| ho | good morning! | 00:00 |
|---|---|---|
| *** esker has quit IRC | 00:00 | |
| *** haomaiwang has quit IRC | 00:01 | |
| *** haomaiwang has joined #openstack-swift | 00:01 | |
| mattoliverau | notmyname: lol | 00:01 |
| mattoliverau | notmyname: at least you tried | 00:01 |
| *** ianbrown__ has quit IRC | 00:10 | |
| *** ianbrown_ has joined #openstack-swift | 00:10 | |
| *** ianbrown_ has quit IRC | 00:16 | |
| *** ianbrown_ has joined #openstack-swift | 00:17 | |
| *** Jeffrey4l has quit IRC | 00:17 | |
| *** dmorita has quit IRC | 00:19 | |
| *** dmorita has joined #openstack-swift | 00:20 | |
| *** ianbrown_ has quit IRC | 00:21 | |
| *** ianbrown_ has joined #openstack-swift | 00:22 | |
| *** dslevin has joined #openstack-swift | 00:23 | |
| *** dmorita_ has joined #openstack-swift | 00:23 | |
| *** dmorita has quit IRC | 00:24 | |
| *** dslevin has quit IRC | 00:27 | |
| *** Guest35433 has quit IRC | 00:29 | |
| *** km has joined #openstack-swift | 00:29 | |
| *** km is now known as Guest6019 | 00:29 | |
| *** pgbridge has quit IRC | 00:38 | |
| *** thumpba has joined #openstack-swift | 00:45 | |
| *** thumpba has quit IRC | 00:45 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Add X-Delete-After and X-Delete-At functional test https://review.openstack.org/255626 | 00:47 |
| *** chlong has joined #openstack-swift | 00:52 | |
| *** haomaiwang has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-swift | 01:01 | |
| *** dslevin has joined #openstack-swift | 01:17 | |
| *** dslevin has quit IRC | 01:21 | |
| *** ianbrown_ has quit IRC | 01:32 | |
| *** ianbrown_ has joined #openstack-swift | 01:32 | |
| *** david-lyle has joined #openstack-swift | 01:37 | |
| *** dmorita_ has quit IRC | 01:38 | |
| *** dmorita has joined #openstack-swift | 01:40 | |
| *** zhill has quit IRC | 01:41 | |
| *** janonymous_ has quit IRC | 01:58 | |
| *** noark9 has joined #openstack-swift | 02:00 | |
| *** haomaiwang has quit IRC | 02:01 | |
| *** haomaiwang has joined #openstack-swift | 02:01 | |
| *** jamielennox is now known as jamielennox|away | 02:01 | |
| *** ianbrown_ is now known as ianbrown | 02:05 | |
| *** Jeffrey4l has joined #openstack-swift | 02:07 | |
| *** lyrrad has quit IRC | 02:08 | |
| *** ianbrown has quit IRC | 02:15 | |
| *** ianbrown has joined #openstack-swift | 02:15 | |
| *** petertr7_away is now known as petertr7 | 02:17 | |
| *** dmorita has quit IRC | 02:23 | |
| *** dmorita has joined #openstack-swift | 02:26 | |
| *** jamielennox|away is now known as jamielennox | 02:29 | |
| *** dewsday has quit IRC | 02:57 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| *** sanchitmalhotra has joined #openstack-swift | 03:03 | |
| *** bill_az has joined #openstack-swift | 03:12 | |
| openstackgerrit | Merged openstack/swift: Unification of manpages and conf-samples (default values, etc) https://review.openstack.org/251396 | 03:20 |
| openstackgerrit | hgangwx@cn.ibm.com proposed openstack/swift: Fix some inconsistency in docstrings https://review.openstack.org/256197 | 03:38 |
| *** links has joined #openstack-swift | 03:42 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-swift | 04:01 | |
| *** trifon has joined #openstack-swift | 04:01 | |
| openstackgerrit | Takashi Kajinami proposed openstack/swift: Fix proxy-server's support for chunked transferring in GET object https://review.openstack.org/256201 | 04:04 |
| *** trifon has quit IRC | 04:07 | |
| *** trifon has joined #openstack-swift | 04:10 | |
| mahatic | notmyname: hi. Is it worth updating that, after DELETE, the older version is gone from versions container and back in the actual container, indicating with a curl command? Here: http://docs.openstack.org/developer/swift/overview_object_versioning.html#examples-using-curl | 04:13 |
| mahatic | it just says the older version is gone, and I wondered why until I checked that it is back in the (original) container after a DELETE | 04:14 |
| *** trifon has quit IRC | 04:15 | |
| *** trifon has joined #openstack-swift | 04:25 | |
| *** yatin has joined #openstack-swift | 04:28 | |
| yatin | Hello Folks | 04:29 |
| yatin | I'm new to swift, read swift docs, got one query | 04:29 |
| yatin | after the ring datastructure decides which drive to find the object on..how does it go further from there? | 04:30 |
| *** ppai has joined #openstack-swift | 04:34 | |
| *** garthb has quit IRC | 04:34 | |
| *** esker has joined #openstack-swift | 04:41 | |
| *** proteusguy_ has quit IRC | 04:45 | |
| *** trifon has quit IRC | 04:46 | |
| *** sgundur has left #openstack-swift | 04:47 | |
| *** proteusguy_ has joined #openstack-swift | 04:58 | |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-swift | 05:01 | |
| *** esker has joined #openstack-swift | 05:02 | |
| *** dmorita has quit IRC | 05:03 | |
| *** dmorita has joined #openstack-swift | 05:07 | |
| *** esker has quit IRC | 05:08 | |
| notmyname | mahatic: yeah, it could be more specific | 05:13 |
| notmyname | yatin: still here? still wondering about how swift works? :-) | 05:14 |
| yatin | notmyname: yes.. having some in between gaps | 05:14 |
| mahatic | notmyname: okay, thanks. will update | 05:14 |
| notmyname | mahatic: cool. thanks | 05:14 |
| notmyname | yatin: where are you at now? | 05:14 |
| *** ig0r__ has joined #openstack-swift | 05:15 | |
| notmyname | (in your understanding) | 05:15 |
| yatin | notmyname: so, two datastructures , device dict, and replica2devpart | 05:16 |
| yatin | notmyname: the arrays of arrays stores device-ids | 05:16 |
| yatin | notmyname: this device id is key into devices dict, where it gets ip and drive name | 05:17 |
| notmyname | ring | 05:17 |
| yatin | notmyname: so, after that what happens? | 05:17 |
| notmyname | err...right | 05:17 |
| notmyname | ok, so you've gotten really deep into the ring sturcture | 05:17 |
| notmyname | *structure | 05:17 |
| notmyname | that's great | 05:18 |
| yatin | notmyname: not feeling that great yet, my understanding has landed me into service with specific drive, how to locate object? | 05:18 |
| notmyname | basically, all that is called from the ring class's get_nodes() and get_more_nodes() | 05:19 |
| notmyname | heh | 05:19 |
| yatin | notmyname: *server | 05:19 |
| *** ig0r_ has quit IRC | 05:19 | |
| notmyname | ok, backing up a little more | 05:19 |
| yatin | notmyname: ok. please | 05:19 |
| notmyname | given a path to something in swift, eg /account/container/object, that info (the account, container, and object parts) are passed to the ring | 05:20 |
| notmyname | the ring hashes those parts (along with some salts) and gets a hash value | 05:20 |
| yatin | notmyname: right | 05:20 |
| *** klrmn has quit IRC | 05:20 | |
| notmyname | some number of prefix bits of the resulting hash are used as the "partion" | 05:20 |
| notmyname | the number of bits is the part power number. or actually 2**part_power | 05:21 |
| yatin | notmyname: yes.. md5 bits - part_power.. some formula there | 05:21 |
| notmyname | so essentially, part = hasher(path_parts) << 2**part_power | 05:21 |
| yatin | notmyname: to get hash | 05:21 |
| notmyname | ring | 05:21 |
| *** ig0r__ has quit IRC | 05:21 | |
| notmyname | right | 05:22 |
| notmyname | sorry. muscle memory on my fingers ;-) | 05:22 |
| notmyname | ok, so you get a partition | 05:22 |
| notmyname | partitons are mapped to devices in the cluster | 05:22 |
| yatin | notmyname: :) yes | 05:22 |
| notmyname | so given a partition, you can find where that *should* be stored in the cluster | 05:23 |
| notmyname | and that mapping is The Ring (tm) | 05:23 |
| yatin | notmyname: correct | 05:23 |
| notmyname | with the various data structures you mentioned | 05:23 |
| notmyname | ok | 05:23 |
| yatin | notmyname: 2D array of [replica][partions] | 05:23 |
| notmyname | so the 2 key parts of the ring data structure are the mapping from partition to device (dev_id) and the mapping of device ids | 05:24 |
| *** esker has joined #openstack-swift | 05:24 | |
| *** venkatesh has joined #openstack-swift | 05:24 | |
| notmyname | a device is identified by (IP, port, mount point) | 05:24 |
| yatin | notmyname: true | 05:25 |
| notmyname | so eg if you hash /a/c/cat.jpg, and you get partition 42, partion 42 might be mapped to devices 3, 5, and 7 | 05:25 |
| notmyname | and device 3 is identified as 10.0.1.34:6001/sdf1 | 05:25 |
| notmyname | IP:port/mount | 05:26 |
| notmyname | and that's what's returned to the proxy when it calls the ring with a given URL (or the URL parts) | 05:26 |
| notmyname | so therefore the proxy connects to the given IP:port and sends the mountpoint and the URL and the object server translates that to the on-disk filename | 05:27 |
| notmyname | and, ta-da, you can read and write your objects | 05:27 |
| yatin | notmyname: so, so proxy-server sends ip:port/mount and object name too? | 05:28 |
| yatin | notmyname: because on sdf1, here may be many more objects? | 05:28 |
| notmyname | right | 05:28 |
| notmyname | in fact, many partitions will be assigned to whatever device is mounted ad sdf1 | 05:28 |
| yatin | notmyname: and could be from different containers too, | 05:29 |
| notmyname | absolutely | 05:29 |
| yatin | notmyname: two contrainer can have same objects name | 05:29 |
| notmyname | right | 05:29 |
| notmyname | and 2 different accounts can have the same contaienr name (with objects of the same name too) | 05:29 |
| yatin | notmyname: worst case, could go on same sdf1, | 05:30 |
| *** esker has quit IRC | 05:30 | |
| notmyname | no, that's ok. since the on-disk filename has to do with the hash of "account + contaienr + object" | 05:30 |
| yatin | notmyname: ook...now, its clear | 05:31 |
| yatin | notmyname: thanks a ton | 05:31 |
| venkatesh | Hi all, | 05:31 |
| notmyname | while not *exactly* related to your question, this is a link to a talk I gave a few years ago about how swift deals with failures. it shows some of the on-disk files and how that works. https://www.youtube.com/watch?v=_sUvfGKhaMo | 05:31 |
| venkatesh | is proxy-server sends account HEAD/GET/POST request to every account-server? | 05:32 |
| yatin | notmyname: great! thanks for the url | 05:32 |
| notmyname | it might provides some additional details for you | 05:32 |
| *** Trixboxer has quit IRC | 05:33 | |
| yatin | notmyname: this would really help further to gain deeper understanding. :) | 05:33 |
| notmyname | yatin: happy to help. | 05:33 |
| notmyname | yatin: it's getting late in my timezone, and I'm stepping away now, but please feel free to ask more questions. there's other pople online now, and I'll definitely be back on later | 05:34 |
| venkatesh | Hi all | 05:34 |
| venkatesh | Can someone tell | 05:35 |
| venkatesh | is proxy-server sends account HEAD/GET/POST request to every account-server? | 05:35 |
| *** SkyRocknRoll has joined #openstack-swift | 05:36 | |
| *** petertr7 is now known as petertr7_away | 05:36 | |
| yatin | notmyname: thanks. good night | 05:38 |
| *** Trixboxer has joined #openstack-swift | 05:42 | |
| *** esker has joined #openstack-swift | 05:46 | |
| notmyname | venkatesh: if by "every" you mean all servers in the cluster, it depends on how big your cluster is. but assuming you have a large cluster with many servers, no, requests are not sent to every server | 05:51 |
| *** esker has quit IRC | 05:52 | |
| venkatesh | notmyname: In saio setup, it goes to every server right? | 05:53 |
| venkatesh | *account | 05:54 |
| notmyname | the saio simulates 4 servers. if you have a 3 replica policy and you haven't failed any of the servers, then it will only go to the 3 servers that are handling the request | 05:54 |
| notmyname | for a write (eg PUT/POST) | 05:54 |
| *** bill_az has quit IRC | 05:55 | |
| notmyname | for a read (GET/HEAD), it will only go to one | 05:55 |
| venkatesh | notmyname: I assumed every server handles separate replica, is that correct | 05:57 |
| notmyname | yes | 05:57 |
| *** dmorita has quit IRC | 05:57 | |
| venkatesh | 4 servers, 3 replicas are possible? | 05:58 |
| venkatesh | How the account-db new entries are replicated across multiple account-servers? | 05:59 |
| venkatesh | from the code I understood as make_requests, which sends request to every account-server | 06:00 |
| *** haomaiwang has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-swift | 06:01 | |
| *** Trixboxer has quit IRC | 06:01 | |
| notmyname | venkatesh: you should talk to yatin, who was just going through this too :-) | 06:01 |
| venkatesh | Ok, Thanks for your help | 06:02 |
| venkatesh | yatin: | 06:02 |
| venkatesh | Hi | 06:02 |
| *** trifon has joined #openstack-swift | 06:03 | |
| yatin | venkatesh: Hi | 06:05 |
| venkatesh | 4 accountservers, 3 replicas are possible? | 06:07 |
| yatin | venkatesh: yes | 06:07 |
| venkatesh | yatin : can I repost the above 2 questions? | 06:07 |
| yatin | venkatesh: accountservers has nothing to do with replicas | 06:08 |
| venkatesh | account-server mainatins separate db's right? | 06:08 |
| yatin | venkatesh: i did read your question, as i'm new to swift, don't know what's the exact answer | 06:08 |
| yatin | venkatesh: but you can check api guide | 06:08 |
| yatin | venkatesh: yes.. | 06:09 |
| *** Trixboxer has joined #openstack-swift | 06:09 | |
| yatin | venkatesh: swift calls it metadata, use to get container list | 06:09 |
| venkatesh | here replica means I am saying account-db replica | 06:10 |
| yatin | venkatesh: it;s not replica, it has to distribute account ring to all servers which having accont-server running | 06:10 |
| yatin | venkatesh: use word replica for objects | 06:11 |
| venkatesh | Ok, | 06:11 |
| yatin | venkatesh: do you have setup, if not get swift AIO to begin with.. | 06:11 |
| venkatesh | I installed SAIO and devstack and separate VM's | 06:12 |
| venkatesh | I went through the swift code | 06:12 |
| yatin | venkatesh: nice | 06:12 |
| venkatesh | I will ask my exact doubt w.r.t code | 06:12 |
| yatin | venkatesh: I yet to go through code...i'm just new here...but you can shoot question | 06:13 |
| yatin | venkatesh: i will try | 06:13 |
| venkatesh | https://github.com/openstack/swift/blob/master/swift/proxy/controllers/account.py#L127 | 06:13 |
| venkatesh | this is sending the request to every account-server right? | 06:13 |
| *** chlong has quit IRC | 06:15 | |
| *** bill_az has joined #openstack-swift | 06:16 | |
| * yatin checking code | 06:16 | |
| venkatesh | Ok | 06:17 |
| yatin | venkatesh: yes.. you are right | 06:20 |
| venkatesh | Ok, thanks | 06:20 |
| *** ChubYann has quit IRC | 06:21 | |
| yatin | venkatesh: it gets unique devices, via using seen_ids | 06:22 |
| yatin | venkatesh: def _get_part_nodes() in ring.py | 06:23 |
| *** ppai is now known as ppai|afk | 06:30 | |
| venkatesh | Ok, Thanks, I will check | 06:31 |
| *** esker has joined #openstack-swift | 06:31 | |
| *** jamielennox is now known as jamielennox|away | 06:33 | |
| *** bill_az has quit IRC | 06:36 | |
| *** esker has quit IRC | 06:36 | |
| *** gyee has quit IRC | 06:38 | |
| *** bill_az has joined #openstack-swift | 06:43 | |
| *** silor has joined #openstack-swift | 06:53 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Update versioned_writes doc https://review.openstack.org/256250 | 06:57 |
| *** bill_az has quit IRC | 06:59 | |
| *** silor has quit IRC | 07:00 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** silor has joined #openstack-swift | 07:01 | |
| *** haomaiwa_ has joined #openstack-swift | 07:01 | |
| *** silor1 has joined #openstack-swift | 07:03 | |
| *** silor has quit IRC | 07:05 | |
| *** silor1 is now known as silor | 07:05 | |
| *** bill_az has joined #openstack-swift | 07:07 | |
| *** zhill has joined #openstack-swift | 07:07 | |
| *** SkyRocknRoll has quit IRC | 07:16 | |
| *** zaitcev has quit IRC | 07:22 | |
| *** takashi has joined #openstack-swift | 07:27 | |
| *** ig0r_ has joined #openstack-swift | 07:32 | |
| *** bill_az has quit IRC | 07:52 | |
| *** takashi has quit IRC | 07:57 | |
| *** haomaiwa_ has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| *** ppai|afk is now known as ppai | 08:04 | |
| *** arnox has joined #openstack-swift | 08:05 | |
| *** rledisez has joined #openstack-swift | 08:08 | |
| *** ig0r_ has quit IRC | 08:11 | |
| kota_ | takashi? | 08:12 |
| *** arnox has quit IRC | 08:15 | |
| *** esker has joined #openstack-swift | 08:18 | |
| *** ig0r_ has joined #openstack-swift | 08:19 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:20 | |
| *** esker has quit IRC | 08:23 | |
| *** arnox has joined #openstack-swift | 08:35 | |
| *** arnox has quit IRC | 08:40 | |
| *** esker has joined #openstack-swift | 08:40 | |
| *** arnox has joined #openstack-swift | 08:43 | |
| *** rcernin has joined #openstack-swift | 08:44 | |
| *** esker has quit IRC | 08:46 | |
| *** ig0r_ has quit IRC | 08:53 | |
| openstackgerrit | Peter Lisák proposed openstack/swift: Unification of manpages and swift-* --help https://review.openstack.org/251397 | 08:59 |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-swift | 09:01 | |
| openstackgerrit | Peter Lisák proposed openstack/swift: Change schedule priority of daemon/server in config https://review.openstack.org/238799 | 09:14 |
| *** acoles_ is now known as acoles | 09:19 | |
| *** ig0r_ has joined #openstack-swift | 09:19 | |
| *** Zyric1 has quit IRC | 09:24 | |
| mahatic | kota_: hi, would you happen to know if I should be updating something else for this patch (256250)? the rst picks up from doc string - I remember you mentioning about sphinx build last time on a doc patch | 09:32 |
| *** zhill has quit IRC | 09:32 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Fix ClientException handling in Container Sync https://review.openstack.org/256306 | 09:37 |
| *** jordanP has joined #openstack-swift | 09:38 | |
| *** aix has joined #openstack-swift | 09:42 | |
| *** ig0r_ has quit IRC | 09:47 | |
| acoles | eranrom: re. your question on patch 205803, see http://www.ktcdigital.com/2012/07/whats-up-with-all-the-symbols/ :) | 09:54 |
| patchbot | acoles: https://review.openstack.org/#/c/205803/ - Container-Sync to iterate only over synced containers | 09:54 |
| *** jistr has joined #openstack-swift | 09:55 | |
| tdasilva | good morning | 09:59 |
| *** haomaiwang has quit IRC | 10:01 | |
| mahatic | morning already? | 10:01 |
| *** haomaiwang has joined #openstack-swift | 10:01 | |
| acoles | tdasilva: good morning. you're up early! | 10:01 |
| tdasilva | acoles: 8am here ;) | 10:02 |
| tdasilva | acoles: i'm in Brasil | 10:03 |
| mahatic | oh well, morning! | 10:03 |
| mahatic | :) | 10:03 |
| acoles | tdasilva: now i'm confused. Brasil is only UTC-2 ? | 10:03 |
| janonymous | tdasilva: o/ | 10:04 |
| acoles | i thought it was UTC-4 | 10:04 |
| tdasilva | yeah..BRST is UTC-2. (Brasilia Summer Time) | 10:04 |
| tdasilva | BRT is UTC-3 | 10:05 |
| acoles | tdasilva: oh yeah its summer there :) | 10:05 |
| tdasilva | acoles: yes!!! :-) | 10:05 |
| kota_ | mahatic: i think you need just recheck. Other patches also got similar issue, | 10:06 |
| tdasilva | wow, even kota_ is still around!! I like this time | 10:06 |
| mahatic | kota_: yeah, thanks. I just noticed some email about gate failure | 10:06 |
| kota_ | tdasilva: yeah | 10:07 |
| kota_ | tdasilva : yeah but unfortunately i'm heading for a pub today :P | 10:08 |
| tdasilva | kota_: nice! no worries, enjoy the pub, it's friday night! | 10:09 |
| kota_ | Usually i'm on this time | 10:09 |
| tdasilva | janonymous: o/ hello | 10:09 |
| *** esker has joined #openstack-swift | 10:27 | |
| *** esker has quit IRC | 10:32 | |
| *** hseipp has joined #openstack-swift | 10:33 | |
| *** aix has quit IRC | 10:36 | |
| *** ho has quit IRC | 10:42 | |
| openstackgerrit | Takashi Kajinami proposed openstack/swift: Fix proxy-server's support for chunked transferring in GET object https://review.openstack.org/256201 | 10:43 |
| * mahatic is ooo until wednesday | 10:44 | |
| *** Jeffrey4l has quit IRC | 10:45 | |
| *** esker has joined #openstack-swift | 10:48 | |
| tdasilva | mahatic: enjoy! | 10:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 10:51 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 10:51 |
| mahatic | tdasilva: thanks! you too in brasil | 10:51 |
| tdasilva | mahatic: thx :) | 10:52 |
| tdasilva | mahatic: btw, your changes to the obj. versioning doc look good | 10:53 |
| *** esker has quit IRC | 10:54 | |
| mahatic | tdasilva: ok, apparently the gate is fixed and i left a recheck | 10:56 |
| mahatic | and added you a reviewer too | 10:56 |
| tdasilva | mahatic: ok, cool | 10:56 |
| *** SkyRocknRoll has quit IRC | 10:58 | |
| openstackgerrit | Takashi Kajinami proposed openstack/swift: Fix proxy-server's support for chunked transferring in GET object https://review.openstack.org/256201 | 11:00 |
| *** haomaiwang has quit IRC | 11:01 | |
| *** haomaiwang has joined #openstack-swift | 11:01 | |
| *** ppai has quit IRC | 11:02 | |
| openstackgerrit | Merged openstack/swift: fix mock and assert in test_direct_client https://review.openstack.org/254341 | 11:05 |
| openstackgerrit | Takashi Kajinami proposed openstack/swift: Fix proxy-server's support for chunked transferring in GET object https://review.openstack.org/256201 | 11:06 |
| *** ig0r_ has joined #openstack-swift | 11:08 | |
| *** esker has joined #openstack-swift | 11:11 | |
| *** aix has joined #openstack-swift | 11:12 | |
| *** esker has quit IRC | 11:17 | |
| *** noark9 has quit IRC | 11:19 | |
| *** manous has joined #openstack-swift | 11:22 | |
| *** chlong has joined #openstack-swift | 11:32 | |
| *** jmccarthy has quit IRC | 11:34 | |
| *** jmccarthy has joined #openstack-swift | 11:34 | |
| *** sanchitmalhotra has quit IRC | 11:34 | |
| *** ig0r_ has quit IRC | 11:46 | |
| *** Guest6019 has quit IRC | 11:51 | |
| *** yatin has quit IRC | 11:57 | |
| *** kei_yama has quit IRC | 11:58 | |
| *** haomaiwang has quit IRC | 12:01 | |
| *** haomaiwang has joined #openstack-swift | 12:01 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:04 | |
| *** changbl has quit IRC | 12:04 | |
| openstackgerrit | Stuart McLaren proposed openstack/python-swiftclient: Retry download of object body https://review.openstack.org/159208 | 12:07 |
| *** ig0r_ has joined #openstack-swift | 12:14 | |
| *** rcernin has quit IRC | 12:18 | |
| *** links has quit IRC | 12:19 | |
| *** ig0r_ has quit IRC | 12:24 | |
| *** janonymous_ has joined #openstack-swift | 12:27 | |
| *** janonymous_ has quit IRC | 12:28 | |
| *** janonymous_ has joined #openstack-swift | 12:28 | |
| *** [1]eranrom has joined #openstack-swift | 12:29 | |
| *** eranrom has quit IRC | 12:29 | |
| *** [1]eranrom is now known as eranrom | 12:29 | |
| *** petertr7_away is now known as petertr7 | 12:30 | |
| janonymous_ | kota_ : hey, u around? | 12:30 |
| *** petertr7 is now known as petertr7_away | 12:39 | |
| *** cdelatte has joined #openstack-swift | 12:49 | |
| *** esker has joined #openstack-swift | 12:59 | |
| *** dslevin has joined #openstack-swift | 13:00 | |
| *** haomaiwang has quit IRC | 13:01 | |
| *** haomaiwang has joined #openstack-swift | 13:01 | |
| *** esker has quit IRC | 13:03 | |
| *** jistr is now known as jistr|biab | 13:10 | |
| *** ig0r_ has joined #openstack-swift | 13:16 | |
| *** esker has joined #openstack-swift | 13:19 | |
| *** ig0r_ has quit IRC | 13:25 | |
| *** esker has quit IRC | 13:26 | |
| *** venkatesh has quit IRC | 13:36 | |
| *** haomaiwang has quit IRC | 13:41 | |
| *** haomaiw__ has joined #openstack-swift | 13:41 | |
| *** esker has joined #openstack-swift | 13:44 | |
| *** janonymous_ has quit IRC | 13:44 | |
| *** esker has quit IRC | 13:48 | |
| *** haomaiw__ has quit IRC | 13:52 | |
| *** esker has joined #openstack-swift | 14:04 | |
| *** jistr|biab is now known as jistr | 14:09 | |
| *** esker has quit IRC | 14:09 | |
| *** haomaiwang has joined #openstack-swift | 14:10 | |
| *** CaioBrentano has joined #openstack-swift | 14:10 | |
| *** haomaiwang has quit IRC | 14:11 | |
| *** haomaiwa_ has joined #openstack-swift | 14:11 | |
| *** haomaiwa_ has quit IRC | 14:12 | |
| *** acoles is now known as acoles_ | 14:12 | |
| *** 18WABFL3O has joined #openstack-swift | 14:12 | |
| *** 18WABFL3O has quit IRC | 14:13 | |
| *** haomaiwang has joined #openstack-swift | 14:13 | |
| *** haomaiwang has quit IRC | 14:14 | |
| *** haomaiwa_ has joined #openstack-swift | 14:14 | |
| *** haomaiwa_ has quit IRC | 14:15 | |
| *** haomaiwang has joined #openstack-swift | 14:15 | |
| *** haomaiwang has quit IRC | 14:16 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 14:16 |
| *** haomaiwang has joined #openstack-swift | 14:16 | |
| *** changbl has joined #openstack-swift | 14:16 | |
| *** haomaiwang has quit IRC | 14:17 | |
| *** haomaiwang has joined #openstack-swift | 14:17 | |
| *** haomaiwang has quit IRC | 14:18 | |
| *** haomaiwang has joined #openstack-swift | 14:18 | |
| *** janonymous_ has joined #openstack-swift | 14:18 | |
| *** haomaiwang has quit IRC | 14:19 | |
| *** haomaiwang has joined #openstack-swift | 14:19 | |
| *** breitz has quit IRC | 14:19 | |
| *** breitz has joined #openstack-swift | 14:19 | |
| *** haomaiwang has quit IRC | 14:20 | |
| *** haomaiwang has joined #openstack-swift | 14:20 | |
| wbhuber | good morning, my fellow swifters! | 14:20 |
| *** haomaiwang has quit IRC | 14:21 | |
| janonymous_ | hey wbhuber | 14:21 |
| wbhuber | just wanted to let you know jrichli did really awesome last night at the openstack meetup presenting on swift encryption: http://www.meetup.com/OpenStack-Austin/events/226421866/ | 14:21 |
| *** haomaiwa_ has joined #openstack-swift | 14:21 | |
| wbhuber | a lot of good questions being asked and she articulated at each one.... | 14:21 |
| *** haomaiwa_ has quit IRC | 14:22 | |
| *** 17WAA8LUI has joined #openstack-swift | 14:22 | |
| tdasilva | wbhuber, jrichli: awesome, great job! | 14:22 |
| wbhuber | do any of you guys know the purpose of swift_bytes in this string: 'content_type': 'application/json;swift_bytes=25' | 14:22 |
| *** 17WAA8LUI has quit IRC | 14:23 | |
| tdasilva | got slides to share? | 14:23 |
| wbhuber | it's failling some of SLO unit test that i'm rewriting | 14:23 |
| wbhuber | will ask jrichli to share when she gets on line | 14:23 |
| *** haomaiwa_ has joined #openstack-swift | 14:23 | |
| *** haomaiwa_ has quit IRC | 14:24 | |
| *** 17WAA8LWI has joined #openstack-swift | 14:24 | |
| *** 17WAA8LWI has quit IRC | 14:25 | |
| *** 17WAA8LXI has joined #openstack-swift | 14:25 | |
| *** 17WAA8LXI has quit IRC | 14:26 | |
| *** haomaiwang has joined #openstack-swift | 14:26 | |
| *** haomaiwang has quit IRC | 14:27 | |
| *** haomaiwang has joined #openstack-swift | 14:27 | |
| *** haomaiwang has quit IRC | 14:28 | |
| *** haomaiwa_ has joined #openstack-swift | 14:28 | |
| *** haomaiwa_ has quit IRC | 14:29 | |
| *** haomaiwa_ has joined #openstack-swift | 14:29 | |
| *** haomaiwa_ has quit IRC | 14:30 | |
| *** haomaiwa_ has joined #openstack-swift | 14:30 | |
| *** haomaiwa_ has quit IRC | 14:31 | |
| *** haomaiwa_ has joined #openstack-swift | 14:31 | |
| *** haomaiwa_ has quit IRC | 14:32 | |
| *** haomaiwa_ has joined #openstack-swift | 14:32 | |
| *** haomaiwa_ has quit IRC | 14:33 | |
| *** haomaiwa_ has joined #openstack-swift | 14:33 | |
| *** haomaiwa_ has quit IRC | 14:34 | |
| *** haomaiwang has joined #openstack-swift | 14:34 | |
| *** haomaiwang has quit IRC | 14:35 | |
| *** haomaiwang has joined #openstack-swift | 14:35 | |
| *** haomaiwang has quit IRC | 14:36 | |
| *** haomaiwang has joined #openstack-swift | 14:36 | |
| *** haomaiwang has quit IRC | 14:37 | |
| *** 17WAA8MB6 has joined #openstack-swift | 14:37 | |
| *** 17WAA8MB6 has quit IRC | 14:38 | |
| *** haomaiwa_ has joined #openstack-swift | 14:38 | |
| *** haomaiwa_ has quit IRC | 14:39 | |
| *** 18WABFM0S has joined #openstack-swift | 14:39 | |
| *** 18WABFM0S has quit IRC | 14:40 | |
| *** haomaiwang has joined #openstack-swift | 14:40 | |
| *** haomaiwang has quit IRC | 14:41 | |
| *** ig0r_ has joined #openstack-swift | 14:41 | |
| *** haomaiwang has joined #openstack-swift | 14:41 | |
| *** haomaiwang has quit IRC | 14:42 | |
| *** haomaiwa_ has joined #openstack-swift | 14:42 | |
| *** haomaiwa_ has quit IRC | 14:43 | |
| *** haomaiwang has joined #openstack-swift | 14:43 | |
| *** ig0r_ has quit IRC | 14:43 | |
| *** haomaiwang has quit IRC | 14:44 | |
| *** haomaiwa_ has joined #openstack-swift | 14:44 | |
| *** haomaiwa_ has quit IRC | 14:45 | |
| *** 7F1AA7M98 has joined #openstack-swift | 14:45 | |
| *** 7F1AA7M98 has quit IRC | 14:46 | |
| *** 17WAA8MLS has joined #openstack-swift | 14:46 | |
| *** 17WAA8MLS has quit IRC | 14:47 | |
| *** haomaiwang has joined #openstack-swift | 14:47 | |
| *** diogogmt has quit IRC | 14:47 | |
| *** haomaiwang has quit IRC | 14:48 | |
| *** Jeffrey4l has joined #openstack-swift | 14:48 | |
| *** haomaiwang has joined #openstack-swift | 14:48 | |
| *** haomaiwang has quit IRC | 14:49 | |
| *** 17WAA8MN8 has joined #openstack-swift | 14:49 | |
| *** 17WAA8MN8 has quit IRC | 14:50 | |
| *** haomaiwa_ has joined #openstack-swift | 14:50 | |
| *** diogogmt has joined #openstack-swift | 14:51 | |
| *** haomaiwa_ has quit IRC | 14:51 | |
| *** blmartin has joined #openstack-swift | 14:51 | |
| *** haomaiwa_ has joined #openstack-swift | 14:51 | |
| *** haomaiwa_ has quit IRC | 14:52 | |
| *** haomaiwang has joined #openstack-swift | 14:52 | |
| *** haomaiwang has quit IRC | 14:53 | |
| *** haomaiwang has joined #openstack-swift | 14:53 | |
| *** haomaiwang has quit IRC | 14:54 | |
| *** haomaiwa_ has joined #openstack-swift | 14:54 | |
| *** haomaiwa_ has quit IRC | 14:55 | |
| *** haomaiwa_ has joined #openstack-swift | 14:55 | |
| *** haomaiwa_ has quit IRC | 14:56 | |
| *** haomaiwa_ has joined #openstack-swift | 14:56 | |
| *** haomaiwa_ has quit IRC | 14:57 | |
| *** haomaiwa_ has joined #openstack-swift | 14:57 | |
| *** ig0r_ has joined #openstack-swift | 15:01 | |
| *** haomaiwa_ has quit IRC | 15:04 | |
| *** haomaiwang has joined #openstack-swift | 15:04 | |
| *** ig0r_ has quit IRC | 15:06 | |
| *** sgundur has joined #openstack-swift | 15:06 | |
| *** esker has joined #openstack-swift | 15:07 | |
| *** diogogmt has quit IRC | 15:09 | |
| *** links has joined #openstack-swift | 15:11 | |
| *** diogogmt has joined #openstack-swift | 15:12 | |
| *** diogogmt has quit IRC | 15:13 | |
| *** tongli has joined #openstack-swift | 15:25 | |
| *** SkyRocknRoll has quit IRC | 15:30 | |
| *** tsg has joined #openstack-swift | 15:33 | |
| *** daddyjoseph97 has joined #openstack-swift | 15:33 | |
| *** diazjf has joined #openstack-swift | 15:36 | |
| *** pgbridge has joined #openstack-swift | 15:40 | |
| jrichli | good morning. wbhuber, tdasilva, thanks! I will provide a link to the slides in a moment - they are really high level, though. And not much without the spoken words. | 15:44 |
| jrichli | wbhuber: my understanding is that swift_bytes is used internally only for the container listing so that the total size of the large object is shown. | 15:44 |
| jrichli | also, I am *technically* on vacation today. but will be popping in occasionally | 15:45 |
| *** lpabon has joined #openstack-swift | 15:46 | |
| *** diazjf has quit IRC | 15:48 | |
| jrichli | my presentation is here https://drive.google.com/file/d/0B-zpM_BFwU5ATWMxWXZWUzVJMkU/view?usp=sharing | 15:49 |
| jrichli | The highlight was that at the end, somebody talked to me about how they want to start contributing to swift and asked for some advice :-) | 15:50 |
| *** david-ly_ has joined #openstack-swift | 15:50 | |
| tdasilva | jrichli: great slides | 15:53 |
| *** david-lyle has quit IRC | 15:53 | |
| *** david-ly_ is now known as david-lyle | 15:53 | |
| jrichli | tdasilva: thanks | 15:55 |
| *** janonymous has quit IRC | 15:56 | |
| *** sgundur has quit IRC | 15:57 | |
| janonymous_ | nice slides | 15:57 |
| jrichli | janonymous_: thanks | 15:59 |
| *** diazjf has joined #openstack-swift | 16:01 | |
| *** changbl has quit IRC | 16:03 | |
| wbhuber | jrichi: thumbs up! :) | 16:09 |
| *** links has quit IRC | 16:09 | |
| jrichli | wbhuber: he he, like the monkey | 16:11 |
| wbhuber | yes, fonze the monkey :) | 16:11 |
| janonymous_ | Haha, yes i was thinking of what the monkey in last slide means :P | 16:12 |
| jrichli | the last slide wasn't shown - Resources slide was really last | 16:12 |
| *** diogogmt has joined #openstack-swift | 16:13 | |
| jrichli | the first slide is where I introduced "The Fonze" and a little story | 16:13 |
| janonymous_ | Hahaa, cool :) | 16:14 |
| *** changbl has joined #openstack-swift | 16:16 | |
| *** ig0r_ has joined #openstack-swift | 16:21 | |
| *** jlvillal has quit IRC | 16:22 | |
| *** jlvillal has joined #openstack-swift | 16:22 | |
| *** jlvillal has quit IRC | 16:24 | |
| *** lpabon has quit IRC | 16:24 | |
| *** hseipp has quit IRC | 16:24 | |
| *** jlvillal has joined #openstack-swift | 16:28 | |
| *** jlvillal has quit IRC | 16:30 | |
| *** jlvillal has joined #openstack-swift | 16:31 | |
| *** haomaiwang has quit IRC | 16:34 | |
| *** sgundur has joined #openstack-swift | 16:37 | |
| *** acoles_ is now known as acoles | 16:42 | |
| *** mfalatic has joined #openstack-swift | 16:42 | |
| *** mfalatic has quit IRC | 16:45 | |
| *** jordanP has quit IRC | 16:46 | |
| *** jordanP has joined #openstack-swift | 16:47 | |
| *** ig0r_ has quit IRC | 16:47 | |
| acoles | jrichli: nice! | 16:48 |
| jrichli | acoles: thanks | 16:49 |
| *** diazjf has quit IRC | 16:52 | |
| openstackgerrit | Bill Huber proposed openstack/swift: SLO multipart-manfest=get call returns json in inconsistent format https://review.openstack.org/256580 | 16:52 |
| *** dslevin has quit IRC | 16:52 | |
| wbhuber | tdasilva: timburke: ^^ feel free to make recommendation/feedback. | 16:53 |
| *** dslevin has joined #openstack-swift | 16:56 | |
| *** garthb has joined #openstack-swift | 16:57 | |
| *** esker has quit IRC | 17:01 | |
| *** rledisez has quit IRC | 17:05 | |
| *** klrmn has joined #openstack-swift | 17:07 | |
| *** diazjf has joined #openstack-swift | 17:07 | |
| *** diazjf has quit IRC | 17:08 | |
| *** thumpba has joined #openstack-swift | 17:08 | |
| *** manous has quit IRC | 17:10 | |
| *** lyrrad has joined #openstack-swift | 17:11 | |
| *** gyee has joined #openstack-swift | 17:16 | |
| *** klrmn has quit IRC | 17:17 | |
| *** ianbrown has quit IRC | 17:19 | |
| *** ianbrown has joined #openstack-swift | 17:20 | |
| hrou | jrichli, that's really awesome ! What was the audience like ? New'ish to swift / wanting to get involved (dev or ops side). | 17:21 |
| jrichli | hrou: thanks. It is hard for me to say. I only talked to a few of them. I was advised that the general meetup audience would most likely be new to swift. A couple security folks were there. | 17:23 |
| jrichli | Because there was a "dojo" this week at IBM, a few people who work with Monty Taylor under Angel Diaz were there. That was unexpected for me :-) | 17:23 |
| openstackgerrit | Ondřej Nový proposed openstack/swift: Removed deprecated tox -downloadcache option https://review.openstack.org/256597 | 17:24 |
| hrou | jrichli, Ah Angel and Monty, nice :) That's really great, is it like an openstack meetup with different people presenting about the various projects? | 17:25 |
| jrichli | hrou: well, Angel and Monty were not there :-) just people who work in their team | 17:25 |
| hrou | jrichli, There loss, they should have attended ; ) | 17:25 |
| *** dmorita has joined #openstack-swift | 17:26 | |
| jrichli | hrou: I don't have a lot of experience with meetups, but I take it that the openstack meetups are pretty wide-ranged in topic. As long as its under the openstack umbrella. | 17:27 |
| *** dmorita has quit IRC | 17:27 | |
| hrou | jrichli, yea I just looked it up, interesting, they have them throughout ! | 17:27 |
| *** dmorita has joined #openstack-swift | 17:28 | |
| *** trifon has quit IRC | 17:30 | |
| *** esker has joined #openstack-swift | 17:36 | |
| *** esker has joined #openstack-swift | 17:36 | |
| *** daddyjoseph97 has quit IRC | 17:38 | |
| *** arnox has quit IRC | 17:44 | |
| *** daddyjoseph97 has joined #openstack-swift | 17:44 | |
| *** daddyjoseph97 has quit IRC | 17:45 | |
| *** esker has quit IRC | 17:47 | |
| *** sgundur has quit IRC | 17:47 | |
| *** sgundur1 has joined #openstack-swift | 17:47 | |
| *** esker has joined #openstack-swift | 17:48 | |
| *** robefran has joined #openstack-swift | 17:50 | |
| *** janonymous_ has quit IRC | 17:58 | |
| *** diazjf has joined #openstack-swift | 18:01 | |
| *** diazjf has quit IRC | 18:02 | |
| *** diazjf has joined #openstack-swift | 18:03 | |
| *** jordanP has quit IRC | 18:05 | |
| *** janonymous_ has joined #openstack-swift | 18:08 | |
| openstackgerrit | Ondřej Nový proposed openstack/swift: Deprecated tox -downloadcache option removed https://review.openstack.org/256597 | 18:09 |
| notmyname | good morning | 18:14 |
| notmyname | jrichli: that's great! | 18:14 |
| *** klrmn has joined #openstack-swift | 18:16 | |
| jrichli | notmyname: :-) | 18:16 |
| notmyname | the slides look great | 18:16 |
| *** zaitcev has joined #openstack-swift | 18:16 | |
| *** ChanServ sets mode: +v zaitcev | 18:16 | |
| diazjf | whoop whoop jrichli :) | 18:18 |
| jrichli | notmyname diazjf: thanks. i got into a little more than what was on the slides. | 18:19 |
| *** delattec has joined #openstack-swift | 18:20 | |
| *** cdelatte has quit IRC | 18:23 | |
| *** zhill has joined #openstack-swift | 18:28 | |
| *** nadeem has joined #openstack-swift | 18:29 | |
| * jrichli is stepping out | 18:31 | |
| *** janonymous__ has joined #openstack-swift | 18:34 | |
| *** haomaiwang has joined #openstack-swift | 18:37 | |
| *** acoles is now known as acoles_ | 18:38 | |
| *** janonymous__ has quit IRC | 18:38 | |
| *** janonymous__ has joined #openstack-swift | 18:42 | |
| *** shakamunyi has quit IRC | 18:44 | |
| *** tamizh_geek has quit IRC | 18:45 | |
| *** jistr has quit IRC | 18:45 | |
| *** tamizh_geek has joined #openstack-swift | 18:47 | |
| *** janonymous__ has quit IRC | 18:48 | |
| *** tamizh_geek has quit IRC | 18:52 | |
| *** silor has quit IRC | 18:55 | |
| *** tamizh_geek has joined #openstack-swift | 18:56 | |
| *** robefran has quit IRC | 18:56 | |
| *** robefran has joined #openstack-swift | 18:57 | |
| *** blmartin has quit IRC | 19:00 | |
| *** sgundur1 has quit IRC | 19:02 | |
| *** robefran has quit IRC | 19:02 | |
| *** pberis has joined #openstack-swift | 19:06 | |
| *** sgundur has joined #openstack-swift | 19:07 | |
| *** nadeem has quit IRC | 19:12 | |
| *** hogepodge has quit IRC | 19:18 | |
| *** hogepodge has joined #openstack-swift | 19:18 | |
| *** hogepodge has quit IRC | 19:19 | |
| *** aix has quit IRC | 19:19 | |
| *** hogepodge has joined #openstack-swift | 19:19 | |
| *** diazjf has quit IRC | 19:19 | |
| *** changbl has quit IRC | 19:19 | |
| *** diazjf has joined #openstack-swift | 19:20 | |
| *** diazjf has quit IRC | 19:22 | |
| *** hogepodge has quit IRC | 19:23 | |
| *** hogepodge has joined #openstack-swift | 19:23 | |
| *** garthb_ has joined #openstack-swift | 19:28 | |
| *** garthb has quit IRC | 19:30 | |
| *** changbl has joined #openstack-swift | 19:39 | |
| *** esker has quit IRC | 19:48 | |
| *** shakamunyi has joined #openstack-swift | 19:48 | |
| wbhuber | dfg: timburke: thanx for ur review. im gonna look into a new query argument like get-formatted instead of breaking backward compatibility | 19:49 |
| *** delatte has joined #openstack-swift | 19:52 | |
| *** shakamunyi has quit IRC | 19:55 | |
| *** delattec has quit IRC | 19:55 | |
| *** robefran has joined #openstack-swift | 19:58 | |
| *** diazjf has joined #openstack-swift | 20:02 | |
| *** tsg has quit IRC | 20:02 | |
| robefran | A swift consistency question. Can a sequence of two PUTs followed by a POST (post-as-copy) eventually settle on the data from the first PUT ? I've created an etherpad https://etherpad.openstack.org/p/swift_consistency_questions with the details of the sequnce I'm worried about. Is this a known situation, or a violation of eventual consistency ? | 20:07 |
| *** Zyric has joined #openstack-swift | 20:10 | |
| Zyric | Good morning | 20:11 |
| *** dmorita has quit IRC | 20:33 | |
| *** robefran has quit IRC | 20:35 | |
| *** ig0r_ has joined #openstack-swift | 20:37 | |
| *** hogepodge has quit IRC | 20:47 | |
| *** sgundur has quit IRC | 20:50 | |
| *** ChubYann has joined #openstack-swift | 20:51 | |
| *** sgundur has joined #openstack-swift | 20:53 | |
| *** sgundur1 has joined #openstack-swift | 20:54 | |
| *** rcernin has joined #openstack-swift | 20:56 | |
| *** sgundur1 has quit IRC | 20:56 | |
| *** lcurtis has joined #openstack-swift | 21:08 | |
| *** dewsday has joined #openstack-swift | 21:09 | |
| *** diazjf has quit IRC | 21:10 | |
| *** CaioBrentano has quit IRC | 21:10 | |
| *** hogepodge has joined #openstack-swift | 21:12 | |
| *** dmorita has joined #openstack-swift | 21:12 | |
| *** dewsday has quit IRC | 21:12 | |
| *** tsg has joined #openstack-swift | 21:14 | |
| *** diazjf has joined #openstack-swift | 21:26 | |
| notmyname | I wishe robefran were still here | 21:31 |
| hrou | notmyname, feel free to answer ;) Here I'll relay the message we've been chatting about the subject the last week, or you could answer in the etherpad | 21:34 |
| notmyname | heh | 21:34 |
| notmyname | so the answer is "sortof" | 21:34 |
| notmyname | or rather, yes it might be possible, but it's very very unlikely | 21:34 |
| notmyname | so when swift does a server-side copy, it uses the x-newest:true header to get the source | 21:34 |
| notmyname | which means that it will query all the primaries and use the one with the biggest timestamp | 21:35 |
| hrou | notmyname, Yea we agree about the "that'll probably never" happen part - but to give some context this us trying to understand how things work so we can present a better way forward for symlink ;) | 21:35 |
| *** then3rd has joined #openstack-swift | 21:35 | |
| hrou | notmyname, yep! But x-newest may not be able to find a handoff that's offline atm. | 21:35 |
| notmyname | so his example would only happen if all of the primary nodes with the newer data were unavailable | 21:35 |
| notmyname | totally | 21:35 |
| notmyname | x-newest isn't a guarantee, really. but it significantly reduces the possiblity of getting older data | 21:36 |
| hrou | yep, or if the primarys were down earlier and didn't have all the data in the first place? | 21:36 |
| hrou | notmyname, perfect yea that was our interpretation | 21:37 |
| notmyname | so yeah, we could come up with scenarios where that might happen. failures + ring rebalances + operators who are pushing new rings too fast + etc | 21:38 |
| hrou | notmyname, that's great; That's what we expected and just wanted to ensure there wasn't something else we're missing that would handle it. | 21:39 |
| hrou | hrou, that's good though it'll help up frame what we're proposing with symlink given that's the case here. | 21:39 |
| hrou | notmyname, thanks! | 21:39 |
| *** zhill has quit IRC | 21:42 | |
| *** robefran has joined #openstack-swift | 21:47 | |
| then3rd | Hey all, I have a dumb question I can't seem to find the answer for. I've only a few months experience with swift and am responsible for a previously configured cluster. So far, I haven't made any big changes to the existing environment. | 21:47 |
| notmyname | then3rd: sure. what's up? | 21:48 |
| then3rd | Our current haproxy, swift-proxy, keystone, and storage nodes have both public and private IPs. The storage nodes communicate via a 10G private network that the proxies also have access to. I think it's all fairly standard. I'm going to soon be adding additional storage nodes located in a separate geographic region, so I'll be learning how to implement regions. I've read that generally you want to | 21:48 |
| then3rd | configure either a VPN or MPLS/Ethernet circuit to connect the private networks. Here's my question: Since we already have public IPs, would changing the rsycnd config to bind to the public IPs enable me to synchronize between regions without the need for a vpn? | 21:48 |
| notmyname | then3rd: the two regions need to be routable for each other. eg storage nodes in R1 need to be able to open a connection to nodes in R2 | 21:49 |
| then3rd | So, internet-routable ips would count? | 21:50 |
| notmyname | then3rd: but, and this is important, anything that can talk to the storage nodes directly, including via rsync, will be able to access any and all data on the node | 21:50 |
| notmyname | so you *really* don't want the storage nodes to be accessible by anything other than other trusted nodes in the cluster (eg proxies and maybe admin bastion servers) | 21:50 |
| then3rd | So, a security vector. As long as we trust our keys, is that okay? I'm guessing that's why vpns are prefered. | 21:51 |
| notmyname | trust your keys? what keys? | 21:51 |
| robefran | notmymame: I'm back, but hrou explained the motivation for the question (symlink POST) | 21:51 |
| *** dustins is now known as dustins|out | 21:51 | |
| robefran | before we get too complicated with symlink POST we just want to check that we are not trying to hard to close holes that exist anyways in regular POST | 21:52 |
| then3rd | Hmm.. maybe I'm confused. I thought rsync used ssh keys. | 21:52 |
| notmyname | then3rd: yeah, it can. but rsync isn't the only way nodes are talking to each other | 21:53 |
| then3rd | What other communication happens between them? | 21:53 |
| notmyname | then3rd: there's an internal http api for node to node (or proxy to node) communication | 21:54 |
| then3rd | Okay. Can it be configured for tls like everything else? | 21:55 |
| notmyname | then3rd: in addition to get/put/post/delete, there's also a replicate command | 21:55 |
| then3rd | I'm just trying to feel out my options in the event the boss doesn't want to pay for additional ASAs | 21:55 |
| notmyname | then3rd: not directly. or rather, like the proxy, you really don't want to set it up to use ssl. you want to use something like stud or stunnel to do the TLS | 21:55 |
| then3rd | I think what I'm hearing is: make the private network work. have you heard of anyone trying to do it without? | 21:56 |
| notmyname | without a VPN between the regions? | 21:57 |
| *** sgundur1 has joined #openstack-swift | 21:57 | |
| then3rd | Yeah. | 21:57 |
| notmyname | if you don't have your own connection between the regions, use a VPN | 21:58 |
| then3rd | Cool. I value your time and input, thanks for answering! | 22:00 |
| then3rd | The next step will be learing how to use reigons. The concept hasn't completely clicked with me yet. I need to get hands-on. | 22:01 |
| notmyname | sure | 22:04 |
| notmyname | then3rd: our of curiosity, what are you using swift for? | 22:04 |
| notmyname | I like to hear how people are using swift and what they've done with it | 22:07 |
| notmyname | then3rd: here's an older blog post on global clusters with swift https://swiftstack.com/blog/2012/09/16/globally-distributed-openstack-swift-cluster/ | 22:08 |
| notmyname | then3rd: bah. looks like some images were lost in a blog reformat. here's an older cached one https://web.archive.org/web/20140429163105/https://swiftstack.com/blog/2012/09/16/globally-distributed-openstack-swift-cluster/ | 22:09 |
| then3rd | We're running a CDN for a startup company that's building a media managment platform. Atlys.io and atlys MMP | 22:10 |
| notmyname | cool | 22:11 |
| then3rd | They have their 1.0 client app is available to play with http://atlys.tv and you can mess with the mmp dash at https://atlysmmp.com. | 22:12 |
| notmyname | then3rd: so it's video streaming from swift? | 22:12 |
| notmyname | oh yeah! that video looks like a swift url ;-) | 22:13 |
| notmyname | cool | 22:14 |
| then3rd | Their 2.0 version is launching first quarter next year.. wayyy nicer interface and tons more features. Yep, all straight from swift. They're working on implimenting transcoding and a bunch of other stuff. They're interested in storlets and that terrifys me. | 22:14 |
| notmyname | lol | 22:14 |
| then3rd | s/terrifys/terrifies/ | 22:14 |
| then3rd | I'm like.. I don't do java, I won't write it. But I may be able to enable it. | 22:15 |
| then3rd | I think storlets might be overkill for what they want anyway. who knows. | 22:15 |
| notmyname | depends on what they need | 22:16 |
| notmyname | the main advantage of storlets over "normal" middleware in swift is that the storlet can change without having to restart any processes in the cluster | 22:17 |
| then3rd | I was on this channel a few months back asking how I could track bandwidth usage per-object. Thanks to someones advice here, I went with post-log analysis. logstash+elasticsearch let me pull all kinds of fun stats that we can use for billing. | 22:17 |
| notmyname | so you, in effect, get user-defined middleware | 22:17 |
| notmyname | nice | 22:18 |
| notmyname | yeah, that's the way to do it | 22:18 |
| then3rd | I still don't know what middleware is. | 22:18 |
| then3rd | I'm a noob man.. such noob. | 22:18 |
| notmyname | http://docs.openstack.org/developer/swift/development_middleware.html | 22:19 |
| *** sgundur1 has quit IRC | 22:20 | |
| *** haomaiwang has quit IRC | 22:23 | |
| *** haomaiwang has joined #openstack-swift | 22:23 | |
| then3rd | Okay. I'm still getting my feet wet with python (I don't know why it took me this long to realize how wonderufl it is) but if I understand right.. this lets us do things to the objects passing through the proxy. So I could write middleware to scan for viruses on the fly, or rewrite configurations as they're accessed. | 22:35 |
| notmyname | yeah, exactly | 22:36 |
| notmyname | you can intercept and act on (including modify) the request on the way in and the response on the way out | 22:36 |
| then3rd | Could middleware do transcoding as a video is uploaded? | 22:36 |
| notmyname | absolutely | 22:36 |
| then3rd | Crap, they're not gonna like to hear that. | 22:37 |
| notmyname | the basic hello world version of middleware is resizing an image | 22:37 |
| *** openstackgerrit has quit IRC | 22:37 | |
| notmyname | now, the reason you might not want to transcode on the fly on ingest is because transcoding might take a long time | 22:37 |
| *** rcernin has quit IRC | 22:37 | |
| *** openstackgerrit has joined #openstack-swift | 22:37 | |
| then3rd | i think they're pulling the files out, transcoding, and re-adding as a new object. Though, they want to keep archives of the original quality. | 22:38 |
| notmyname | and you might want to store the non-transcoded (original source format) in the cluster so you can re-transcode again later | 22:38 |
| notmyname | yeah | 22:38 |
| then3rd | ^^] | 22:38 |
| notmyname | either way is possible. depends on what the end goal is | 22:38 |
| then3rd | So, the middleware could be cablable of writing both objects. That would be awesome. | 22:38 |
| *** diazjf has quit IRC | 22:38 | |
| notmyname | yeah, that's possible too | 22:39 |
| notmyname | I've used this use case as an example for storage polices before. one storage policy eg erasure coded or 3 or 4 replicas across multiple regions for the original format. then use something like a regional 2 replica policy for the transcoded version | 22:40 |
| notmyname | so you get full durability of the original that you don't want to replace. but cheaper (and potentially local) storage for the transcoded ones that people access | 22:40 |
| openstackgerrit | Merged openstack/swift: Update versioned_writes doc https://review.openstack.org/256250 | 22:41 |
| then3rd | notmyname: that's exactly what they want too. I knew I would have to figure out something like that. | 22:41 |
| then3rd | Policies are also something I'm going to have to get familiar with | 22:42 |
| notmyname | then3rd: do you have a copy of the oreilly swift book yet? | 22:42 |
| *** thumpba has quit IRC | 22:43 | |
| *** diogogmt has quit IRC | 22:44 | |
| pdardeau | notmyname: yatin: your earlier discussion of swift ring was awesome | 22:45 |
| pdardeau | keeping handy reference to that for future use | 22:45 |
| notmyname | pdardeau: oh, good. it's a pretty complex topic ;-) | 22:45 |
| hrou | then3rd, I'd be curious about hearing more about their uses cases, there's talk (well there always talk ;-) of extending storlets beyond java if that's actually a blocking factor; You're correct though storlets bring in an ecosystem (all the suden you're running docker on your proxy / object server) and all the fun that comes with that | 22:45 |
| notmyname | then3rd: http://www.amazon.com/OpenStack-Swift-Administering-Developing-Storage/dp/1491900822/ | 22:45 |
| *** setmason has joined #openstack-swift | 22:46 | |
| notmyname | then3rd: or, if you want it for free (ie giving some contact details) https://swiftstack.com/books/ | 22:46 |
| then3rd | I don't have the orriley book, though I've read a few chapters of the Sotware Defined Storage with swift by Joe Arnold. | 22:46 |
| setmason | In SSBench, is there docs on the output columns? I’m trying to understand what a backend_error is. | 22:47 |
| *** Trixboxer has quit IRC | 22:49 | |
| *** tsg has quit IRC | 22:49 | |
| notmyname | setmason: yeah, I remember looking for that too | 22:49 |
| then3rd | hrou: One of the devs show me this and said it was what he wanted to do. I think the goal is basically transcoding on the storage nodes? http://superuser.openstack.org/articles/docker-meets-swift-a-broadcaster-s-experience | 22:50 |
| *** shakamunyi has joined #openstack-swift | 22:51 | |
| hrou | then3rd, Yep! That's essentially that use case, some transcoding on the way in or out of object storage | 22:52 |
| then3rd | So, having just learned about middleware, why use storlets to accomplish that? | 22:52 |
| notmyname | setmason: wait. backend_error? I can't find that string anywhere in the ssbench repo | 22:53 |
| notmyname | then3rd: because you can update a storlet implementation without getitng the cluster operator to redeploy middleware and update configs (if necessary) and restart the servers | 22:54 |
| hrou | then3rd, on paper you could indeed accomplish just about anything with middleware ;) Heck storlets are middleware. The project gives a mechanism for people to easily add the functionality though without having to get the operator to make any changes, or really be familiar with swift at all. | 22:54 |
| hrou | or what notmyname said ;) | 22:54 |
| notmyname | hrou: on the other hand, the operator overhead of managing the storlet stuff isn't zero either ;-) | 22:55 |
| then3rd | Oh! I get it.. storlets are objects.. so they could just build and upload. | 22:55 |
| notmyname | right | 22:55 |
| hrou | notmyname, 100% agreed I think that's going to be one of the problems to overcome with storlets, it can get complex, there's thread managment to worry about, how to allocate resources between accounts, not free at all. | 22:55 |
| hrou | notmyname, then3rd one other very big point / advantage; | 22:55 |
| hrou | Even if the operator was expecting of generic python middleware code (scary thought), there are a lot security concerns, storlets only allow you to mess with the input / output stream, you can't do anything else (no communication to any other part swift / outside world). | 22:56 |
| notmyname | right | 22:57 |
| notmyname | whereas with middleware, you better trust the developer of that code. ie know what you're running | 22:57 |
| hrou | notmyname, yep exactly | 22:58 |
| *** thumpba has joined #openstack-swift | 22:58 | |
| *** nadeem has joined #openstack-swift | 22:58 | |
| then3rd | You've both been enlightening. | 22:59 |
| hrou | then3rd, shameless plug time, I'd be curious to here about more use cases, there's a storlet IRC channel btw ! #openstack-storlets, which is starting to get a bit more active as of late thanks to kota_ and other folks. | 23:00 |
| *** shakamunyi has quit IRC | 23:00 | |
| then3rd | I'm sure I'll be back soon | 23:00 |
| then3rd | Can't say I have a lot of info about use cases though. | 23:01 |
| hrou | then3rd, oh no worries, thanks ! | 23:01 |
| then3rd | Well, I've been up for at least 24 hours at this point. I think it's time to go home and start my (on-call) weekend. | 23:02 |
| hrou | then3rd, "starting the weekend" always a good decision ;) | 23:02 |
| hrou | see ya | 23:02 |
| *** bkmz_ has joined #openstack-swift | 23:02 | |
| then3rd | Btw, any of you guys have friends at lehi Mirantis? I have some old bluehost buddies that all went and got jobs there. | 23:03 |
| *** August1914 has joined #openstack-swift | 23:03 | |
| *** bkmz has quit IRC | 23:06 | |
| *** diogogmt has joined #openstack-swift | 23:07 | |
| notmyname | then3rd: I notice that you're running an older version of swift. (in addition to all the other good stuff added recently) you'll need to upgrade to take advantage of storage policies | 23:19 |
| *** shakamunyi has joined #openstack-swift | 23:26 | |
| Zyric | notmyname: Speaking of new features/storage policies, is it possible to get eyes on https://review.openstack.org/#/c/241978/ ? I'm not sure what to do with it at this point, I know the goal was to have it done this week. From what I can tell the cluster failures aren't due to the patch code. | 23:27 |
| *** thumpba has quit IRC | 23:28 | |
| notmyname | Zyric: I think you're good for now. it's up to torgomatic and tdasilva to look at it and come to an agreement. and I should look at it again to see if my previous +2 still should be there | 23:28 |
| notmyname | Zyric: how's the rest of your first week? what do you need? what are you looking at? | 23:29 |
| Zyric | notmyname: Cool. It's been quite slow going for me the past few days, mostly just reading code and docs and trying to wrap my head around A) How Swift works and B) How my project fits into it, much like we discussed in our start-of-project meeting. | 23:29 |
| Zyric | I have a much better understanding of the system than I did a few days ago, but I'm still feeling pretty overwhelmed. | 23:29 |
| Zyric | I noticed myself falling into a kind of trap where I'll be reading about something relevant to my project, discover terms or concepts I'm unfamiliar with so I'll Google those so I can understand the relevant content better, then find more things I don't know about those things and research those and on and on it goes | 23:30 |
| notmyname | fun! | 23:30 |
| notmyname | Zyric: don't worry. that all sounds right :-) | 23:30 |
| notmyname | and don't try to understand everything 100% before trying to play with stuff. | 23:30 |
| notmyname | Zyric: have you had a chance to look at the auditor hook patch or slogging? | 23:30 |
| Zyric | notmyname: Yeah I have figured that out eventually. It's only the first week and a few days of learning and confusion are expected but it's still frustrating, The theory is also pretty dry so I think for the rest of the internship I'll do my best to learn what I need on the go as required so it's moving on time. | 23:31 |
| *** thumpba_ has joined #openstack-swift | 23:31 | |
| *** shakamunyi has quit IRC | 23:32 | |
| Zyric | notmyname: Sort of, I have tried but ran into some problems setting up slogging - namely creating a new account. I have read the code for the patch and slogging respository but it hasn't quite clicked yet. | 23:32 |
| notmyname | Zyric: maybe if he weren't such a jerk the maintainer of that repo would stop by and answer your questions ;-) | 23:33 |
| Zyric | notmyname: Heh, that would be nice. Speaking of questions, I haven't been very good at asking them this week, somewhat due to shyness and somewhat due to the fact the docs are pretty good! Thanks for all the great docs and video tutorials Swift Community :) | 23:36 |
| *** August1914 has quit IRC | 23:36 | |
| *** thumpba_ has quit IRC | 23:49 | |
| *** tongli has quit IRC | 23:50 | |
| gmmaha | notmyname: i sent you an email about the testr patch. when you get a chance, your thoughts on it would be awesome. | 23:53 |
| gmmaha | definitely not urgent. getting pulled down by end of the year administrative stuff.. :( | 23:53 |
| notmyname | gmmaha: actually that is pretty important. I do want to look at it again verys oon | 23:54 |
| mattoliverau | gmmaha: aren't we all :P | 23:54 |
| *** Jeffrey4l has quit IRC | 23:54 | |
| gmmaha | notmyname: :D | 23:54 |
| notmyname | already 4pm here and I was looking at a couple of other things, but I hope to look at it before monday | 23:54 |
| notmyname | we need it to land soon | 23:54 |
| gmmaha | mattoliverau: :D that too with barely 3 weeks left to close of the year | 23:54 |
| gmmaha | notmyname: got it. thanks a lot.. and hope i am not being a pain :D | 23:55 |
| *** setmason has quit IRC | 23:55 | |
| mattoliverau | I'll spend my Monday (your Sunday) looking at it :) | 23:55 |
| notmyname | quite the opposite | 23:55 |
| notmyname | mattoliverau: thanks | 23:55 |
| gmmaha | mattoliverau: woohoo.. awesome | 23:55 |
| mattoliverau | gmmaha: no thank you so much for all the awesome effort you've put it, it's been so usful! | 23:55 |
| gmmaha | mattoliverau: thanks :) i have a strong feeling the solution is "duh, you did it wrong" but thats good | 23:56 |
| gmmaha | :D | 23:56 |
| mattoliverau | I'm just sorry I was pulled away for most of this week, when I was coming to the Sydney office, I suspected (like in the past) I'd still have a chance to do actualy work.. but that kinda failed. | 23:56 |
| gmmaha | mattoliverau: aah travels are always fun!! never once have i gotten to do actual work when i travel | 23:57 |
| mattoliverau | lol | 23:57 |
| gmmaha | you are online and thats more than what i could have accomplished | 23:57 |
| *** lcurtis has quit IRC | 23:57 | |
| mattoliverau | well I'm in a cafe in Sydney ATM waiting for my wife to arrive, so trying to get some work in beofre then ;) | 23:58 |
| *** zhill has joined #openstack-swift | 23:58 | |
| gmmaha | thank god you dont have to sign into a VPN to work | 23:58 |
| * gmmaha needs to get himself a VPS hosted znc | 23:59 | |
| *** zhill has quit IRC | 23:59 | |
| mattoliverau | gmmaha: nah, I don't even use my rackspace email for openstack work. Everything in the open, just the way I like it :) | 23:59 |
| gmmaha | niceeee.... | 23:59 |
| notmyname | mattoliverau: I don't use my rackspace email either | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!