Thursday, 2016-05-19

*** jamielennox is now known as jamielennox|away00:04
*** jamielennox|away is now known as jamielennox00:04
*** gyee has quit IRC00:08
*** gyee has joined #openstack-swift00:08
*** diogogmt has quit IRC00:23
*** Suyi has quit IRC00:25
*** sams-gleb has joined #openstack-swift00:26
*** diogogmt has joined #openstack-swift00:27
*** asettle has joined #openstack-swift00:27
*** asettle has quit IRC00:33
*** gyee has quit IRC00:34
*** lyrrad has quit IRC00:34
*** gyee has joined #openstack-swift00:34
*** sams-gleb has quit IRC00:35
*** diogogmt has quit IRC00:36
*** duobei has joined #openstack-swift01:03
*** bill_az has joined #openstack-swift01:05
*** NM has quit IRC01:06
*** NM has joined #openstack-swift01:12
*** _JZ_ has quit IRC01:13
openstackgerritTim Burke proposed openstack/swift: Allow concurrent bulk deletes
*** diogogmt has joined #openstack-swift01:14
*** Jeffrey4l has joined #openstack-swift01:30
*** sams-gleb has joined #openstack-swift01:31
*** saltsa has quit IRC01:35
*** saltsa has joined #openstack-swift01:37
*** sams-gleb has quit IRC01:39
*** bill_az has quit IRC01:41
*** duobei has quit IRC01:49
*** duobei has joined #openstack-swift01:56
openstackgerritMatthew Oliver proposed openstack/swift: disable_fallocate also disables fallocate_reserve
*** StraubTW has joined #openstack-swift02:01
*** tqtran has quit IRC02:09
*** StraubTW has quit IRC02:10
*** tqtran has joined #openstack-swift02:18
*** tqtran has quit IRC02:26
*** links has joined #openstack-swift02:35
*** sams-gleb has joined #openstack-swift02:36
*** sams-gleb has quit IRC02:44
*** lakshmiS has quit IRC02:52
*** adu has joined #openstack-swift03:04
*** ozialien10 has quit IRC03:18
*** ozialien10 has joined #openstack-swift03:19
*** tqtran has joined #openstack-swift03:23
*** tqtran has quit IRC03:27
*** psachin has joined #openstack-swift03:38
*** kei_yama has joined #openstack-swift03:40
*** sams-gleb has joined #openstack-swift03:41
*** duobei has quit IRC03:46
*** sams-gleb has quit IRC03:50
*** ouchkernel has joined #openstack-swift03:52
*** hosanai has joined #openstack-swift03:54
*** ChanServ sets mode: +v hosanai03:54
*** _JZ_ has joined #openstack-swift03:55
*** nadeem has joined #openstack-swift03:56
*** links has quit IRC04:00
*** rickyrem has quit IRC04:01
*** chlong has joined #openstack-swift04:04
*** psachin has quit IRC04:05
*** links has joined #openstack-swift04:16
*** Jeffrey4l has quit IRC04:16
*** psachin has joined #openstack-swift04:20
*** can8dnSix has joined #openstack-swift04:21
mahaticgood morning. -meeting scrollback done04:21
*** adu has quit IRC04:29
*** adu has joined #openstack-swift04:29
openstackgerritTony Breeds proposed openstack/swift: [WIP] Testing latest u-c
mattoliveraumahatic: morning04:41
mahaticmattoliverau: o/04:41
*** adu has quit IRC04:43
*** sams-gleb has joined #openstack-swift04:47
*** eranrom has quit IRC04:52
*** sams-gleb has quit IRC04:54
*** dmorita has quit IRC05:09
*** dmorita has joined #openstack-swift05:11
*** can8dnSix has quit IRC05:12
*** ppai has joined #openstack-swift05:12
*** dmorita has quit IRC05:16
*** chlong has quit IRC05:17
*** zaitcev has quit IRC05:20
*** tqtran has joined #openstack-swift05:24
openstackgerritKota Tsuyuzaki proposed openstack/swift: Fix probe failure and small things
*** ChubYann has quit IRC05:28
*** tqtran has quit IRC05:28
*** chlong has joined #openstack-swift05:30
openstackgerritKota Tsuyuzaki proposed openstack/swift: Fix probe failure and small things
*** duobei has joined #openstack-swift05:48
*** sams-gleb has joined #openstack-swift05:51
*** gyee has quit IRC05:56
*** Jeffrey4l has joined #openstack-swift05:58
*** sams-gleb has quit IRC06:00
*** chlong has quit IRC06:05
*** rcernin has joined #openstack-swift06:13
*** Jeffrey4l has quit IRC06:18
*** chlong has joined #openstack-swift06:22
*** _JZ_ has quit IRC06:23
*** duobei has quit IRC06:39
*** duobei has joined #openstack-swift06:45
*** chlong has quit IRC06:54
*** jmccarthy has quit IRC06:54
*** sams-gleb has joined #openstack-swift06:56
*** sams-gleb has quit IRC06:58
*** sams-gleb has joined #openstack-swift06:58
*** jmccarthy has joined #openstack-swift07:06
*** rledisez has joined #openstack-swift07:10
*** duobei has quit IRC07:29
*** ouchkernel has quit IRC07:35
*** admin0 has joined #openstack-swift07:45
*** jordanP has joined #openstack-swift07:46
*** dmk0202 has joined #openstack-swift07:50
*** duobei has joined #openstack-swift07:57
*** jamielennox is now known as jamielennox|away08:05
openstackgerritDavanum Srinivas (dims) proposed openstack/swift: [WIP] Testing latest u-c
*** daemontool has joined #openstack-swift08:21
*** asettle has joined #openstack-swift08:25
*** jordanP has quit IRC08:27
*** ozialien10 has quit IRC08:27
*** ozialien10 has joined #openstack-swift08:27
*** hezhiqiang has joined #openstack-swift08:27
*** jistr has joined #openstack-swift08:30
*** jordanP has joined #openstack-swift08:38
*** mmcardle has joined #openstack-swift08:39
*** mmcardle has quit IRC08:48
*** mmcardle has joined #openstack-swift08:52
openstackgerritMerged openstack/python-swiftclient: Use application/directory content-type for dir markers
*** nadeem has quit IRC08:59
*** duobei has quit IRC09:06
*** duobei has joined #openstack-swift09:08
*** acoles_ is now known as acoles09:09
acolesmahatic: hi - did you see the discussion re. post-as-copy of unencrypted object in the meeting log?09:10
acolesconclusion was that its ok for post-as-copy to encrypt the object data, so nothing for us to fix.09:11
*** mmcardle has quit IRC09:15
*** mmcardle has joined #openstack-swift09:27
*** mmcardle has quit IRC09:31
*** hseipp has joined #openstack-swift09:33
*** ZZelle has joined #openstack-swift09:36
*** hezhiqiang has quit IRC10:00
*** porunov has joined #openstack-swift10:01
porunovHello everyone! I am trying to install swift without others products but I have problem with adding drives to the swift ring. I use this command: "swift-ring-builder account.builder add --region USA_east --zone New_York --ip --port 6002 --device d1 --weight 100" . It shows me: "Usage: swift-ring-builder [options]. swift-ring-builder: error: option --region: invalid integer value: 'USA'". What's wrong with a name of the region? How can10:07
porunovI add new regions/zones into swift? Please help me10:07
*** mvk_ has quit IRC10:15
*** mmcardle has joined #openstack-swift10:16
*** duobei has quit IRC10:22
*** d0ugal has quit IRC10:24
*** d0ugal has joined #openstack-swift10:28
*** duobei has joined #openstack-swift10:30
*** d0ugal has quit IRC10:31
*** d0ugal has joined #openstack-swift10:31
*** jordanP has quit IRC10:36
*** duobei has quit IRC10:39
*** duobei has joined #openstack-swift10:45
openstackgerritKota Tsuyuzaki proposed openstack/swift: CompositeRing Support
acolesporunov: region and zone values need to be integers. it's not well documented, but there is a mention for zones here
*** mvk_ has joined #openstack-swift10:50
*** duobei has quit IRC10:50
kota_acoles: yeah, that's what I sent a response for porunov E-mail in openstack-dev mailing list.10:54
kota_I'm too late :/10:54
*** sams-gleb has quit IRC10:55
*** sams-gleb has joined #openstack-swift10:55
acoleskota_: hi! I didn't see the email yet :)10:56
kota_acoles: :-)10:57
kota_acoles: small notice, I updated probe test failure fix at patch 31591810:57
patchbotkota_: - swift - Fix probe failure and small things10:57
mahaticacoles: yeah, I did. thanks. less code and less testing10:58
acoleskota_: yes I saw, I will try to review again today10:58
kota_acoles: thank you very much!10:58
acoleskota_: I filed bug 158355110:58
openstackbug 1583551 in OpenStack Object Storage (swift) "docs should say more about ring regions and that regions and zones are ints" [Undecided,New]
*** mmcardle has quit IRC10:58
kota_acoles: great. yeah, we need better docs always.10:59
acolesmahatic: so I suggest, if you can, to focus on patches 315588 and the test you were working on to verify encrypted data on disk11:00
acolesmahatic: fyi  I am probably not here tomorrow11:00
*** jamielennox|away is now known as jamielennox11:03
mahaticacoles: oic. okay. have a fun long weekend!11:04
*** sams-gleb has quit IRC11:04
kota_is heading for home11:04
acolesmahatic: I was thinking more about the test to verify on disk content, and realised that the test_context has a tuple with all the test servers including the object servers, so you could just make a GET request direct to the first object server to get the object back.11:04
acoleskota_: good night11:04
acolesmahatic: sorry I didn't think of that before, i was thinking how could you get a handle on the object server DiskFileRouter, then realised you could just make a request!11:05
*** mmcardle has joined #openstack-swift11:07
mahaticacoles: hmm would i still not need the node port etc? to be able to access ondisk data (and not pass on the proxy_app/crypto_app)?11:12
acolesmahatic: no. similar to this line you should be able to find the obj server app then use it in a similar way to the requests to put object to the proxy e.g. line 7711:15
patchbotacoles: patch 318167 - swift (feature/crypto) - crypto - test varying proxy pipeline configs11:15
*** sams-gleb has joined #openstack-swift11:16
acolesmahatic: in fact you could create a _direct_get_object method as a helper11:16
acolesmahatic: I haven't tried it but I think that should work11:16
*** jordanP has joined #openstack-swift11:17
mahaticacoles: oic! makes sense, we have access to obj server. I will try that. thanks much!11:20
acolesmahatic: unlike the requests to the proxy app, these requests to obj server don't have /v1 at start of the path11:24
mahaticacoles: oic11:26
*** porunov has quit IRC11:27
mahaticacoles: why is that the case though?11:30
*** psachin has quit IRC11:33
acolesmahatic: the 'v1' part is removed in proxy server so requests to obejct servers don't have it11:34
*** mmcardle has quit IRC11:35
acolesmahatic: but you DO need the device and partition at start of path (sorry I may have isunderstood you earlier question about needing node & port)11:35
mahaticacoles: right, I thought something like this: http_connect(node['ip'], node['port'], node['device'], part, 'GET', path)11:37
mahaticacoles: which the direct_get_object does (in direct_client11:37
acolesmahatic: that's what you'd need connect to the server socket, but you have a handle for the server app so you're just calling the __call__ method of the server's ObjectController11:39
acolesif you follow through the http_connect function you'll see it builds a path /dev/part/a/c/o - that's what you would need a swob Request path to be11:40
acolessomething like swob.Request(/dev/part/a/c/o).get_response(object_server_app)11:43
acoles(not literally that of course!)11:43
mahaticyeah, (I'm following through the http_connect)11:44
*** NM has quit IRC11:46
*** Shashikant86 has joined #openstack-swift11:54
*** hosanai has quit IRC11:56
*** psachin has joined #openstack-swift12:00
*** SkyRocknRoll has joined #openstack-swift12:04
admin0hi all .. i I see a db does not exit error like this:  — what could be the error ?12:05
admin0DatabaseConnectionError(self.db_file, "DB doesn't exist")#012DatabaseConnectionError: DB connection error12:06
admin0DB connection error (/srv/node/data13/accounts/311400/0e7/980d086aa111aa719d770f136d6cc0e7/980d086aa111aa719d770f136d6cc0e7.db, 0):#012DB doesn't exist12:07
*** jistr has quit IRC12:08
*** raildo-afk is now known as raildo12:15
*** kei_yama has quit IRC12:19
*** jistr has joined #openstack-swift12:28
*** jistr is now known as jistr|bbl12:28
*** mmcardle has joined #openstack-swift12:32
*** CaioBrentano has quit IRC12:35
*** links has quit IRC12:36
*** Shashikant86 has quit IRC12:43
*** duobei has joined #openstack-swift12:50
*** Shashikant86 has joined #openstack-swift12:50
*** klamath has joined #openstack-swift12:52
*** klamath has quit IRC12:57
*** klamath has joined #openstack-swift12:57
*** hseipp has quit IRC13:00
*** ppai has quit IRC13:11
*** mmcardle has quit IRC13:12
*** vinsh_ is now known as Vinsh13:16
*** pauloewerton has joined #openstack-swift13:16
*** diogogmt has quit IRC13:17
*** duobei has quit IRC13:18
*** _JZ_ has joined #openstack-swift13:20
pdardeaugood morning13:25
*** mmcardle has joined #openstack-swift13:30
*** ouchkernel has joined #openstack-swift13:32
*** BigWillie has joined #openstack-swift13:37
openstackgerritOpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements
*** vint_bra has joined #openstack-swift13:40
tdasilvaworth reading:
*** ametts has joined #openstack-swift13:52
jrichlitdasilva: thanks!13:53
*** Shashikant86 has quit IRC13:57
openstackgerritAlfredo Moralejo proposed openstack/swift: Fix locale directory in
*** jistr|bbl is now known as jistr14:00
*** bill_az has joined #openstack-swift14:02
*** mmcardle has quit IRC14:04
*** Shashikant86 has joined #openstack-swift14:06
*** Shashikant86 has quit IRC14:06
*** mmcardle has joined #openstack-swift14:06
*** vint_bra1 has joined #openstack-swift14:08
*** dmorita has joined #openstack-swift14:11
*** vint_bra has quit IRC14:12
*** pauloewerton has quit IRC14:13
*** pauloewerton has joined #openstack-swift14:13
*** dmorita has quit IRC14:15
jrichlihello everyone.  the crypto code has been in flux for quite some time, but has finally become stable enough to now get confirmation on our encryption scheme.14:25
jrichliif you are knowledgeable about cryptography, we would like to hear your opinion.14:26
jrichliplease visit and let us know your thoughts.  torgomatic : would like to know your thoughts in particular.14:27
* jrichli will be away for the next couple hours14:30
*** diogogmt has joined #openstack-swift14:33
admin0hi all .. if i see a line like   "GET /data09/159382/AUTH_66797544cf82404bb4122b97b5e2c934/handleidingen/pdfjs-viewer/web/images/toolbarButton-search.png" 200 503 "GET http:14:47
admin0is this a 200 or a 503 ?14:47
admin0where can i learn more about swift log formats14:48
*** fesp has joined #openstack-swift14:54
*** sams-gleb has quit IRC15:06
*** sams-gleb has joined #openstack-swift15:06
*** links has joined #openstack-swift15:11
*** sams-gleb has quit IRC15:15
notmynamegood morning15:19
notmynameoh boy
*** links has quit IRC15:26
*** dmk0202 has quit IRC15:33
ahalea wild mailng list thread appears..15:34
openstackgerritAlistair Coles proposed openstack/swift: crypto - add overview doc (WIP)
*** asettle has quit IRC15:37
*** psachin has quit IRC15:39
*** zaitcev has joined #openstack-swift15:40
*** ChanServ sets mode: +v zaitcev15:40
*** admin0 has quit IRC15:48
acolestimburke: you have a +1 on patch 313640, were there remaining issues to address to convert you to a +2? there was some discussion about using ECB and re-use of IV, but we have those on trello so I'm wondering if you feel we could land this and follow up with any change in strategy?15:53
patchbotacoles: - swift (feature/crypto) - crypto - wrap random body key15:53
* swiftfan goes to get some popcorn15:53
openstackgerritMerged openstack/swift: Use the same key for memcache and env['swift.infocache']
tdasilvaacoles: i was just reviewing that too and running tests now15:54
*** lyrrad has joined #openstack-swift15:54
openstackgerritAlistair Coles proposed openstack/swift: crypto - make some probe tests compatible with crypto
tdasilvaI think it is good to go15:54
acolestdasilva: great, thanks15:54
* notmyname needs to step out for a little bit for a kid's school thing15:56
*** joeljwright has joined #openstack-swift15:57
*** ChanServ sets mode: +v joeljwright15:57
*** SkyRocknRoll has quit IRC16:01
*** lakshmiS has joined #openstack-swift16:01
bkeller`so i saw that with container and account metadata, you can set it to empty to remove the metadata, but with objects it just creates an empty metadata property16:01
bkeller`is that supposed to happen?16:01
ahaleyou can send a X-Remove-Account-Meta-* header, about halfway down
*** marcusvrn_ has quit IRC16:03
*** rcernin has quit IRC16:04
bkeller`yeah, but it seems weird that accounts/containers act differently than objects for that header16:08
bkeller`especially as it's a documented feature for accounts/containers16:08
acolesbkeller`: Object metadata behaves differently to containers and accounts. Container/account metadata items can be individually updated or removed. Sending an empty meta header with an account or container POST will delete that header name from the account or container metadata.16:10
acolesobject metadata is always replaced as a whole on each POST to the object.16:10
acolesI'd never noticed that the empty metadata was reported back in the object's case.16:10
bkeller`that makes sense then16:11
bkeller`i'm guessing the empty object metadata wasn't intentional then?16:12
*** sams-gleb has joined #openstack-swift16:12
bkeller`test_POST_update_meta actually tests for it though16:16
acolesI think it's more that the container/account handling of the empty meta header is different in order to support the act of deleting that item. Since an empty header in that case means to delete, it would not make sense to then return the empty meta. There's no such delete semantic with objects, so you just get back what you POSTed.16:17
acolesthat's my interpretation.16:17
bkeller`alright, thanks16:19
*** xenthree3 has joined #openstack-swift16:20
*** xenthree3 has left #openstack-swift16:20
*** sams-gleb has quit IRC16:20
*** rledisez has quit IRC16:24
*** fesp has quit IRC16:29
*** klrmn has joined #openstack-swift16:32
openstackgerritAlistair Coles proposed openstack/swift: Fix probe failure and small things
*** jordanP has quit IRC16:38
*** vint_bra has joined #openstack-swift16:42
*** vint_bra1 has quit IRC16:43
*** vint_bra1 has joined #openstack-swift16:43
*** vint_bra has quit IRC16:47
*** tqtran has joined #openstack-swift16:50
*** vint_bra1 has quit IRC16:50
*** mvk_ has quit IRC16:51
*** admin0 has joined #openstack-swift16:56
* notmyname back17:07
*** dmorita has joined #openstack-swift17:11
*** admin0 has quit IRC17:13
*** dmorita has quit IRC17:15
*** admin0 has joined #openstack-swift17:15
*** sams-gleb has joined #openstack-swift17:17
*** nadeem has joined #openstack-swift17:18
*** dmorita has joined #openstack-swift17:19
*** joeljwright has quit IRC17:25
*** sams-gleb has quit IRC17:25
*** zzelle_ has joined #openstack-swift17:26
*** mmcardle has quit IRC17:26
zzelle_zaitcev, hi17:31
*** nadeem has quit IRC17:33
*** nadeem has joined #openstack-swift17:38
clayganyone have time for a quick-y?  lp bug #1583719 would be some good experience with db replicators (which are *always* fun!)17:47
openstackLaunchpad bug 1583719 in OpenStack Object Storage (swift) "empty db hash/suffix directories can pile up" [Medium,New]
*** daemontool has quit IRC17:47
ntata|notmyname, do you have interesting stuff to share about saio on ubuntu 16.04?17:49
claygit's kinda funny tracking that bug history back - we've "fixed" empty db dir cleanup at least three times ;)17:49
ntata|it was by far the most smoothest saio installation for me17:49
notmynamentata|: not really. pretty much everything just worked17:49
claygnotmyname: ntata|: nowhey!17:49
notmynamentata|: the only "gotcha" is that to enable rsync across restarts I I had to do `sudo systemctl enable rsync` instead of (just) setting the /etc/defaults/rsync RSYNC_ENABLE=true flag17:50
openstackgerritMerged openstack/swift: crypto - wrap random body key
*** Suyi has joined #openstack-swift17:50
notmynameand with the new bindep stuff in the swift repo, installing dependencies is `bindep -b | xargs sudo apt-get install -y`17:50
notmynameso I think there are a few cleanups or reordering of the SAIO doc in order, but pretty much it Just Works17:51
ntata|oh cool! I didn't try rsync yet.. will do now17:51
ntata|Yeah I believe, the right pip version solved so many problems I think17:52
notmynamewell, I still couldn't use the distro-packaged virtualenv (I think) because it didn't support the version designation on the requirements file. I think that was the one. point is, all that stuff we have to upgrade you still have to upgrade in xenial17:56
claygcrazy systemctl17:56
notmynameI mean, you might think that a distro released a few weeks ago would be "new enough" for openstack dependencies, but you'd be wrong ;-)17:57
acolestimburke: tdasilva: thanks17:59
acolesI'll be out tomorrow18:01
*** ChubYann has joined #openstack-swift18:18
*** ChubYann has quit IRC18:18
*** sams-gleb has joined #openstack-swift18:22
*** admin0 has quit IRC18:23
tdasilvaacoles: have a good weekend! :-)18:24
*** admin0 has joined #openstack-swift18:25
*** ChubYann has joined #openstack-swift18:26
*** dosaboy has quit IRC18:26
*** dmellado has quit IRC18:26
*** mahatic has quit IRC18:26
*** j_king has quit IRC18:26
*** lifeless has quit IRC18:26
*** timur has quit IRC18:26
*** j_king has joined #openstack-swift18:26
*** dosaboy has joined #openstack-swift18:26
*** lifeless has joined #openstack-swift18:26
*** mahatic has joined #openstack-swift18:27
*** dmellado has joined #openstack-swift18:27
tdasilvawhat's the hummingbird jenkins url again?18:28
*** sams-gleb has quit IRC18:30
*** nottrobin has quit IRC18:31
*** zhiyan has quit IRC18:31
*** gmmaha has quit IRC18:31
*** lcurtis has joined #openstack-swift18:33
*** timur has joined #openstack-swift18:33
*** nottrobin has joined #openstack-swift18:33
*** zhiyan has joined #openstack-swift18:34
notmynametdasilva: bah. I thought I had it, but it seems I lost that bookmark18:35
tdasilvai've got a saio setup running, but it's failing some functional tests, so I don't know if that's "expected" or not...18:35
*** ouchkernel has quit IRC18:36
*** ouchkernel has joined #openstack-swift18:38
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements
*** admin0 has quit IRC18:42
*** jistr has quit IRC18:43
*** gmmaha has joined #openstack-swift18:44
*** rcernin has joined #openstack-swift18:46
*** admin0 has joined #openstack-swift18:52
ahalegoogling for openstack swift libraries for apple swift is pretty tricky18:53
tdasilvayou might get some taylor swift results ;)18:54
klrmnahale: i think the answer is 'no', but i've been wrong before18:54
ahaletdasilva: you think maybe she wrote such a library ?18:55
*** diogogmt has quit IRC18:55
notmynameshe's a pretty sharp security expert18:55
tdasilvaahale: hehe, never know, she might be a geek hidden in superstar front18:55
zaitcevIt's not like it requires super extensive libraries if you only want basics (e.g. no automatic segmentation with SLO). A simple driver fo CloudFiles was 50 lines in C.18:55
zaitcevUnlike S3 you don't need to calculate HMAC and get it exactly right. Just pass a couple of strings around between HTTP requests.18:56
tdasilvanotmyname: lol, i knew it! :)18:57
*** klrmn has quit IRC18:57
openstackgerritTim Burke proposed openstack/swift: Make get_info requests useful with recheck_*_existence == 0
clayghas anyone been talking about a swift 2.8?18:58
notmynameclayg: we did a little at the meeting yesterday18:58
*** bzurawski has joined #openstack-swift18:59
claygnotmyname: are we going to wait for crypto - or cut 2.8 just before?18:59
zaitcevoh crap, the meeting!18:59
notmynameclayg: summary is get probetests passting, land patch 317475 and do a release18:59
patchbotnotmyname: - swift - Send correct size in POST async update for EC object18:59
notmynameI really don't want to wait for crypto for the next release19:00
openstackgerritTim Burke proposed openstack/swift: Make get_info requests useful with recheck_*_existence == 0
notmynameI think we should do one soon19:00
claygnotmyname: we should get a patch chain to merge crypto soon!19:00
notmynamealso mentioned in yesterday's meeting19:00
notmynameacoles and jrichli are on it. they have some reviews that need landing, and they're almost there for getting a patch chain ready19:01
notmynamepersonally, I'm hoping for next week sometime, but I haven't told them that yet. shhhh19:01
acolesnotmyname: need those reviews...19:02
notmynamethe 3 crypto patches listed on are the stuff that's top of the list for them19:02
acolesoh, one just got approved, let me update that19:02
timburkeacoles: too bad!19:04
timburkealready done19:04
acolestimburke: thanks!19:04
acolesI love people that clear up after a job is done :D19:05
clayghooray tdasilva zaitcev timburke for merging patch 280978 !!!19:08
patchbotclayg: - swift - Fix up get_account_info and get_container_info (MERGED)19:08
zaitcevclayg: I have to say, I did it purely because of elves19:09
zaitcevI mean, spent time reviewing it19:09
zaitcevAnd now I feel like venting about Ceph's build and test system19:09
claygzaitcev: you preach brother!19:10
zaitcevI see Tim found something lucking with it though...19:10
clayghooray zaitcev and mattoliverau for merging patch 31060819:11
patchbotclayg: - swift - Use smaller quorum size in proxy for even numbers ... (MERGED)19:11
claygzaitcev: is on a roll!19:11
zaitcevUmm, that one is super old. We discussed it in person in Austin.19:11
zaitcevIt sounded convincing19:12
*** diogogmt has joined #openstack-swift19:12
zaitcevclayg: the rolling 90-day says that acoles is far and far away the champion with 157 reviews19:13
*** diogogmt has quit IRC19:15
zaitcevMy official highest priority for today is to buy a workstation with 32 GB because our OpenStack installer will not install anything, including Swift, on the 12GB box that I have now.19:17
zaitcevMan, I sure wish 640 KB were enough for everybody.19:18
notmynamethat sounds like you're getting into the "real server + thin client" realm19:18
*** dmorita has quit IRC19:19
zaitcevTo tell you the truth, I have a 64 GB box in the lab, but I hosed it and accessing it remotely requires installing Fedora 12 (because it has a proper 32-bit Java that's compatible with Dell DRAC5 console).19:19
notmynamezaitcev: I'm sure you saw watch out for skylake (at least for mobile stuff)19:20
openstackgerritAlistair Coles proposed openstack/swift: Make SLO manifest copies retain correct content-type
acoleszaitcev: I'm slipping - done so little reviewing recently :/19:24
zaitcevnotmyname: I'd need to know what Skylake was first, before that post begins to matter to me. Also, if the CPU melts down because Linux wasn't idling it? I win a new workstation.19:24
notmynameskylake is the current newest generation of motherboards from intel (or motherboard platform? I'm not sure exactly how that works)19:25
zaitcevneither do I :-)19:25
clarkbit is the codename for the intel microarchitecture19:25
zaitcevThat much is obvious.19:25
notmynameyeah, I don't know if it's an actual intel product line or just a intel-defined standard19:25
clarkbit describes a family of cpus all built on the same process19:26
notmynamefrom what I've seen, it's pretty nice. unless you're needing to run linux on a laptop, I guess19:26
*** sams-gleb has joined #openstack-swift19:27
zzelle_zaitcev, hi, could you have a look to ? It needs an extra +2 since a month19:31
*** ouchkernel has quit IRC19:31
*** ouchkernel has joined #openstack-swift19:33
zaitcevzzelle_: But I don't know a thing about client certs. Not to mention the whole idea seems odd when applied to Swift. Why in the world are you authenticating with a certificate in addition to a password?19:34
*** sams-gleb has quit IRC19:34
*** diogogmt has joined #openstack-swift19:37
zzelle_zaitcev, it's for double-factor authentication like credit card with passwords or signature19:37
claygmattoliverau: lol @ ETOOEARLY19:40
clayglogs are fun19:41
claygcrap - i was hoping to get the update from tdasilva on liberasure?19:41
acolesnotmyname: i updated the priority reviews for crypto.19:41
claygthe *stupid* error/warning about he plugin's not loading is *soooo* annoying19:41
claygtimur: fixed it ages ago - I think tsg even merged it - but no new version tagged/released19:42
claygI'm going to end up just packaging some random liberasure sha <grrr>19:42
acolesnotmyname: there's a couple there that do not require great familiarity with the feature/crypto to review19:42
acolesclayg: everything will be ok ;)19:44
claygthanks bro19:44
* acoles back on Monday19:44
*** dmorita has joined #openstack-swift19:45
zzelle_zaitcev, two-factor auth increases security by using different ways to auth (what i own, what i know, what i am) which increases the difficult to hack someone auth as each way requires a different approach to steal/falsify it19:45
*** acoles is now known as acoles_19:48
*** McMurlock1 has joined #openstack-swift19:49
*** McMurlock1 has quit IRC19:49
zaitcevzzelle_: I know what 2FA is.19:50
*** diogogmt has quit IRC19:55
tdasilvaclayg: on liberasure code, things have been moving...slowly ;)19:56
tdasilvaclayg: both projects have been moved over19:56
tdasilvabut no builds yet on the gate19:56
tdasilvai'm currently working on pyeclib gate jobs19:56
tdasilvaand then will work on liberasurecode19:56
*** ouchkernel has quit IRC19:58
notmynameacoles_: thanks19:58
notmynametdasilva: ah, thanks for the info19:59
*** gyee has joined #openstack-swift20:02
zaitcevzzelle_: You wrote in the bug 1565112 that "other clients (nova/neutron/glanceclient) allows to specify them using --os-cert, --os-key". But a quick look at python-novaclient fails to identify where it actually happens. It seems that nova only takes CA certificate (for the server). Do you know where it is actually done? Depths of Oslo somewhere?20:02
openstackbug 1565112 in python-swiftclient "Client certificate/key are not supported " [Undecided,In progress] - Assigned to Cedric Brandily (cbrandily)20:02
zzelle_zaitcev, iirc, it's provided by keystonemiddleware or keystoneauth20:03
zzelle_zaitcev, let me check20:03
zaitcevzzelle_: What do you mean? We're discussing the client now, not middleware.20:03
*** diogogmt has joined #openstack-swift20:04
zaitcevzzelle_: Anyway, all I need is to make sure we're not going to rename --os-cert soonish because the rest of OpenStack implements something like --os-client-certificate20:06
zzelle_zaitcev, iirc --os-key and --os-cert is inherited by novaclient from keystoneauth/keystonemiddleware20:08
zaitcevthat just makes no sense20:08
zaitcevclient cannot inherit from middlewares, or can it now20:08
*** diogogmt has quit IRC20:08
zaitcevWell, I see that your change for the common client went in, so we have that.20:09
*** diogogmt has joined #openstack-swift20:09
*** ouchkernel has joined #openstack-swift20:11
bzurawskisorry for interrupting but i want to ask, what do you think about removing all loggers from swiftclient/ as it's causing problems for users, it's bug 121317920:12
openstackbug 1213179 in python-swiftclient "swiftclient logs all ClientException (404) at ERROR (with traceback)" [High,Confirmed] - Assigned to Bartek Żurawski (bartekzurawski1)20:12
openstackgerritMerged openstack/swift: Fix locale directory in
zaitcevoh, wait20:12
zzelle_zaitcev, novaclient inherits it through
zaitcevzzelle_: thanks20:13
zzelle_and define --os-key/cert in
zaitcevzzelle_: thanks... I also checked that our shipped nova takes those20:14
zzelle_i doubt the option will change has some clients had different options for client certs and moved to --os-key/cert20:15
timburkebzurawski: removing all loggers seems like a bad way to go; we'd lose a lot of info/debug logging that can be useful, especially on the CLI. i think i'd be on board with not logging an exception when we're about to reraise, though20:24
*** diogogmt has quit IRC20:25
bzurawskitimburke: but isn't swiftclient/ is using for CLI not I can be wrong, but is used mostly by devs who want's to create they're own solution to talk to swift20:28
bzurawskitimburke: shell is using swiftclient/ only for importing logger config, so maybe we can move this logger definition to
*** diogogmt has joined #openstack-swift20:29
claygbzurawski: service api is also for applications - not just CLI20:31
bzurawskitimburke: but if we will be on board with not logging as exception, you mean then change logger.exception to logger.debug for example ?20:31
*** sams-gleb has joined #openstack-swift20:31
zzelle_zaitcev, thx20:32
*** daemontool has joined #openstack-swift20:32
timburkebzurawski: all of service goes through client. i mean that i very much *want* the logging at info/debug going on in http_log - see
timburkebut i think we can remove *all* of the logger.exception calls, as we only ever do it before raising20:33
claygtimburke: I think it's weird that we re-raise this one w/o first translating to ClientException ->
claygtimburke: hurrmrm... this one too ->
claygidk, the whole thing probably requires a lot of thought - maybe I haven't put in enough - changing to debug seems like an easy way to cut down the noise without having to think about it too hard20:37
timburkeclayg: yeah, probably not the best. i've always found those a little weird; i guess i could see it if there's a timeout or something, but i don't think retrying will help with DNS failures or refused connections20:37
claygtimburke: the other thing it would be *nice* to have logs around is like what was the history of retried attempts that lead to a "failed with XXX after X attempts"20:39
*** dmk0202 has joined #openstack-swift20:39
claygbut logging the traceback on a valid wellformed well known http client error status seems so stupid :'(20:39
timburketrue. they may not have all been the same types of failures20:40
*** sams-gleb has quit IRC20:40
*** tqtran has quit IRC20:40
*** zzelle_ has quit IRC20:49
bzurawskiclayg: timburke so maybe removing logger.exception from function _retry from ClientException block, would be good as a first step, but what about this re-raise things, should we do something with it ?20:57
*** BigWillie has quit IRC20:58
*** klrmn has joined #openstack-swift20:58
*** tqtran has joined #openstack-swift21:00
*** raildo is now known as raildo-afk21:10
claygyeah anytime you see `logger.exception(); raise` it smells21:14
bzurawskiclayg: timburke Oki, so I'm starting fixing this issue, if any question or troubles i will be asking :) BTW. thanks for you're help with understanding how it can be corrected :)21:20
claygbzurawski: thanks for working on it!21:20
*** dmk0202 has quit IRC21:22
*** pauloewerton has quit IRC21:26
ntata|clayg, I'm working on bug - and am wondering if the "[]" need to be added only to the ipv6 addresses and not for ipv4?21:36
openstackLaunchpad bug 1567105 in OpenStack Object Storage (swift) "swift-ring-builder output display bad column alignment with ipv6" [Wishlist,New] - Assigned to Nandini (nandini-tata)21:36
*** sams-gleb has joined #openstack-swift21:37
zaitcevntata: What makes you think that brackets are needed in that bug, at all? The bug is that the column alignment is bad. The [] syntax is only needed when a colon may be adjacent to the address, as in URI's netloc. This is not the case in the outout of swift-ring-builder.21:41
zaitcevntata: actually, never mind, I see Clay's comment now.21:42
*** sams-gleb has quit IRC21:45
*** lakshmiS has quit IRC21:45
*** diogogmt has quit IRC21:50
*** daemontool has quit IRC21:57
notmynamemaybe next week will have less 'epic mailing list threads' in it, /me says hopefully21:58
timburkenotmyname: time to go camping for a week?21:59
notmynamelol. I actually am going camping not this weekend but the next :-)22:01
*** rickyrem has joined #openstack-swift22:05
*** nadeem has quit IRC22:07
mattoliverauLol, notmyname epic thread master22:13
jrichlimattoliverau: i know, I didn't realize swift was so controversial :-)22:14
notmynamemattoliverau: no. please no. nooooo. don't make me do another one22:15
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements
tdasilvais the new dashboard missing a scrollbar or is it just me?22:17
notmynametdasilva: like mattoliverau, you must have a really wide browser window ;-)22:17
tdasilvai think it's a 22 or 2422:18
notmynametdasilva: I've got a stylesheet on it that over a certain width it reformats to be good on a wall-mounted TV. since that doesn't have any keyboard/mouse, I have the overflow: hidden on it to prevent scroll bars22:18
notmynameso if you narrow your browser window, it will look more natural22:19
mattoliverautdasilva: notmyname doesn't believe in scroll bars. He was going to write a mailing list thread on it :p22:19
notmynameand since you're the 2nd person to say something, it's probably actually a bug22:19
*** diogogmt has joined #openstack-swift22:19
* tdasilva goes to refresh mailing list page22:19
bkeller`there's an @media tv {  } you can use if your tv actually advertises itself correctly22:20
openstackgerritMerged openstack/python-swiftclient: Support client certificate/key
mattoliveraujrichli: yeah, we're the black sheep of the OpenStack world, probably cause ops know our stuff just works ;)22:20
notmynamebkeller`: yeah. just need to actually figure out if the raspberrypi driving the tv display has anything like that22:20
clarkbmattoliverau: you only have to look about 14 messages above this one to find a real example of why people may feel that way ;)22:22
clarkb(I get the reason for that but it does make swift different)22:22
*** rcernin has quit IRC22:23
pdardeaunotmyname: so are your epic ML threads the result of you trashing the movie script? ;-)22:30
mattoliverauclarkb: it's not different it's called special :)22:30
*** _JZ_ has quit IRC22:39
*** sams-gleb has joined #openstack-swift22:42
*** bzurawski has quit IRC22:42
*** sams-gleb has quit IRC22:50
tdasilvaif anybody is interested, I just updated my ansible saio script to add support for hummingbird:
*** diogogmt has quit IRC22:56
*** klamath has quit IRC23:03
*** ametts has quit IRC23:04
openstackgerritMerged openstack/swift: Fix probe failure and small things
openstackgerritMerged openstack/swift: disable_fallocate also disables fallocate_reserve
*** ChubYann has quit IRC23:16
*** ouchkernel has quit IRC23:26
*** sorrison_laptop has joined #openstack-swift23:27
*** lcurtis has quit IRC23:27
*** admin0 has quit IRC23:27
*** ChubYann has joined #openstack-swift23:28
*** ouchkernel has joined #openstack-swift23:28
*** kei_yama has joined #openstack-swift23:31
*** sams-gleb has joined #openstack-swift23:47
*** sorrison_laptop has quit IRC23:49
*** jamielennox is now known as jamielennox|away23:49
*** sorrison_laptop has joined #openstack-swift23:49
*** sorrison_laptop has quit IRC23:52
*** sorriso__ has joined #openstack-swift23:52
*** sams-gleb has quit IRC23:55
*** sorrison_laptop has joined #openstack-swift23:56
*** sorriso__ has quit IRC23:56
*** jmccarthy has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!