Tuesday, 2016-03-22

*** garthb has quit IRC00:08
*** lyrrad has quit IRC00:16
*** Zyric has quit IRC00:19
*** Zyric has joined #openstack-swift00:25
*** daemontool_ has joined #openstack-swift00:26
*** daemontool has quit IRC00:27
kota_good morning at office.00:35
*** MVenesio has joined #openstack-swift00:38
*** MVenesio has quit IRC00:43
claygsorry nm00:55
*** furlongm has quit IRC00:55
*** furlongm_ has joined #openstack-swift00:55
*** Jeffrey4l__ has quit IRC01:11
*** htaccess has left #openstack-swift01:14
*** klrmn has quit IRC01:16
*** rickyrem has quit IRC01:42
*** baojg has joined #openstack-swift01:43
*** rickyrem has joined #openstack-swift01:43
*** mrmoje has quit IRC02:02
*** panda has quit IRC02:05
*** mingdang1 has joined #openstack-swift02:05
*** panda has joined #openstack-swift02:06
*** Zyric has quit IRC02:10
*** klrmn has joined #openstack-swift02:10
*** dmorita has quit IRC02:12
mingdang1@kota_ hello, excuse me02:33
kota_mingdang1: hello02:34
kota_mingdang1: sorry I didn't notice on your message yesterday.02:34
kota_mingdang1: how's going?02:35
mingdang1oh,nothing,I have a question. where can i read the swift log?02:35
*** wanghua has joined #openstack-swift02:35
kota_mingdang1: I guess you can see /var/log/syslog in default (setting with SAIO docs)02:36
mingdang1i can see log in /var/log/message02:36
mingdang1but why not in /var/log/swift/proxy.log02:37
mingdang1can i setup where is write?02:37
kota_you can make a config via syslog and server conf (e.g. proxy-sever.conf, object-server.conf)02:39
*** MVenesio has joined #openstack-swift02:39
kota_you can make the setting via "log_facility" in the each conf file.02:40
kota_and then, make rsyslog conf file like as "dump the facility to here (e.g. /var/log/swift/proxy.log)02:41
kota_mingdang1: For example, (my VM setup)02:42
-kota_- local1.*;local1.!notice /opt/stack/data/swift/logs/proxy.log02:42
-kota_- local1.notice /opt/stack/data/swift/logs/proxy.error02:42
-kota_- local1.* ~02:42
kota_is in /etc/rsyslog.d/10-swift.conf02:42
kota_that means LOG_LOCAL1 facility will dump into /opt/stack/data/swift/logs/proxy.log02:43
kota_mingdang1: Is it what you are looking for?02:43
mingdang1en,maybe,I will try you said.:)02:44
*** MVenesio has quit IRC02:44
kota_mingdang1: :)02:44
*** asettle has quit IRC02:44
mingdang1local1 means what?02:46
kota_mingdang1: reserved word in rsyslog, maybe.02:48
kota_not have clear memory in my head.02:49
kota_ah, i mean 'LOG_LOCAL' for the reserved word.02:49
mingdang1@kota_ you means i first read from /var/log/syslog and choose something i need write to the file i want?02:49
kota_0~7 for the value.02:49
*** esker has quit IRC02:50
kota_mingdang1: you may see your current settting /etc/ryslog.conf or the files in /etc/rsyslog.d/ directory.02:50
mingdang1in ryslog.conf I can see "*.info;mail.none;authpriv.none;cron.none                /var/log/messages"02:53
zaitcevThat's fine, the ~ takes care of that duplication. You don't need to edit that line.02:55
kota_notmyname: done for patch 27944002:56
patchbotkota_: https://review.openstack.org/#/c/279440/ - swift - Skip already checked partitions when auditing obje...02:56
zaitcevAnd you would not have had the trouble if you installed official RDO RPMs for openstack-swift. They include the necessary configuration for local1 & local2.02:56
*** asettle has joined #openstack-swift02:56
mingdang1I want to choose proxy-server info to /var/log/swift/proxy.log02:57
kota_if you use default setting in proxy-server (i.e. log_facility = LOG_LOCAL0), adding a line "local0.* /var/log/swift proxy.log" to somewhere in the rsyslog conf file may work.03:01
zaitcevThat may be a little difficult without a basic understanding how syslog facility works.03:04
kota_mingdang1: rsyslog online manual is here, http://www.rsyslog.com/doc/v8-stable/index.html03:04
kota_zaitcev: maybe, I still sometimes be confused for the setting :/03:05
kota_specific document may be http://www.rsyslog.com/doc/v8-stable/historical/multi_ruleset_legacy_format_samples.html?highlight=local03:05
mingdang1@kota_ @zaitcev okay, thanks, i will try it03:08
*** 20WAALIJF has joined #openstack-swift03:17
*** links has joined #openstack-swift03:19
*** 20WAALIJF has quit IRC03:28
*** rickyrem has quit IRC03:29
*** haomaiwang has joined #openstack-swift03:30
*** rickyrem has joined #openstack-swift03:33
*** sgundur has left #openstack-swift03:35
openstackgerritMerged openstack/swift: Skip already checked partitions when auditing objects after a restart  https://review.openstack.org/27944003:43
mahaticgood morning03:52
kota_mahatic: morning03:58
mingdang1@kota_ have you ever used the EC policy?04:00
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-swift04:01
kota_mingdang1: yup, in test deployment though.04:01
*** aswadr_ has joined #openstack-swift04:01
mingdang1replicator and EC and use at the same time?04:03
*** rickyrem has quit IRC04:04
kota_you can use both (replicator/reconstructor) at the same time.04:04
mingdang1I need to create two kinds of ring?04:05
kota_each process checks the policy when collecting the replication/reconstruction jobs.04:05
kota_mingdang1: exactly04:05
*** ho_ has joined #openstack-swift04:06
*** dmorita has joined #openstack-swift04:06
openstackgerritDavid Liu proposed openstack/swift: test submit patch.  https://review.openstack.org/29561604:06
kota_mingdang1: you can set the policies at /etc/swift/swift.conf04:08
kota_mingdang1: in detail, http://docs.openstack.org/developer/swift/overview_policies.html04:08
mingdang1@kota_ do you have any guide for EC?04:09
kota_mingdang1: if you make more rings, you can deploy multiple replicaiton/EC policies into Swift04:09
kota_mingdang1: wait04:09
*** dmorita has quit IRC04:10
kota_mingdang1: http://docs.openstack.org/developer/swift/overview_erasure_code.html04:10
kota_mingdang1: community docs but it might be close to development side rather than operator side.04:11
kota_mingdang1: ah, specifically for users, http://docs.openstack.org/developer/swift/overview_erasure_code.html#using-an-erasure-code-policy04:12
*** chlong|wfh has quit IRC04:12
*** nakagawamsa has joined #openstack-swift04:12
kota_this section might be helpful.04:12
mingdang1yesterday, I try this guide,and get some error?maybe I will try again carefully04:12
kota_mingdang1: what's error appreared?04:13
mingdang1"ec_type = liberasurecode_rs_vand" it said liberasurecode_rs_vand is not a valid type04:14
kota_ah, ok. it looks the difference for the PyECLib version.04:14
kota_mingdang1: what version do you use?04:15
*** nakagawamsa has quit IRC04:15
kota_you can confirm your env via 'pip freeze | grep PyECLib'04:15
kota_alright, liberasurecode_rs_vand is not included yet in the version.04:16
kota_so you can set 'ec_type=jerasure_rs_vand' instead.04:16
mingdang1ok , i will try it04:18
kota_mingdagn1: you're welcome :)04:18
*** klrmn has quit IRC04:18
mingdang1@kota_ swift-ring-builder object-1.builder create 10 14 1  I only need to create a builder file?04:23
*** nadeem has joined #openstack-swift04:23
kota_mingdang1: no? you need object-1.ring.gz via swift-ring-builder object-1.builder rebalance04:24
*** treaki__ has joined #openstack-swift04:24
kota_as well as replicated.04:24
*** chlong has joined #openstack-swift04:25
*** treaki_ has quit IRC04:27
mingdang1@kota_ I need container.ring and account.ring?04:35
kota_mingdang1: not needed. same account/container ring used.04:35
*** furlongm_ is now known as furlongm04:36
*** MVenesio has joined #openstack-swift04:40
*** takashi has joined #openstack-swift04:40
takashigood afternoon :-)04:41
kota_takashi: !!!04:41
takashikota_: o/04:41
mingdang1@kota_  get error: pyeclib.Error: Invalid arguments passed to liberasurecode_instance_create04:42
kota_mingdang1: could you tell me the conf for EC policy?04:42
mingdang1name = ec10404:43
mingdang1policy_type = erasure_coding04:43
mingdang1ec_type = jerasure_rs_vand04:43
mingdang1ec_num_data_fragments = 1004:43
mingdang1ec_num_parity_fragments = 404:43
mingdang1ec_object_segment_size = 104857604:43
kota_looks like default value of example conf.04:43
*** MVenesio has quit IRC04:44
kota_can you see the error message from liberasurecode at /var/log/syslog?04:45
mingdang1liberasurecode_backend_open: dynamic linking error libJerasure.so: cannot open shared object file: No such file or directory04:46
kota_ah, you don't have jerasure binary.04:46
mingdang1can i use yum install?04:47
kota_hmm... I think PyECLib 1.0.7 have the installation package for jerasure (AFAIK) so 'pip install PyECLib==1.0.7' will install it.04:48
kota_mingdang1: wait a sec, check out the repo.04:48
kota_mingdang1: looks like there is a binary in the repo, hmm....04:49
kota_mingdang1: how did you setup the PyECLib?04:49
mingdang1I don't setup it04:49
kota_what swift version do you use?04:52
*** daemontool_ has quit IRC04:53
*** daemontool__ has joined #openstack-swift04:53
kota_if possible, I recommend >=2.6.0 to enable liberasurecode_rs_vand04:53
mingdang1you means pyeclib >= 2.6.0?04:54
*** furlongm has quit IRC04:54
kota_mingdang1: no, it's Swift version, L is 2.5.004:54
notmynamegood evening04:54
notmynamekota_: thanks for getting that review04:54
kota_notmyname: \o/04:54
notmynameIMO I think we're good for a release. anything else is gravy04:55
kota_notmyname: it was easy because just 2 lines diff from my last +204:55
notmynamealthough https://review.openstack.org/#/c/282578/ would be a nice improvement to EC to land :-)04:55
patchbotnotmyname: patch 282578 - swift - Set backend content length for fallocate - EC Policy04:55
kota_notmyname: yup, it's my favorite :)04:55
kota_mingdang1: so we have a couple of choices04:56
kota_mingdang1: 1. use L and setup jerasure by hand or by pip tool04:57
kota_mingdang1: 2. use >= 2.6.0 and liberasurecode_rs_vand.04:57
mingdang1if 1, how can i do04:58
kota_I think yum package repo have the newer liberasurecode binary and it will enables to use liberasurecode_rs_vand but04:58
kota_PyECLib==1.0.7 cap in L doesn't allow to use it.04:59
kota_mingdang1: I think easy way may be  'pip install PyECLib==1.0.7' but not sure a bit because of the lack of knowlege for your actuall env.05:00
kota_there might be more things to do for setting up.05:00
*** haomaiwang has quit IRC05:01
mingdang1my env PyECLib==1.0.705:01
*** haomaiwang has joined #openstack-swift05:01
kota_I don't know how yum package set up the PyECLib but setting via pip will kick the jerasure installation so...05:02
kota_yum remove PyECLib and then, pip install PyECLib==1.0.7?05:03
kota_ok, if you want to install just only jerasure05:03
kota_see http://lab.jerasure.org/jerasure/jerasure/blob/master/README05:04
kota_The way to install by hand is described in the docs.05:05
*** Zyric has joined #openstack-swift05:17
notmynamehttps://wiki.openstack.org/wiki/Swift/PriorityReviews has been slightly updated to reflect the status of patches for the release05:45
notmyname(thanks also to those who have been helping keep it up to date)05:45
notmynameI'll propose the authors/changelog first-draft tomorrow, most likely. and we're on track for the mitaka release this week05:46
notmynameand with that...good night05:46
kota_notmyname: thanks for great jobs, good night :)05:53
takashinotmyname: good night!05:55
*** baojg has quit IRC05:56
*** baojg has joined #openstack-swift05:57
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: WIP: Swift Automated Tiering  https://review.openstack.org/28705705:58
*** haomaiwang has quit IRC06:01
*** haomaiwang has joined #openstack-swift06:01
*** rminmin has joined #openstack-swift06:02
*** silor has joined #openstack-swift06:05
*** silor1 has joined #openstack-swift06:10
*** mingdang1 has quit IRC06:10
*** mingdang1 has joined #openstack-swift06:11
*** silor has quit IRC06:12
*** silor1 is now known as silor06:12
*** mingdang1 has quit IRC06:16
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Imported Translations from Zanata  https://review.openstack.org/29563406:16
*** ChubYann has quit IRC06:25
*** MVenesio has joined #openstack-swift06:41
*** MVenesio has quit IRC06:45
*** haomaiwang has quit IRC07:01
*** haomaiwang has joined #openstack-swift07:01
*** nadeem has quit IRC07:16
*** tesseract has joined #openstack-swift07:28
*** tesseract is now known as Guest9125107:28
*** McMurlock1 has joined #openstack-swift07:33
*** rcernin has joined #openstack-swift07:39
*** mingdang1 has joined #openstack-swift07:41
*** asettle has quit IRC07:45
*** proteusguy has quit IRC07:45
openstackgerritMerged openstack/swift: Imported Translations from Zanata  https://review.openstack.org/29563407:47
*** daemontool__ has quit IRC07:50
*** asettle has joined #openstack-swift07:57
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-swift08:01
*** asettle has quit IRC08:02
*** rledisez has joined #openstack-swift08:06
mingdang1@kota_ I install gf-complete,the source code README is said to "./configure",but it don't have ./configure? how can i do that08:21
kota_'bash autogen.sh'08:22
kota_mingdang1: the script via autoconf tools to generate configure file08:22
mingdang1ok ,thanks08:23
mingdang1Can't exec "aclocal": No such file or directory at /usr/share/autoconf/Autom4te/FileUtils.pm line 326.08:24
mingdang1autoreconf: failed to run aclocal: No such file or directory08:24
kota_Did you install automake/autoconf tools?08:25
mingdang1ok , i should install automake first08:26
*** permalac has quit IRC08:33
*** openstackgerrit has quit IRC08:33
*** openstackgerrit has joined #openstack-swift08:34
*** daemontool has joined #openstack-swift08:34
*** pcaruana has joined #openstack-swift08:40
*** geaaru has joined #openstack-swift08:41
mingdang1@kota_ I have install the libJerasure.so,but log get error "dynamic linking error libJerasure.so: cannot open shared object file"08:41
*** MVenesio has joined #openstack-swift08:41
mingdang1should I setup someing in local?08:41
kota_mingdang1: can you check the binary in /usr/loca/lib/libJerasure.so ?08:42
kota_mingdang1: is there/08:42
mingdang1yes ,it's there08:43
mingdang1[root@controller ~]# cd /usr/local/lib08:43
mingdang1[root@controller lib]# ls libJerasure.so08:43
kota_ok, install looks finished successfully.08:43
kota_could you try 'sudo ldconfig'08:43
mingdang1yes , i try it08:44
kota_could you try to start Swift again?08:44
mingdang1yes ,I restart swift08:45
kota_mingdang1: does it works?08:45
mingdang1no ,log get same error08:45
kota_maybe you need to add /usr/local/lib as seachable path.08:46
*** MVenesio has quit IRC08:46
mingdang1where is the searchable path08:46
kota_add /usr/local/lib to /etc/ld.so.conf08:47
mingdang1it has one line "include ld.so.conf.d/*.conf"08:47
mingdang1i only add /usr/local/lib to it ?08:48
kota_yup, not replace just add08:48
*** jmccarthy has quit IRC08:48
kota_and then, sudo ldconfig again.08:48
mingdang1add "/usr/local/lib" or "include /usr/local/lib"08:49
kota_only "/usr/local/lib"08:49
*** jmccarthy has joined #openstack-swift08:49
kota_"include" means the ld process search another config in the dir.08:50
kota_(i.e. it will search ls.so.conf.d/*.conf in default)08:50
mingdang1nice,it works fine08:51
kota_mingdang1: great!08:51
mingdang1I need to repeat this to other storage node08:51
*** jordanP has joined #openstack-swift08:52
kota_mingdang1: that's too bad :/08:52
*** geaaru has quit IRC08:52
mingdang1other node is also need libJerasure.so08:53
mingdang1I am in test env08:53
*** ho_ has quit IRC08:53
*** daemontool has quit IRC08:57
*** geaaru has joined #openstack-swift08:58
*** haomaiwang has quit IRC09:01
*** 92AAALWH6 has joined #openstack-swift09:01
*** mrmoje has joined #openstack-swift09:02
openstackgerritMerged openstack/swift: Set backend content length for fallocate - EC Policy  https://review.openstack.org/28257809:13
mahaticcschwede: you're on a roll. since yesterday, I had two patches which i was reviewing and the moment i wanted to drop a review, it's already merged :) nice work!09:24
cschwedemahatic: Hello! Thx - yes, I want to make sure the priority reviews are getting done before the release. Just writing another review ;)09:25
mahaticcschwede: great :)09:26
cschwedemahatic: but then, sorry to „steal“ your review time. you can always add comments even if it is merged, especially if you think something looks questionable09:28
cschwedefor today i’m done with reviews ;)09:28
cschwedeif anyone has some spare review cycles - i would be happy to see patch 257502 landed in Mitaka09:30
patchbotcschwede: https://review.openstack.org/#/c/257502/ - swift - Fix full_listing in internal_client09:30
*** dmorita has joined #openstack-swift09:30
cschwedethat’s affecting swift-dispersion-report with more than 10.000 entries (which is sometimes useful if you run dispersion with 100% coverage)09:30
mahaticcschwede: i agree and oh it's not stealing at all, it's still learning for me. glad to know you're done for today, i can slowly chew on the ones that aren't ;)09:31
mingdang1@kota_ it all work fine.how can i upload a object use EC policy09:31
*** takashi has quit IRC09:31
kota_mingdang1: create EC policy container via X-Storage-Policy: <your EC policy name> header.09:32
kota_probably ec10409:32
mingdang1like this "swift upload contain2 start.sh -H "X-Storage-Policy: ec104""09:33
*** dmorita has quit IRC09:34
mingdang1get message 'put_object(u'eccontainer', u'start.sh', ...) failure and no ability to reset contents for reupload'09:35
kota_mingdang1: though I'm not so familier with swift command but "swift post container2 -H 'X-Storage-Policy: ec104'" seems safe09:35
mingdang1yes, i use 'swift post eccontainer  -H "X-Storage-Policy: ec104"' create a container succeessful09:36
mingdang1but when i upload a object to this contaner09:36
mingdang1it get 'put_object(u'eccontainer', u'start.sh', ...) failure and no ability to reset contents for reupload.'09:37
*** Jeffrey4l has joined #openstack-swift09:37
kota_please remove the X-Storage-Policy header when uploading the object.09:37
kota_it might be a bug, though.09:37
mingdang1this 'swift upload eccontainer start.sh'09:38
mingdang1i do this,but get same error09:38
kota_wait a minute, I am trying to do this09:39
kota_mingdang1: looks work well in my env.09:43
kota_mingdang1: how does it works via 'swift stat eccontainer'?09:43
mingdang1@kota_ it works well09:50
mingdang1 Account: AUTH_98fe1a2907e54d9592886688079ad47e09:51
mingdang1       Container: eccontainer09:51
mingdang1         Objects: 009:51
mingdang1           Bytes: 009:51
mingdang1        Read ACL:09:51
mingdang1       Write ACL:09:51
mingdang1         Sync To:09:51
mingdang1        Sync Key:09:51
mingdang1   Accept-Ranges: bytes09:51
mingdang1X-Storage-Policy: ec10409:51
mingdang1     X-Timestamp: 1458639284.1685309:51
mingdang1      X-Trans-Id: tx9ad589b0d4794c60ae049-0056f1156609:51
mingdang1    Content-Type: text/plain; charset=utf-809:51
kota_mingdang1: ok, looks like you ware able to create the container successfully09:51
kota_maybe it's time to see proxy log09:52
mingdang1proxy log get nothing error or warning09:53
kota_ah, really?09:53
kota_mingdang1: can you make sure what's error happend during the upload?09:55
mingdang1i can use --debug09:55
mingdang1@kota_ http://pastebin.com/Lwd3ziL609:58
*** 92AAALWH6 has quit IRC10:01
*** 14WAAJIS3 has joined #openstack-swift10:01
kota_DEBUG:requests.packages.urllib3.connectionpool:"PUT /v1/AUTH_98fe1a2907e54d9592886688079ad47e/eccontainer/start.sh HTTP/1.1" 503 11810:01
*** rminmin has quit IRC10:02
kota_looks like your PUT failed so probably proxy dumps something into the log10:02
mingdang1i see /var/log/message, it have nothing ..10:04
kota_mingdang1: did you try to change the log place today?10:11
kota_i thought I saw in irc...10:11
mingdang1i change log faild10:13
mingdang1@kota_  proxy log get "Mar 22 18:19:27 controller proxy-server: Object PUT returning 503, 1/11 required connections (txn: tx86e28ead9a134b93ba903-0056f11c2f) (client_ip:"10:20
kota_mingdang1: looks like you added only 1 drive for that?10:21
kota_got it10:21
kota_you need 14 drives (at least) for EC if you set data=10, parity=4.10:22
kota_swift will attempt to store 14 actual data stored into object-severs uniquely.10:23
kota_the 14 data (we call fragments) are different each other.10:23
mingdang1ok, i try to add others10:24
kota_but... looks odd because 14 replica ring and only 1 device seems failed at rebalance...10:24
*** acoles_ is now known as acoles10:25
kota_mingdang1: if it's hard to set many drives, you can choose the parameter like data=2, parity=1 (Note that you should set this only testing deployment) for setting up.10:26
kota_time to back home.10:31
kota_see you tommorow10:32
*** 14WAAJIS3 has quit IRC10:36
*** kei_yama has quit IRC10:36
*** 7F1AAKFVO has joined #openstack-swift10:36
*** 7F1AAKFVO has quit IRC10:36
*** haomaiwang has joined #openstack-swift10:36
*** haomaiwang has quit IRC10:37
mingdang1see you10:39
*** MVenesio has joined #openstack-swift10:42
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: WIP: Swift Automated Tiering  https://review.openstack.org/28705710:44
*** MVenesio has quit IRC10:47
*** asettle has joined #openstack-swift10:56
*** mrmoje has quit IRC10:57
*** mvk_ has joined #openstack-swift11:12
*** daemontool has joined #openstack-swift11:13
*** mvk has quit IRC11:15
openstackgerritAlistair Coles proposed openstack/swift: Remove unused code from container sync  https://review.openstack.org/29572611:15
*** km has quit IRC11:19
*** mingdang1 has quit IRC11:26
*** Zyric has quit IRC11:35
*** baojg has quit IRC11:41
openstackgerritAlistair Coles proposed openstack/swift: Docs: Container sync does not require POST-as-COPY  https://review.openstack.org/29573911:47
*** div_ has joined #openstack-swift11:49
div_hi swift11:49
div_issue facing during devstack install11:49
div_openstack object store account set --property Temp-URL-Key=password 2016-03-21 09:36:49.425 | Not Found (HTTP 404)11:49
*** karthi_ has joined #openstack-swift11:50
karthi_Hi all11:50
karthi_i am trying to install mitaka devstack for ironic baremetal driver testing11:50
karthi_it is failing with switch error11:50
div_hello karthi11:51
karthi_this is the error encountered during switch installation11:51
karthi_openstack object store account set --property Temp-URL-Key=password11:51
karthi_please anybody give some clue to fix this issue11:51
karthi_sorry guys *swift11:52
*** cdelatte has quit IRC11:52
openstackgerritAlistair Coles proposed openstack/swift: Add .eggs/* to .gitignore  https://review.openstack.org/29574311:55
openstackgerritMerged openstack/swift: Container sync nodes shuffle cleanup  https://review.openstack.org/29407711:59
*** mvk_ has quit IRC12:07
karthi_guys any suggestion?12:08
*** cdelatte has joined #openstack-swift12:13
*** delattec has joined #openstack-swift12:18
*** MVenesio has joined #openstack-swift12:20
*** delattec has quit IRC12:21
*** cdelatte has quit IRC12:21
*** cbartz has joined #openstack-swift12:31
*** wanghua has quit IRC12:44
*** rvasilets__ has joined #openstack-swift12:45
*** mmcardle has quit IRC12:48
*** links has quit IRC13:01
*** daemontool has quit IRC13:07
*** daemontool has joined #openstack-swift13:09
*** Jeffrey4l has quit IRC13:25
*** mmcardle has joined #openstack-swift13:27
*** mvk_ has joined #openstack-swift13:27
*** rickyrem has joined #openstack-swift13:30
*** sgundur has joined #openstack-swift13:35
*** sgundur has left #openstack-swift13:37
*** sgundur has joined #openstack-swift13:38
*** ametts has joined #openstack-swift13:44
*** mingdang1 has joined #openstack-swift13:46
*** daemontool has quit IRC13:48
*** daemontool has joined #openstack-swift13:54
*** karthi_ has quit IRC14:00
*** asettle has quit IRC14:04
*** links has joined #openstack-swift14:05
*** baojg has joined #openstack-swift14:05
openstackgerritTakashi Kajinami proposed openstack/swift: Remove useless checking about broker instance  https://review.openstack.org/29585514:21
*** div_ has quit IRC14:27
*** mingdang1 has quit IRC14:28
*** openstackgerrit has quit IRC14:33
*** baojg has quit IRC14:34
*** openstackgerrit has joined #openstack-swift14:34
*** mrmoje has joined #openstack-swift14:34
*** baojg has joined #openstack-swift14:35
*** baojg has quit IRC14:36
*** asettle has joined #openstack-swift14:36
openstackgerritTakashi Kajinami proposed openstack/swift: Remove useless checking about broker instance  https://review.openstack.org/29585514:41
*** asettle has quit IRC14:41
*** links has quit IRC14:42
*** takashi has joined #openstack-swift14:43
*** rvasilets__ has quit IRC14:46
*** rickyrem has quit IRC14:47
jrichlicschwede: thanks for the review and merge!14:48
openstackgerritTakashi Kajinami proposed openstack/swift: Remove useless checking about a broker instance in container-sync  https://review.openstack.org/29585514:48
openstackgerritTakashi Kajinami proposed openstack/swift: Pass logger instances to AccountBroker/ContainerBroker  https://review.openstack.org/29587514:48
claygcschwede: I think I could get behind rsync_timeout + 900 for the auditor tunable, but the rsync_timeout config option would have to move to default I think - or else the auditor's config dict wouldn't see it?14:52
*** baojg has joined #openstack-swift14:53
*** rickyrem has joined #openstack-swift14:57
*** baojg has quit IRC15:05
*** klrmn has joined #openstack-swift15:07
openstackgerritMerged openstack/swift: Remove unused code from container sync  https://review.openstack.org/29572615:16
*** garthb has joined #openstack-swift15:17
notmynamegood morning15:20
openstackgerritTakashi Kajinami proposed openstack/swift: [WIP] Pass logger instances to AccountBroker/ContainerBroker  https://review.openstack.org/29587515:23
takashinotmyname: good morning15:23
*** _JZ_ has joined #openstack-swift15:40
*** NM has joined #openstack-swift15:44
*** cbartz has quit IRC15:47
openstackgerritAlistair Coles proposed openstack/swift: EC: Avoid conflicts when ssync'ing fragments  https://review.openstack.org/18140715:55
*** kevinc has joined #openstack-swift16:11
kevincif i run a proxy process with swauth auth on port 8080 and another proxy process with keystone auth on port 8081 on the same server, can I use the same memcache for both? or do I have to run multiple  memcache processes?16:13
*** vinsh has joined #openstack-swift16:17
notmynameI don't see why the same memcache wouldn't work for both. swift namespaces all of the values it puts into memcache16:17
*** takashi has quit IRC16:26
*** dmorita has joined #openstack-swift16:29
*** Jeffrey4l has joined #openstack-swift16:30
*** lyrrad has joined #openstack-swift16:32
acolesnotmyname: you don't get .eggs ?? interesting16:34
acolesnotmyname: do you have html5lib installed ?16:37
acolesi don't (didn't) and that seems to be why i got .eggs/16:37
acolesdoes anyone else see a .eggs/ dir created when they run 'python setup.py build_sphinx' ?16:38
acolesor, does anyone run 'python setup.py build_sphinx' ? :)16:39
*** geaaru has quit IRC16:39
*** Zyric has joined #openstack-swift16:43
*** haomaiwang has joined #openstack-swift16:50
tdasilvaacoles: btw i wasn't able to duplicate that either16:51
tdasilvaacoles: no .eggs dir16:52
*** pcaruana has quit IRC16:52
*** joeljwright has joined #openstack-swift16:52
*** ChanServ sets mode: +v joeljwright16:52
*** joeljwright has quit IRC16:52
acolestdasilva: thanks. it seems i just needed html5lib to be installed16:52
*** joeljwright has joined #openstack-swift16:53
*** ChanServ sets mode: +v joeljwright16:53
*** rickyrem has quit IRC16:53
acolesjoeljwright: you're back?16:53
joeljwrightI am now16:53
joeljwright48 hours to get home after the air traffic controller strikes16:53
joeljwrighthad to take a bus from Toulouse -> Gatwick in the end :(16:54
joeljwrighthow can I help?16:54
*** rickyrem has joined #openstack-swift16:54
acolesjoeljwright: oh, did the snow make it worthwhile?16:54
joeljwrightthe snow was great :)16:54
*** nadeem has joined #openstack-swift16:54
acolesjoeljwright: don't need any help, just saying hi16:55
*** haomaiwang has quit IRC16:55
joeljwrightacoles: :)16:55
joeljwrightit is nice to be back home16:55
acolesjoeljwright: some people pay good money for a bus tour of France ;)16:56
kevincnotmyname: thank you!16:56
joeljwrightacoles: we took in all the sights, the Paris Peripherique at rush hour…, 3 service stations…16:56
*** nadeem has quit IRC16:56
joeljwrightacoles: the Calais jungle and fences16:56
claygacoles: cschwede: i'm so sorry to hear about Brussels - sending some love across the pond16:56
*** nadeem has joined #openstack-swift16:57
joeljwrightclayg: thanks! was quite a shock to wake up to that!16:57
* acoles has been in a bubble and just checked news... :(16:58
acolesjoeljwright: hmmm...17:00
acolesclayg: thanks.17:01
*** Guest91251 has quit IRC17:02
acolesclayg: i'm looking at patch 293177. do the rsync temp files ever get re-used after rsync fails? i have been reading around to try to understand if a subsequent rsync finds the temp file and uses its content. or is it just garbage.17:03
patchbotacoles: https://review.openstack.org/#/c/293177/ - swift - Auditor will clean up stale rsync tempfiles17:03
notmynamea1|away: do you want to add the .gitignore entry or not. if so, I'll approve the patch.17:13
*** McMurlock1 has quit IRC17:14
notmynameacoles:  do you want to add the .gitignore entry or not. if so, I'll approve the patch.17:17
acolesnotmyname: it does no harm and may save someone else needing to figure out why it happens17:17
acolesnotmyname: or save us documenting that html5lib needs to be installed, so yes.17:18
acolesnotmyname: it was one of those "this has been bugging me for months" moments !17:19
notmynamehmmm...I don't seem to have that installed17:19
*** daemontool has quit IRC17:22
*** klamath has joined #openstack-swift17:23
klamathi can have multiple storage policies on the same disks correct?17:23
notmynameyes, absolutely17:23
klamathawesome, thank you17:23
*** mingdang1 has joined #openstack-swift17:29
*** rledisez has quit IRC17:32
*** mingdang1 has quit IRC17:33
klamathfor multi-region replication is it possible to only write to the primary nodes and not any handoff nodes?17:37
notmynameassuming you don't have any hardware failures at the time of the request?17:38
klamathwith write affinity on17:39
notmynamethen yes. the default behavior (which includes write affinity turned off) is to write only to the primary nodes when there are no failures in the cluster17:39
notmynameby turning write affinity on, you are specifically asking swift to write to handoffs instead of primary locations17:40
*** daemontool has joined #openstack-swift17:42
*** arch-nemesis has joined #openstack-swift17:42
*** pcaruana has joined #openstack-swift17:44
notmynameitlinux just informed me that the italian translation for swift is now 100% complete17:45
claygacoles: just garbage as best I can tell as a new .tmp file is created after the failure if you set your rsync_timeout = 117:49
*** jordanP has quit IRC17:49
clayg... at least on largish files - it's a bit of a race obviously17:49
*** _JZ_ has quit IRC17:49
*** cathR has joined #openstack-swift17:55
*** cathR has left #openstack-swift17:55
clayggah, i'm still not sure how to handle the rsync_timeout config option :\17:56
claygit seems like the simplest thing to do is add a new auditor section option called "clean_rsync_tempfiles" and suggest that it be larger than the replicators rsync_timeout - with a default of say 30 mins 1800?17:57
*** d0ugal has quit IRC17:58
claygI agree with cschwede that it'd be nice if the auditor could just do the right thing based on the replicators rsync_timeout - but I'd have to re-parse the config unless we move it to default17:58
claygbut it sure does seem silly to add another knob - gah!  it's very vexxing17:59
*** d0ugal has joined #openstack-swift17:59
*** dmorita has quit IRC17:59
notmynameclayg: https://s-media-cache-ak0.pinimg.com/736x/ed/12/62/ed1262fa244b0f6f3fd3d61b46526ce0.jpg ?17:59
*** dmorita has joined #openstack-swift18:01
*** silor has quit IRC18:01
*** kevinc has quit IRC18:02
*** silor has joined #openstack-swift18:04
acolesclayg: k. thanks.18:04
*** panda has quit IRC18:04
*** panda has joined #openstack-swift18:05
acolesclayg: why did you choose the auditor to do the cleanup and not do it in cleanup_ondisk_files?18:05
notmynameitlinux: how can I check on the state of translations for swift?18:05
* acoles has feeling that might be a stupid question :/18:05
acolesnotmyname: oops - *my* question not yours ;)18:06
notmynameI didn't even read it that way :-)18:06
*** d0ugal has quit IRC18:09
claygacoles: lol at "phew" - so I didn't think it appropriate to add all those stat calls into PUT paths?18:12
claygacoles: I definately considered it?  Obvs. a judgement call - you think i was off base?18:13
acolesclayg: no, no i am just playing catch up with the thinking18:13
claygi get lucky sometimes18:15
acolesclayg: um, yeah, cleanup_ondisk_files isn't a good place18:15
*** ChubYann has joined #openstack-swift18:16
claygacoles: idk, if you say so18:19
*** daemontool has quit IRC18:25
*** joeljwright has quit IRC18:25
acolesclayg: would it be silly to use reclaim_age as condition to clean up the temp files rather than rsync_timeout+whatever? just wondering if it helps with the config conundrum? and I guess reclaim_age should be > rsync_timeout18:26
*** chsc has joined #openstack-swift18:28
claygacoles: yeah it would definately be longer but we're talking week(s) instead of minutes in the default case!18:28
claygacoles: I think 1 hr rsync and 2 hr rsync tempfile would be closer to reasonable - but I like reclaim age being bigish - i worry more about some downed node coming to life and spewing .data files back into the cluster after being offline for 10 days than I do having some tombstones pile up for a few weeks18:29
*** rickyrem has quit IRC18:30
claygacoles: i suppose it depends on how much churn you have in your cluster18:30
claygi generally assume the life time of .data file is more than a few weeks - so keeping a tombstone around for a month or more is no biggy18:30
openstackgerritMerged openstack/swift: Add .eggs/* to .gitignore  https://review.openstack.org/29574318:30
claygthe wrose issue with tombstones piling up is that inactive suffixes don't have their hash_suffix_cleanup called and we have tombstones that are many many months past reclaim age - but there's another open bug for that18:31
*** rickyrem has joined #openstack-swift18:31
acolesclayg: agree.18:31
acolesclayg: yes i have that bug on my radar, but its a distant blip right now18:31
claygacoles: IKR!18:31
acolesidk, i was just sharing your pain over adding a new config option that relies on user setting it to other option plus something18:32
claygthere's another one about pushing up object updates (async_pending) after a container has been delted - that... well someone had a patch - so that's more on my radar that the other broken that doesn't have a change sitting in the queue!18:32
claygacoles: IKR!  there has to be a better way!18:32
clayggawd why am I so stupid i can't think of it18:33
claygredbo_: you're a ton smarter than most of us - what do you guys set rsync_timeout too - how do you think we should plumb that value into the auditor so that it can cleanup the garbage files!?18:33
acolesclayg: gotta go, will think som emore and finish up review on that patch tomorrow18:34
*** joeljwright has joined #openstack-swift18:34
*** ChanServ sets mode: +v joeljwright18:34
*** StraubTW has joined #openstack-swift18:35
*** acoles is now known as acoles_18:36
*** joeljwright has quit IRC18:36
*** tongli has joined #openstack-swift18:37
*** dmorita has quit IRC18:38
*** vint_bra has quit IRC18:45
*** vint_bra has joined #openstack-swift18:46
*** dmorita has joined #openstack-swift18:47
*** haomaiwang has joined #openstack-swift18:49
*** haomaiwang has quit IRC18:54
*** dmorita has quit IRC18:57
claygoh oh oh i think i have an idea!18:57
klrmnuh oh, clayg has an idea… =)19:00
* klrmn couldn't resist19:00
*** dmorita has joined #openstack-swift19:02
*** chsc has quit IRC19:05
*** rickyrem has quit IRC19:07
redbo_clayg: who's using rsync anymore?19:07
*** redbo_ is now known as redbo19:07
*** ChanServ sets mode: +v redbo19:07
*** mrmoje has quit IRC19:10
*** aswadr_ has quit IRC19:10
claygredbo: good one!19:11
*** mrmoje has joined #openstack-swift19:12
*** ametts has quit IRC19:17
*** stantonnet has quit IRC19:25
*** dmorita has quit IRC19:28
*** mrmoje has quit IRC19:33
*** joeljwright has joined #openstack-swift19:37
*** ChanServ sets mode: +v joeljwright19:37
*** stantonnet has joined #openstack-swift19:40
openstackgerritMerged openstack/swift: Make the object auditor's run-once mode run once.  https://review.openstack.org/21644919:47
notmynameoh cool. that's something to add to the changelog :-)19:51
*** rickyrem has joined #openstack-swift19:55
*** rickyrem has quit IRC19:56
*** david-lyle_ has joined #openstack-swift19:56
*** swifterdarrell has quit IRC19:57
claygyay auto config!19:57
*** david-lyle has quit IRC19:57
*** swifterdarrell has joined #openstack-swift19:58
*** ChanServ sets mode: +v swifterdarrell19:58
*** ekarlso- has quit IRC19:58
*** jeblair has quit IRC19:58
*** jeblair has joined #openstack-swift19:58
*** jaakkos has quit IRC19:58
*** jlvillal has quit IRC19:58
*** tanee has quit IRC19:58
openstackgerritClay Gerrard proposed openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317719:59
claygdoes everyone really wrap their commits at 70 chars?  remote: (W) b06ccd4: too many commit message lines longer than 70 characters; manually wrap lines19:59
claygpeople gripe so much about 80 in python - and they take off another 10 chars in commit messages?  I do try to generally wrap long blocky text at 72 chars - but 70 seems excessive?20:00
clayg... still, I don't want to be some old curmudgeon if all the kids have their editors configured to wrap commit messages at 70 chars...20:00
*** david-lyle_ is now known as david-lyle20:02
notmynameclayg: where are you seeing that error?20:03
*** jlvillal has joined #openstack-swift20:03
*** tanee has joined #openstack-swift20:04
claygfrom `git review` - maybe I need to upgrade or something?20:05
notmynameoh, interesting20:06
*** dmorita has joined #openstack-swift20:06
*** dmorita has quit IRC20:06
*** dmorita has joined #openstack-swift20:06
timburkeclayg: i think it's coming from the server. so, more something that *openstack* wants rather than swift in particular20:09
*** ekarlso- has joined #openstack-swift20:11
*** jaakkos has joined #openstack-swift20:18
*** ev1lkry has joined #openstack-swift20:18
*** mvk_ has quit IRC20:18
openstackgerritClay Gerrard proposed openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317720:19
ev1lkryalright folks on my deep dive studying the COA certification objectives - i'm looking for documentation as it related to "Monitor space available for Object Storage" can anyone point me to best practices for this?20:20
notmynamewhat is COA?20:22
*** joeljwright has quit IRC20:22
ev1lkryCertified OpenStack Administrator through the OpenStack foundation20:23
notmynameuse swift-recon as a start. then standard ops tools like mount and df20:26
ev1lkrythank you notmyname I will take a look at swift-recon20:27
*** jmccarthy1 has joined #openstack-swift20:31
claygev1lkry: swift-recon is fine, but something like standard system metrics - df from collectd is probably more popular/useful - memeory and cpu utilization too - all the stuff you monitor in any distributed system you monitor in swift too20:46
ahalesometimes its useful to measure usage in a few ways, like what alll the account databses think is used , and what df/recon says.. but yeah its just standard stuff20:48
claygev1lkry: ahale makes a good point - once nice thing that recon gets your for free is correlation of your systems disk usage with their ring assignments - if you're just dumping everything into graphite you have to have something to groups your account/container server disk metrics together and seperate from your object metrics20:51
*** asettle has joined #openstack-swift20:53
ev1lkrygreat points thanks guys20:53
*** rickyrem has joined #openstack-swift21:03
*** haomaiwa_ has joined #openstack-swift21:08
notmynamecurrent fire: the openstack-tc wants to strip swift of the "supports rolling upgrade" tag (https://review.openstack.org/#/c/292334/). While I think this is patently ridiculous since swift has supported this for over 6 years now, the result is that swift will get further dinged on it's "maturity" score listed on https://www.openstack.org/software/releases/liberty/components/swift21:11
patchbotnotmyname: patch 292334 - governance - Remove rolling-upgrade tag from swift/ceilometer21:11
notmynameso the reasoning for removing the tag is "swift doesn't have a gate job that runs tests against a partially upgraded cluster"21:11
*** haomaiwa_ has quit IRC21:12
notmynamewe've got probetests that run against every commit, but that's on the community qa cluster, not in the -infra gate queue21:14
ahalejust change how defcore defines swift, make it about the api and include 3rd party ones like ceph rather than designated code :)21:28
ahalethen you'll easily get that adoption point back21:28
*** NM has quit IRC21:31
*** nadeem has quit IRC21:32
*** MVenesio has quit IRC21:33
*** mrmoje has joined #openstack-swift21:33
*** silor has quit IRC21:34
*** rickyrem has quit IRC21:38
*** CaioBrentano has quit IRC21:39
*** CaioBrentano has joined #openstack-swift21:39
*** daemontool has joined #openstack-swift21:41
*** tongli has quit IRC21:42
*** openstackgerrit has quit IRC21:48
*** openstackgerrit has joined #openstack-swift21:49
*** panda has quit IRC22:04
*** panda has joined #openstack-swift22:05
*** vint_bra has quit IRC22:09
*** hogepodge has quit IRC22:12
*** darrenc is now known as darrenc_afk22:14
*** dmorita has quit IRC22:17
*** dmorita has joined #openstack-swift22:21
*** dmorita has quit IRC22:22
*** dmorita has joined #openstack-swift22:22
*** rickyrem has joined #openstack-swift22:29
*** darrenc_afk is now known as darrenc22:35
*** haomaiwang has joined #openstack-swift22:39
*** hogepodge has joined #openstack-swift22:39
*** NM has joined #openstack-swift22:43
*** haomaiwang has quit IRC22:43
*** km has joined #openstack-swift22:59
*** StraubTW has quit IRC23:18
*** NM1 has joined #openstack-swift23:20
*** NM has quit IRC23:22
*** Jeffrey4l has quit IRC23:23
*** NM1 has quit IRC23:24
*** NM has joined #openstack-swift23:24
*** rickyrem has quit IRC23:24
*** kei_yama has joined #openstack-swift23:31
*** MVenesio has joined #openstack-swift23:33
kota_good morning23:34
*** MVenesio has quit IRC23:38
notmynamehttps://gist.github.com/notmyname/4081362a8f20ca28e61d is the minimal config I can use for a proxy server so that it passes ./.functests. that's based on very very early versions of the proxy server config found in git (specifically, from september 1, 2010)23:49
notmynameI had to add/change the auth to tempauth. I had to add account_autocreate=True. I had to explicitly set bind_port. but that's it23:49
notmynamethe impetus for this was to find the oldest version of code where configs still worked to see about a rolling upgrade test23:50

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!