| *** rcernin_ has joined #openstack-swift | 00:04 | |
| *** rcernin has quit IRC | 00:07 | |
| *** rcernin_ has quit IRC | 00:19 | |
| DHE | timburke: 56 million objects and counting... | 00:22 |
|---|---|---|
| DHE | the target is 60 million or so | 00:23 |
| *** gyee has quit IRC | 00:24 | |
| *** rcernin has joined #openstack-swift | 00:27 | |
| timburke | out of curiosity, is it a one-time load, or is it expected to turn over somewhat like zuul-logs do? | 01:00 |
| DHE | most of it is a one time load, but I'm not expecting it to go over, say, 70 million in a year | 01:05 |
| DHE | .. still considered experimental? | 01:38 |
| timburke | it needs a bit of hand-holding. fwiw, we've got a bunch of customers using it, though, and (iirc) some of them have an order of magnitude more objects in the container | 02:20 |
| timburke | one-time load is kind of an ideal use-case for sharding, though. should be able to manually manage shard ranges a couple times, then mostly just leave it alone | 02:21 |
| *** diablo_rojo has quit IRC | 02:51 | |
| *** psachin has joined #openstack-swift | 03:32 | |
| *** diablo_rojo has joined #openstack-swift | 03:48 | |
| mattoliverau | if only we had auto-sharding finshed ;P | 04:04 |
| mattoliverau | tdasilva: I hope your not dying from all the smoke in Sydney. Sorry you come visit and then we surround the city your in with fire. At least we keep it exciting for you :P | 04:06 |
| tdasilva | hehe, feels like it's extra smokey today | 04:07 |
| tdasilva | but should get better later in the afternoon | 04:08 |
| tdasilva | weekend was pretty good thou, so that worked out well | 04:08 |
| tdasilva | mattoliverau: how is it for you? | 04:15 |
| mattoliverau | yeah not too bad. One of the main roads has reopened. Mornings are very smokey, but clears up by the arvo. Lot's of ash washing up on the beaches which is interesting | 04:27 |
| *** diablo_rojo has quit IRC | 05:00 | |
| viks___ | clayg: Ok.. thanks | 05:06 |
| *** new_student1411 has joined #openstack-swift | 05:08 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: WIP: prevent cont-sync when versioning enabled https://review.opendev.org/698139 | 05:35 |
| *** pawan-gupta has joined #openstack-swift | 06:37 | |
| *** psachin has quit IRC | 06:42 | |
| *** psachin has joined #openstack-swift | 06:50 | |
| *** diablo_rojo has joined #openstack-swift | 07:01 | |
| *** new_student1411 is now known as rahul__ | 07:11 | |
| *** rahul__ has quit IRC | 07:13 | |
| *** rs_goyal_ has joined #openstack-swift | 07:25 | |
| *** pcaruana has joined #openstack-swift | 07:39 | |
| *** rcernin has quit IRC | 07:54 | |
| *** tkajinam has quit IRC | 08:06 | |
| *** rs_goyal_ has quit IRC | 08:11 | |
| *** rdejoux has joined #openstack-swift | 08:13 | |
| *** tesseract has joined #openstack-swift | 08:27 | |
| *** rpittau|afk is now known as rpittau | 08:41 | |
| *** spsurya has joined #openstack-swift | 09:25 | |
| *** pawan-gupta has quit IRC | 10:15 | |
| *** spsurya has quit IRC | 12:35 | |
| *** rdejoux has quit IRC | 12:48 | |
| manuvakery | Hi.. I am having a small swift cluster (1 proxy and 3 storage nodes) . Each storage nodes has a capacity of 16GB RAM, 4 cpu cores and 11 HDDs( 2 for account and container servers and rest for object servers). For a 500 MB file upload response time ranges from 5.5 secs to 9 secs. Considering that we don't have much load on the cluster why the response time varies too much. The test is ran from a node which is | 13:04 |
| manuvakery | on the same network and we have 10gig network. Does the partition power has something to do with the performance. | 13:04 |
| *** rs_goyal has joined #openstack-swift | 13:20 | |
| *** rdejoux has joined #openstack-swift | 13:21 | |
| rs_goyal | I am trying to set an acl on object using s3 API. I have set `x-amz-acl` to `public-read` and tried to access the the same using url `https://s3.<domain>/<bucket-name>/file.txt`, but it gives 400 bad request. The `s3_acl` option is true and I can see the object acl to be set using third party tool. According to compatibility matrix(https://docs.openstack.org/swift/latest/s3_compat.html), `object/bucket` | 13:30 |
| rs_goyal | `put/get ACLs` should work. | 13:30 |
| alecuyer | manuvakery: assuming you're using a replication policy, your file data is not being split among different disks, but written entirely to multiple disks. That gives us 55 to 90MB/s in your case, which doesn't sound too bad? Check iostat on your object-server to confirm, you're probably close to 100% busy on those drives. If it's possible in your use(case, try to upload multiple files to get better overall performance | 13:43 |
| alecuyer | as to why it varies so much, 90MB/s may be close to the maximum your hard drive can deliver for sequential writes. even a light activity on the drive would slow it down as the heads need to seek. But I'm only guessing, check iostat output to confirm | 13:52 |
| *** tkajinam has joined #openstack-swift | 14:01 | |
| manuvakery | @alecuyer yes am using replication policy . Will try uploading multiple files and check iostat | 14:24 |
| manuvakery | the iostat uitl is not very high https://usercontent.irccloud-cdn.com/file/WzDwH0Bw/iostat.png | 14:26 |
| manuvakery | response times for multiples files , ran the script for 10 times https://usercontent.irccloud-cdn.com/file/7z0IlEJI/response.png | 14:29 |
| alecuyer | I meant, upload files in parallel, not sequentially (sorry if that's what you're doing already) - also make sure the object names are different | 14:33 |
| alecuyer | You never saw a peak in iostat while running the upload test ? indeed it does not look busy at all in your screenshot | 14:35 |
| *** pcaruana has quit IRC | 14:36 | |
| manuvakery | not much a peak in iostat so far . Will run some beanchmarking and monitor the iostat | 14:38 |
| manuvakery | @alecuyer running late for home . Will try and update :p | 14:40 |
| *** tkajinam has quit IRC | 14:52 | |
| alecuyer | manuvakery: no worries. I'm not always here but I'll catch up if you post test results. good luck! | 14:54 |
| *** pcaruana has joined #openstack-swift | 15:01 | |
| *** zaitcev has joined #openstack-swift | 15:27 | |
| *** ChanServ sets mode: +v zaitcev | 15:27 | |
| openstackgerrit | Romain LE DISEZ proposed openstack/swift master: Replace all "with Chunk*Timeout" by a watchdog https://review.opendev.org/697653 | 15:34 |
| *** gyee has joined #openstack-swift | 16:37 | |
| *** rdejoux has quit IRC | 16:49 | |
| *** tesseract has quit IRC | 16:58 | |
| *** rpittau is now known as rpittau|afk | 17:29 | |
| *** psachin has quit IRC | 19:02 | |
| *** efried is now known as efried_afk | 19:06 | |
| *** gmann is now known as gmann_afk | 19:40 | |
| *** rcernin has joined #openstack-swift | 20:16 | |
| *** pcaruana has quit IRC | 20:16 | |
| *** sorrison has quit IRC | 20:24 | |
| *** sorrison has joined #openstack-swift | 20:26 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Add proxy-server option to quote-wrap all ETags https://review.opendev.org/695131 | 21:24 |
| mattoliverau | morning | 22:13 |
| *** efried_afk is now known as efried | 22:25 | |
| *** tkajinam has joined #openstack-swift | 23:05 | |
| *** gmann_afk is now known as gmann | 23:06 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Have slo tell the object-server that it wants whole manifests https://review.opendev.org/697739 | 23:43 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
