either | 13:28 |
| ctennis | what process is driving the high load, the replicator? | 13:31 |
| sweeper | it looks like it's the object server | 13:31 |
| sweeper | disk usage has gone way up as well | 13:32 |
| sweeper | so it may be part and parcel | 13:32 |
| ctennis | at first glance it sounds normal, things are replicating to get whole again | 13:32 |
| *** marzif_ has quit IRC | 13:33 | |
| sweeper | usage averages do seem to be slightly up... | 13:33 |
| *** marzif_ has joined #openstack-swift | 13:33 | |
| ctennis | the swift logs may show what's going on behind the scenes, depending on hwere you're sending them. We put them in /var/log/swift/all.log, you can check the progress of the replicator | 13:34 |
| *** petertr7_away is now known as petertr7 | 13:34 | |
| sweeper | herm | 13:35 |
| *** mahatic has quit IRC | 13:35 | |
| sweeper | it seems to be trying to replicate into unmounted directories.. | 13:36 |
| ctennis | do you have the mount check enabled? | 13:36 |
| sweeper | yea | 13:36 |
| sweeper | there's a bug in that fwiw :P | 13:36 |
| ctennis | I've seen a bug there with respect to rsync yeah | 13:36 |
| *** SkyRocknRoll has joined #openstack-swift | 13:38 | |
| sweeper | it's weird. the object replicator reports the drive as unmounted, but persists in trying to rsync | 13:41 |
| *** mahatic has joined #openstack-swift | 13:48 | |
| *** NM has joined #openstack-swift | 13:48 | |
| sweeper | any idea why swift-recon would show only one unmounted drive, but object-replicator reports 3? | 13:55 |
| wbhuber | notmyname: yes, i am still working on the bug. i have a couple of Q's for clayg related to the probe test changes he put in for us to test. | 13:55 |
| *** doxavore has joined #openstack-swift | 13:57 | |
| *** petertr7 is now known as petertr7_away | 13:59 | |
| *** jrichli has joined #openstack-swift | 13:59 | |
| *** annegentle has quit IRC | 14:00 | |
| *** joeljwright has joined #openstack-swift | 14:00 | |
| *** ChanServ sets mode: +v joeljwright | 14:00 | |
| *** petertr7_away is now known as petertr7 | 14:00 | |
| *** tmcpeak has joined #openstack-swift | 14:03 | |
| tmcpeak | hey all - I have a question about the Swift deployment guide, is this an appropriate place to ask? Specifically I'd like to find out why this setting is recommended: net.ipv4.tcp_syncookies = 0 . This has the potential to have a negative impact on security, so I'd like to understand if it's required | 14:04 |
| *** lcurtis has joined #openstack-swift | 14:08 | |
| *** bkopilov has joined #openstack-swift | 14:12 | |
| *** jlhinson has joined #openstack-swift | 14:16 | |
| *** pbourke has quit IRC | 14:22 | |
| *** pbourke has joined #openstack-swift | 14:22 | |
| *** psanchez_ has quit IRC | 14:24 | |
| peluse | tmcpeak, this is a good place to ask - I'm not the right one to answer though :( I suspect folks that can answer will be around sometime soon though (few hours) | 14:24 |
| *** minwoob has joined #openstack-swift | 14:26 | |
| *** samueldmq has quit IRC | 14:26 | |
| *** agentle_ has joined #openstack-swift | 14:28 | |
| ctennis | tmcpeak: it's just a preference. most people aren't building publicly faceable swift clouds so the need for syn cookies isn't as important | 14:29 |
| *** thurloat is now known as thurloat_isgone | 14:30 | |
| tmcpeak | ctennis: sure, that makes sense. Is there a reason why disabling it is preferable though? | 14:30 |
| tmcpeak | seems like even internally having DoS protection is a good thing, if it isn't harming anything may as well leave it in? | 14:31 |
| ctennis | well it only protects against specifically crafted DoS attacks that likely wouldn't originate from your internal network | 14:32 |
| ctennis | my best guess is that it's more performant without syn cookies on, and given the amount of TCP requests that happen in swift it's just a way of eeking out a bit of a performance gain | 14:32 |
| ctennis | it's not harming anything by leaving it on though | 14:33 |
| tmcpeak | ctennis: ok, thank you | 14:36 |
| mahatic | ctennis, hello, I think peluse mentioned about the S3-tests yesterday. Would you happen to know what should be placed in "access_keys" and "secret_keys"? in here https://github.com/ceph/s3-tests | 14:39 |
| mahatic | ctennis, are those just the login credentials? or something else? | 14:39 |
| ctennis | mahatic: yeah if you're using swiftstack auth or tempauth then access_key is just your username and secret_key is the password | 14:39 |
| ctennis | also mahatic and peluse, confirmed that it was the ceph/s3-tests suite we used on those results I gave you | 14:43 |
| *** tmcpeak has left #openstack-swift | 14:45 | |
| mahatic | ctennis, I'm not sure if I got that. Yes, I'm using the ceph/s3-tests currently | 14:45 |
| mahatic | ctennis, oh I guess, I missed the conv yesterday. Got it | 14:47 |
| mahatic | ctennis, I used the login credentials and my config file looks like this: http://paste.openstack.org/show/374535/ Could you let me know if I'm missing something there? | 14:48 |
| mahatic | ctennis, With that config, it gives me a connection refused when I run the tests (http://paste.openstack.org/show/374552/) | 14:48 |
| peluse | ctennis, she's jsut using an SAIO (not swiftstack installed) for the S3 tests | 14:48 |
| ctennis | mahatic: I don't think your error is credential related, more HTTPS when expecting HTTP related | 14:49 |
| peluse | mahatic, make sure you undo your env proxy settings for actualling using your SAIO | 14:49 |
| mahatic | ctennis, peluse hmm, I did remove the env proxy settings while trying that out | 14:51 |
| *** jonb2 has quit IRC | 14:53 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 14:57 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 14:58 |
| peluse | mahatic, not sure... maybe hugokuo or timburke have tried this setup and seen something similar? | 15:04 |
| *** ccavanna has quit IRC | 15:08 | |
| *** vinsh_ is now known as Vinsh | 15:09 | |
| *** ig0r__ has joined #openstack-swift | 15:10 | |
| *** ig0r_ has quit IRC | 15:10 | |
| hugokuo | mahatic: Paste 1) proxy-server.conf 2) "netstat -antulp" of the proxy server 3) use boto to talk to swift3 for testing to verify your Swift3 is setting properly. | 15:17 |
| *** ig0r_ has joined #openstack-swift | 15:17 | |
| *** petertr7 is now known as petertr7_away | 15:18 | |
| hugokuo | mahatic: I suggest that to use http on proxy server as the beginning. I agree with ctennis's speaking about https/http | 15:18 |
| *** ig0r__ has quit IRC | 15:19 | |
| mahatic | hugokuo, I did and 1) and 3) of what you mentioned. And not sure where to change the http/https? | 15:19 |
| *** mahatic_ has joined #openstack-swift | 15:21 | |
| hugokuo | mahatic: May I have the link of proxy-server.conf ? Does boto work for you? If yes, paste your boto script to pastebin please | 15:22 |
| hugokuo | mahatic: regarding to the http/https it's in /etc/swift/proxy-server.conf | 15:23 |
| hugokuo | mahatic: comment out these two lines https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L34-L36 | 15:24 |
| hugokuo | how's your ST_AUTH looks like | 15:24 |
| mahatic | hugokuo, http://paste.openstack.org/show/374626/ (proxy-server.conf) | 15:25 |
| hugokuo | mahatic: For any configuration changes, you need restart proxy by $swift-init proxy restart | 15:25 |
| mahatic | ah alright. Thanks for that info. Also I installed boto, but didn't specifically run a script, I thought the s3-tests took care of the conenction? I'm just running the s3-tests directly | 15:26 |
| hugokuo | mahatic: confirmed. Your proxy-server listen on 8080 port w/o SSL. But your s3-test suite configuration specifies to use SSL http://paste.openstack.org/show/374535/ L12 , is_secure | 15:28 |
| hugokuo | the value should be "no" | 15:28 |
| mahatic | hugokuo, oh okay, my bad! thanks much, I'll run the tests with that changed. Is there anything else I'm missing? | 15:31 |
| hugokuo | mahatic: dunno yet. You can try it and let's see how's the result looks like | 15:32 |
| mahatic | hugokuo, sure | 15:33 |
| hugokuo | mahatic: btw, Swift3 wont't pass all test cases of ceph/s3-test | 15:33 |
| hugokuo | There's portion of cases rely on Ceph's implementation. | 15:33 |
| mahatic | oh okay | 15:33 |
| *** petertr7_away is now known as petertr7 | 15:35 | |
| *** mahatic__ has joined #openstack-swift | 15:36 | |
| *** mahatic has quit IRC | 15:37 | |
| *** fthiagogv has joined #openstack-swift | 15:39 | |
| *** ig0r__ has joined #openstack-swift | 15:40 | |
| *** ig0r_ has quit IRC | 15:43 | |
| *** lyrrad has quit IRC | 15:45 | |
| petertr7 | Are logs for this channel saved online somewhere? | 15:46 |
| *** blmartin has joined #openstack-swift | 15:47 | |
| hrou | yep: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/ | 15:47 |
| petertr7 | Thank you! | 15:50 |
| *** marzif_ has quit IRC | 15:52 | |
| *** marzif_ has joined #openstack-swift | 15:52 | |
| robefran | Hi, I'm new here, and I'm interested in helping out with symlinks. Who should I contact? | 15:53 |
| mahatic__ | hugokuo, http://paste.openstack.org/show/374659/ that's how the output looks. Still a forbidden error | 15:54 |
| hugokuo | mahatic__: Perhpas you can read the log of proxy-server about why it's been denied | 15:56 |
| hugokuo | wrong API entry or incorrect credential | 15:57 |
| mahatic__ | yeah, I'm doing that. Just checking if you got something off the top of your mind :) | 15:57 |
| *** mahatic_ has quit IRC | 15:57 | |
| hugokuo | I ran s3-test about 1year ago. Can't remember all the details of the conf file tho | 15:57 |
| mahatic__ | hmm okay | 15:57 |
| charz | mahatic__: try `sudo S3TEST_CONF=mys3.conf ./virtualenv/bin/nosetests -v` | 15:58 |
| hugokuo | mahatic__: at least we figured out the first issue of http protocol :) | 15:59 |
| mahatic__ | hugokuo, yup :) | 15:59 |
| *** marzif_ has quit IRC | 15:59 | |
| hugokuo | charz: hi champion , had you conduct s3-test suite recently. I remember that you did that around 2 months ago ? | 15:59 |
| *** jlhinson has quit IRC | 16:01 | |
| *** jlhinson has joined #openstack-swift | 16:01 | |
| mahatic__ | charz, it tells me fuzz tests are ok. And the rest of the log gives error on functional tests (on __init__) like it's shown previously - forbidden | 16:03 |
| charz | hugokuo: I ran that a couple of months ago, I forgot. | 16:03 |
| charz | mahatic__: I think you need to look into the code of test, see how it work and also check the proxy's log. | 16:05 |
| mahatic__ | yeah | 16:06 |
| charz | mahatic__: not all tests will pass, because s3-tests is develop for ceph, that's some behaviour are different than swift3. | 16:06 |
| mahatic__ | charz, yeah. I'll dig into this issue, it showed at least show me it failed and probably not some error like this right? | 16:07 |
| mahatic__ | it should at least* | 16:08 |
| timburke | mahatic__: what all users are defined for tempauth? my understanding is that user_id and display_name should both be of the form <account>:<username>, while access_key is just <username> | 16:11 |
| timburke | mahatic__: (though i must confess, i've not actually run the tests against tempauth) | 16:11 |
| *** lyrrad has joined #openstack-swift | 16:11 | |
| charz | mahatic__: sometime swift3 return some errors code that's not s3-tests expect. So you will got some response 404 or 403 from swift3. | 16:13 |
| mahatic__ | timburke, oh okay. tempauth does not have my login credentials though (the ones I mentioned in S3 config) http://paste.openstack.org/show/374735/ | 16:18 |
| *** wbhuber has quit IRC | 16:25 | |
| *** thurloat_isgone is now known as thurloat | 16:34 | |
| *** thurloat is now known as thurloat_isgone | 16:38 | |
| *** robefran has quit IRC | 16:44 | |
| notmyname | good morning | 16:44 |
| *** ccavanna has joined #openstack-swift | 16:46 | |
| notmyname | joeljwright: did I catch you or is it already too late? | 16:49 |
| joeljwright | notmyname: I'm here | 16:51 |
| notmyname | hmm..so swob implements HTTPServerError and HTTPInternalServerError. both map to a 500 status code. why do we do the same thing differently? | 16:51 |
| notmyname | joeljwright: re swiftclient release | 16:52 |
| notmyname | joeljwright: I looked yesterday. I don't see anything that should block a swiftclient release. did you find anything | 16:52 |
| notmyname | joeljwright: if not, then I'll do that today | 16:52 |
| joeljwright | notmyname: I was planning on going through those tonight | 16:54 |
| *** hseipp has quit IRC | 16:54 | |
| joeljwright | but I'll have a look now and get back to you in ~30mins | 16:54 |
| notmyname | joeljwright: thanks | 16:54 |
| notmyname | joeljwright: I'm only looking for stuff that we can't release without instead of stuff that would be nice to have. there's a lot that would be nice to have | 16:55 |
| notmyname | (also, as part of prioritizing swiftclient better, I want to do more frequent releases) | 16:55 |
| *** aix has quit IRC | 16:59 | |
| *** mgarza_ has joined #openstack-swift | 17:00 | |
| *** robefran has joined #openstack-swift | 17:01 | |
| *** doxavore has quit IRC | 17:02 | |
| *** doxavore has joined #openstack-swift | 17:05 | |
| *** petertr7 is now known as petertr7_away | 17:06 | |
| notmyname | email subject line: "What is a distributed database & do you need one?" | 17:09 |
| joeljwright | notmyname: there doesn't seem to be anything waiting that should postpone the release | 17:09 |
| notmyname | if you're asking me what a distributed DB is, then don't also ask me if you need one ;-) | 17:09 |
| joeljwright | notmyname: but an increase in release frequency would be nice | 17:09 |
| notmyname | joeljwright: great, thanks! | 17:09 |
| notmyname | joeljwright: I'll write up some release notes for an announce email today and make the tag | 17:10 |
| joeljwright | :) | 17:10 |
| robefran | Is anyone looking into tiering? I'm starting an investigation of tiering between containers with different policies (using symlinks) | 17:13 |
| *** jistr has quit IRC | 17:14 | |
| peluse | robefran, there are multiple specs in review/discussion to implmenet tiering - I am on my way out for a few hours but if you look at our specs repo you will see them. There is also one that doesn't show up because of a -2 ; when I get back I can paste some links in if someone else hasn't already | 17:22 |
| robefran | peluse: I've checked out the NetApp tiering proposal, but it got a lukewarm response. From the symlink proposal it looks like tiering based on symlinks would get more traction. | 17:23 |
| peluse | robefran, yeah, we talked at the summit and none of the specs "as is" is really what we're after. Its a combination of a bunch of them and that considlation/summary work hasn't been done... | 17:26 |
| * peluse heads out for a few hrs... | 17:27 | |
| *** NM1 has joined #openstack-swift | 17:28 | |
| *** NM has quit IRC | 17:30 | |
| robefran | peluse: I'm interested in helping out with symlinks, and with moving the tiering proposals forward. | 17:35 |
| notmyname | robefran: that's great to hear! | 17:37 |
| robefran | notmyname: Who is driving the symlink proposal? Who should I contact? | 17:39 |
| notmyname | robefran: torgomatic first proposed it, and acoles_away and cschwede have also expressed interest. however, acoles_away is on vacation and torgomatic has been looking at a few other short-term things (ring improvements) | 17:41 |
| notmyname | robefran: what would you like (from me) in order to get involved? | 17:41 |
| *** NM1 has quit IRC | 17:42 | |
| robefran | notmyname: I see two parts to tiering, mechanics and decisions. I've been focusing on mechanics, and how to make sure that eventual consistency is maintained in the window between copying the hot-object to the cold-container, and creating the stub (symlink) in the hot container. | 17:47 |
| robefran | notmyname: is there a current spec for tiering that is moving forward? | 17:48 |
| notmyname | robefran: no. the only one I recall is the one from NetApp, which, as you said, had a "lukewarm response". like peluse mentioned, the mechanics are a combination of some different ideas, including symlinks, that haven't been gathered in one place | 17:50 |
| *** NM has joined #openstack-swift | 17:51 | |
| *** openstackgerrit has quit IRC | 17:56 | |
| *** openstackgerrit has joined #openstack-swift | 17:56 | |
| robefran | notmyname: I could draft a tiering proposal based on symlinks, but I'm new to the community, and I could use some guidance on how to get a proposal started. | 17:57 |
| openstackgerrit | Michael Barton proposed openstack/swift: go: fix expect 100-continue in replicator https://review.openstack.org/201690 | 17:57 |
| notmyname | robefran: specs are simply text documents describing an idea that should get input before you start coding. it's not really a full-on up-front design spec. it's more of a way to have a conversation about an idea while we're on opposite sides of the world | 18:00 |
| notmyname | robefran: but backing up a bit, are you familiar with the mechanics of contributing to any openstack project? | 18:01 |
| *** agentle_ has quit IRC | 18:02 | |
| *** agentle_ has joined #openstack-swift | 18:02 | |
| mahatic__ | hugokuo, charz , timburke I gave the SAIO test credentials instead of my login in S3 config and it worked | 18:04 |
| robefran | notmyname: this will be my first contribution to openstack. I've setup the SAIO, and my first step was to propose some additional functional tests for COPY-to-same-name. Then I planned to help out with the symlink implementation. | 18:04 |
| charz | mahatic__: oh, cool! | 18:05 |
| notmyname | robefran: great! you've already proposed a patch? which one is it? | 18:05 |
| robefran | notmyname: I haven't proposed the COPY-to-same-name functest patch yet, that was going to be my first step | 18:06 |
| notmyname | ah ok | 18:06 |
| notmyname | robefran: so you've got a SAIO. that's good. have you set up git-review and your credentials yet? | 18:06 |
| *** petertr7_away is now known as petertr7 | 18:07 | |
| timburke | mahatic__: oh good! you might want to tweak some of the middleware config options; turning on s3_acl and turning off dns_compliant_bucket_names help a decent bit. fwiw, the best test runs i've seen still have ~100 failures | 18:07 |
| robefran | notmyname: I've registered at UbuntuOne and OpenStack (id robefran@ca.ibm.com for both). I've already added a review comment on the symlink proposol. | 18:09 |
| notmyname | ok, that's good | 18:10 |
| robefran | notmyname: I'll get to work on the functest and see how that goes. An additional functest should be harmless. | 18:11 |
| notmyname | robefran: so now, with any openstack project you want to contribute a patch to, the process is (one time up front) do `git review -s` to set it up, then when you're on your branch with a single commit to propose, do `git review` | 18:11 |
| notmyname | robefran: great. that would be a really good place to start | 18:11 |
| notmyname | robefran: the only "odd" thing, IMO, with contributing to swift is that you need to propose your change as one patch, not a series of patches. that means you'll propose something, get feedback, update your local patch, and then have to do a `git commit --amend` before another `git review` | 18:12 |
| mahatic__ | timburke, thanks for me letting me know. w/o acl, this what it looks like - SKIP=4, errors=88, failures=73 | 18:12 |
| *** janonymous_ has joined #openstack-swift | 18:23 | |
| janonymous_ | Could someone please have a look at these patches and review these : https://review.openstack.org/#/q/status:open+project:openstack/swift+branch:master+topic:py3,n,z | 18:24 |
| janonymous_ | Thanks . | 18:24 |
| *** wasmum- has joined #openstack-swift | 18:24 | |
| *** distroguy has joined #openstack-swift | 18:25 | |
| *** SkyRocknRoll has quit IRC | 18:26 | |
| notmyname | robefran: and I just drew all that up on the whiteboard for amandap here in the office. seems she's working on improving upstread docs on the process :-) | 18:27 |
| notmyname | robefran: amandap: here's the swift contributing doc: https://github.com/openstack/swift/blob/master/CONTRIBUTING.md | 18:28 |
| *** doxavore_ has joined #openstack-swift | 18:28 | |
| *** mgagne_ has joined #openstack-swift | 18:29 | |
| *** kcaj_ has joined #openstack-swift | 18:30 | |
| *** doxavore has quit IRC | 18:30 | |
| *** blmartin has quit IRC | 18:30 | |
| *** chlong-afk has quit IRC | 18:30 | |
| *** alpha_ori has quit IRC | 18:30 | |
| *** Abyss_ has quit IRC | 18:30 | |
| *** csmart has quit IRC | 18:30 | |
| *** Protux has quit IRC | 18:30 | |
| *** kcaj has quit IRC | 18:30 | |
| *** zhiyan has quit IRC | 18:30 | |
| *** timburke has quit IRC | 18:30 | |
| *** wasmum has quit IRC | 18:30 | |
| *** mgagne has quit IRC | 18:30 | |
| *** sileht has quit IRC | 18:30 | |
| *** mrda-away has quit IRC | 18:30 | |
| *** timburke_ has joined #openstack-swift | 18:30 | |
| *** CaioBrentano has joined #openstack-swift | 18:31 | |
| *** alpha_ori has joined #openstack-swift | 18:31 | |
| *** mgagne_ has left #openstack-swift | 18:33 | |
| *** timburke_ is now known as timburke | 18:34 | |
| *** SkyRocknRoll has joined #openstack-swift | 18:40 | |
| *** mgagne_ has joined #openstack-swift | 18:41 | |
| *** kcaj_ has quit IRC | 18:47 | |
| *** kcaj has joined #openstack-swift | 18:49 | |
| *** mahatic__ has quit IRC | 18:51 | |
| *** AJS_ has quit IRC | 18:51 | |
| *** openstackgerrit has quit IRC | 18:56 | |
| *** openstackgerrit has joined #openstack-swift | 18:56 | |
| zackmdavis | janonymous_, (speaking of the Py3 crusade, I notice the Python34 gate jobs on those patches are failing on importing rfc822; I had some thoughts earlier about what we might have to do about that, but haven't gotten around to writing a patch (as torgomatic pointed out, we're relying on rfc822 functionality that doesn't really exist in the more modern 'email' package)) | 19:00 |
| *** Guest1999 has joined #openstack-swift | 19:01 | |
| Guest1999 | Question about ring file. Did anyone try to hack the ring file to change zone information of a device. | 19:02 |
| * torgomatic is interested to see what happens with the code using rfc822 | 19:02 | |
| torgomatic | Guest1999: you have to remove the device and then re-add it with its new zone | 19:03 |
| Guest1999 | and there is no other way I guess then. because the assignment is based on that value right? | 19:03 |
| torgomatic | Guest1999: yeah, if you just edited the builder to change the zone in-place, your dispersion could get screwed up | 19:04 |
| *** zaitcev has joined #openstack-swift | 19:04 | |
| *** ChanServ sets mode: +v zaitcev | 19:04 | |
| torgomatic | like if a partition has replicas in zones 1, 2, and 4, but then you change a disk's zone and now it's 1, 4, 4 | 19:04 |
| torgomatic | so that's why | 19:04 |
| Guest1999 | Thank torgomatic got it thats what I thought. Also did you ever experience high CPU usage with affinity. I did an experiment with large parallel requests with small objects and affinity for read and write . In both cases the CPU get a very high utilization than non-affinity runs | 19:05 |
| Guest1999 | i am referring to proxy | 19:05 |
| torgomatic | Guest1999: nope, I haven't observed that behavior | 19:05 |
| Guest1999 | is this something that should happen with affinity i.e. CPU increase. | 19:06 |
| *** mutoulbj has quit IRC | 19:07 | |
| janonymous_ | torgomatic : please tell me if i could do something to help | 19:07 |
| torgomatic | Guest1999: not to my knowledge | 19:07 |
| torgomatic | janonymous_: figure out if there's something in py3 stdlib that can do what `rfc822.Message(fp, 0)` does :) | 19:08 |
| Guest1999 | ok this is very strange behaviour . I ran profiler to check where the time is spent. most was in ring.py (get_more_nodes).... in case of affinity | 19:09 |
| janonymous_ | I'll search :) | 19:09 |
| torgomatic | janonymous_: I tried to figure out how to do it with email.parser, but couldn't get it to stop slurping document bodies into memory; that was in response to someone saying rfc822 was deprecated in a review comment, though | 19:10 |
| janonymous_ | I have kept on hold furthur patches as earlier were not reviewed. should i continue to submit patches ? | 19:10 |
| *** joeljwright has quit IRC | 19:11 | |
| torgomatic | janonymous_: probably let the existing ones drain a bit first, then continue; otherwise you're constantly rebasing. Easier to have just a couple outstanding at a time. | 19:11 |
| janonymous_ | Ohh.. i see | 19:12 |
| *** Guest1999 has left #openstack-swift | 19:13 | |
| zackmdavis | torgomatic, janonymous_, re header parsing without touching the rest of the file, I was thinking we're doomed to just include the 20ish lines of code we need to do it ourselves; you can see where it happens inside of email.feedparser, but email.feedparser is not modular enough to let us call just that: https://github.com/python/cpython/blob/master/Lib/email/feedparser.py#L220-236 | 19:15 |
| torgomatic | zackmdavis: yeah, that'd work. it's either that or expose that as a method on email.feedparser, but that'd require patches to Python itself | 19:17 |
| *** wbhuber has joined #openstack-swift | 19:20 | |
| *** agentle__ has joined #openstack-swift | 19:20 | |
| *** agentle_ has quit IRC | 19:24 | |
| janonymous_ | :o | 19:25 |
| *** petertr7 is now known as petertr7_away | 19:28 | |
| *** barker has joined #openstack-swift | 19:31 | |
| *** fthiagogv has quit IRC | 19:31 | |
| *** petertr7_away is now known as petertr7 | 19:31 | |
| *** fthiagogv has joined #openstack-swift | 19:36 | |
| *** mgarza_ has quit IRC | 19:37 | |
| *** jrichli has quit IRC | 19:37 | |
| *** mgarza_ has joined #openstack-swift | 19:38 | |
| openstackgerrit | Michael Barton proposed openstack/swift: go: fix expect 100-continue in replicator https://review.openstack.org/201690 | 19:46 |
| amandap | @notmyname Thanks :) Everyone seems to do it a bit different but referring to the HowTo wiki and the Infra page seems most common. | 19:47 |
| *** mgarza_ has quit IRC | 19:49 | |
| *** Abyss_ has joined #openstack-swift | 19:49 | |
| *** chlong-afk has joined #openstack-swift | 19:49 | |
| *** blmartin has joined #openstack-swift | 19:49 | |
| *** mgarza_ has joined #openstack-swift | 19:49 | |
| *** sileht has joined #openstack-swift | 19:49 | |
| *** mrda-away has joined #openstack-swift | 19:52 | |
| openstackgerrit | janonymous proposed openstack/swift: Adding bandit for security static analysis testing in swift https://review.openstack.org/196395 | 19:54 |
| *** mgarza_ has quit IRC | 19:57 | |
| *** mgarza_ has joined #openstack-swift | 19:57 | |
| *** Protux has joined #openstack-swift | 19:57 | |
| *** barker has quit IRC | 20:02 | |
| *** agentle__ is now known as annegentle | 20:02 | |
| *** zhiyan has joined #openstack-swift | 20:02 | |
| *** SkyRocknRoll has quit IRC | 20:03 | |
| *** robefran has quit IRC | 20:16 | |
| openstackgerrit | paul luse proposed openstack/swift: Add ability for GET path to see/select alternate fragment archives https://review.openstack.org/201283 | 20:28 |
| notmyname | peluse: that patch is resolving one of the EC bugs right? | 20:28 |
| peluse | torgomatic, clayg ^ ready for some eyes. Still needs more tests but please review the implementation | 20:28 |
| peluse | notmyname, maybe :) | 20:28 |
| peluse | depends | 20:28 |
| peluse | possibly | 20:28 |
| peluse | we'll see | 20:29 |
| notmyname | peluse: heh, ok. if it is, a bug reference int he commit message would be nice. I remember you saying you were working on something when I was asking about one of the bugs. just don't remember the details | 20:29 |
| peluse | notmyname, yeah. all kidding aside this is something we talked about a while ago and since then some bugs have come up that are most definately related but ebcause I started this becfore those bugs were filled I'm not 100% sure | 20:30 |
| peluse | after adding some probe tests it should be clearer | 20:30 |
| notmyname | good | 20:31 |
| peluse | notmyname, so it *might* have material impact on bug #1457691 | 20:34 |
| openstack | bug 1457691 in OpenStack Object Storage (swift) "Insufficient Fragments occurred on EC GET object" [High,Confirmed] https://launchpad.net/bugs/1457691 - Assigned to Bill Huber (wbhuber) | 20:34 |
| peluse | will start by adding a suggestd probe test from clayg written up on that report | 20:34 |
| wbhuber | peluse: I've tried running the probe test that clayg wrote up on the report. it hung and i've been trying to zero in on where the train wrecks at. | 20:35 |
| *** ccavanna has quit IRC | 20:36 | |
| *** janonymous has quit IRC | 20:36 | |
| peluse | wbhuber, is there already a probe test written? I just see a scenario in english. if you can post what you have I'll try it on my stuff | 20:37 |
| wbhuber | peluse: sure. | 20:38 |
| peluse | wbhuber, also please take a look at https://review.openstack.org/#/c/201283/ when you get a chance | 20:38 |
| *** sanchitmalhotra has quit IRC | 20:38 | |
| *** mgarza_ has quit IRC | 20:38 | |
| *** mgarza has joined #openstack-swift | 20:39 | |
| janonymous_ | could somone please confirm my understanding as mentioned in last comment on bug : https://bugs.launchpad.net/swift/+bug/1439816 | 20:39 |
| openstack | Launchpad bug 1439816 in OpenStack Object Storage (swift) "If PID file becomes empty, swift-init throws exception" [Undecided,In progress] - Assigned to janonymous (janonymous) | 20:39 |
| torgomatic | janonymous_: yep | 20:41 |
| notmyname | torgomatic: can you please mark it as fix committed? | 20:42 |
| torgomatic | SO MUCH CLICKING | 20:42 |
| notmyname | indeed | 20:42 |
| notmyname | you'll get launchpad karma | 20:42 |
| *** mgarza has quit IRC | 20:42 | |
| notmyname | that and $2 will get you a cup of coffee | 20:43 |
| peluse | crappy coffee me thinks | 20:44 |
| notmyname | better coffee would require more LP karma | 20:45 |
| openstackgerrit | Minwoo Bae proposed openstack/swift: EC: Handoff node to push existing fragment to the correct location. https://review.openstack.org/196848 | 20:45 |
| peluse | more clicky cicky clicky | 20:45 |
| torgomatic | we need an update to RFC 2324 with the POUROVER verb; the response SHOULD include an Origin header | 20:47 |
| notmyname | unfortunately, the response is always a 402 | 20:48 |
| *** sanchitmalhotra has joined #openstack-swift | 20:48 | |
| *** janonymous_ has quit IRC | 20:50 | |
| *** mgagne_ is now known as mgagne | 20:50 | |
| *** janonymous has joined #openstack-swift | 20:51 | |
| janonymous | coffee , wow :D | 20:53 |
| notmyname | seems that I no longer have the ability to push a release for swiftclient. working through the bureaucracy now | 21:09 |
| *** petertr7 is now known as petertr7_away | 21:12 | |
| klrmn | notmyname: wow, if not you, then who did they expect to do it? | 21:14 |
| notmyname | seems I missed the announcement that it's a whole managed thing now | 21:15 |
| *** sanchitmalhotra1 has joined #openstack-swift | 21:17 | |
| *** sanchitmalhotra has quit IRC | 21:19 | |
| *** janonymous has quit IRC | 21:20 | |
| notmyname | so here's another thing that should be looked at by people in swift https://review.openstack.org/#/c/164035 | 21:20 |
| *** blmartin has quit IRC | 21:22 | |
| *** fthiagogv has quit IRC | 21:31 | |
| redbo | the only sucky thing would be how often we use "with Timeout" | 21:33 |
| notmyname | redbo: not anything about having to have enough threads to handle 15k req/sec? | 21:34 |
| openstackgerrit | Nadeem Syed proposed openstack/swift: go: adding middleware & routing framework https://review.openstack.org/195279 | 21:35 |
| redbo | Well, I figured it could be a regular wsgi application and we could still run it on eventlet if we wanted. | 21:35 |
| *** jlhinson has quit IRC | 21:36 | |
| redbo | I'm assuming the people suggesting threads have never tried it. | 21:36 |
| notmyname | lol | 21:36 |
| notmyname | that spec and the discovery that I can no longer make swiftclient releases makes me glad I bought a bottle of booze at lunch | 21:37 |
| torgomatic | Threads work really well compared to eventlet when you have a C-level mysql connection that blocks the eventlet hub constantly. | 21:37 |
| torgomatic | *cough* hypothetically | 21:38 |
| redbo | That's fair. They might work well on our container server too. | 21:38 |
| notmyname | torgomatic: guess where the major point on server contention is in most other openstack projects ;-) | 21:39 |
| redbo | The GIL is supposed to be better in python 3, but on something like the proxy where it's all networking and python code, eventlet is still probably going to win. | 21:42 |
| notmyname | python-swiftclient release is delayed at least until tomorrow while I figure out the new state of things | 21:44 |
| *** NM has quit IRC | 21:50 | |
| *** delattec has quit IRC | 21:52 | |
| zaitcev | I submitted a talk for M summit using the joke I made during IRC meeting and redbo seconded. If only I could use the IRC log somehow (well, assuming it gets accepted) | 21:56 |
| *** shakamunyi has joined #openstack-swift | 21:58 | |
| *** doxavore_ has quit IRC | 22:00 | |
| openstackgerrit | Bill Huber proposed openstack/swift: EC: Fix Insufficient Fragment on secondary GET call https://review.openstack.org/201791 | 22:06 |
| wbhuber | peluse: here's a probe test written that is hung. at your earliest convenience, https://review.openstack.org/#/c/201791/ | 22:09 |
| *** hrou has quit IRC | 22:12 | |
| *** kutija has quit IRC | 22:13 | |
| *** minwoob has quit IRC | 22:14 | |
| *** cdelatte has joined #openstack-swift | 22:14 | |
| *** delattec has joined #openstack-swift | 22:14 | |
| *** annegentle has quit IRC | 22:18 | |
| *** kutija has joined #openstack-swift | 22:20 | |
| *** sanchitmalhotra has joined #openstack-swift | 22:28 | |
| *** sanchitmalhotra1 has quit IRC | 22:29 | |
| *** zaitcev has quit IRC | 22:33 | |
| *** darrenc is now known as darrenc_afk | 22:35 | |
| *** kutija_ has joined #openstack-swift | 22:36 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 22:38 | |
| *** kutija has quit IRC | 22:39 | |
| *** sanchitmalhotra has quit IRC | 22:40 | |
| mattoliverau | Morning | 22:44 |
| *** ccavanna has joined #openstack-swift | 22:46 | |
| *** wbhuber has quit IRC | 22:48 | |
| *** darrenc_afk is now known as darrenc | 22:56 | |
| *** km has joined #openstack-swift | 23:00 | |
| *** lcurtis has quit IRC | 23:00 | |
| *** hrou has joined #openstack-swift | 23:01 | |
| *** pgbridge has quit IRC | 23:18 | |
| *** delattec has quit IRC | 23:29 | |
| *** cdelatte has quit IRC | 23:29 | |
| *** bitblt has joined #openstack-swift | 23:31 | |
| *** barra204 has joined #openstack-swift | 23:35 | |
| *** shakamunyi has quit IRC | 23:36 | |
| *** cdelatte has joined #openstack-swift | 23:47 | |
| *** cdelatte has quit IRC | 23:47 | |
| *** ho has joined #openstack-swift | 23:50 | |
| ho | good morning! | 23:50 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!