| openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Fix up probetests check on feature/deep https://review.openstack.org/550291 | 00:02 |
|---|---|---|
| kei-ichi | sorry this is caused by my lower level of English... | 00:03 |
| kei-ichi | I think this is "my" misunderstanding..(and I could not express what I want to say) Once abandoned these. | 00:03 |
| kei-ichi | I don't want to make swift team in trouble, so if it is better not to locate project under swift, so is it better to create this as kind of top level project ? | 00:04 |
| *** Sukhdev has joined #openstack-swift | 00:04 | |
| kei-ichi | (like you told me about https://github.com/openstack/liberasurecode .) | 00:04 |
| kei-ichi | > I don't want to make swift team in trouble, so if it is better not to locate project under swift, so is it better to create this as kind of top level project ? | 00:05 |
| kei-ichi | sorry I made mistake in sending message ... | 00:05 |
| kei-ichi | I don't want to make swift team in trouble, so is it better not to locate project under swift ? | 00:06 |
| kei-ichi | I think that is okay, so I'll fix my patch. | 00:10 |
| kei-ichi | Sorry for your incovenience and my misunderstanding.. | 00:10 |
| notmyname | kei-ichi: I do not think you made a very big mistake at all, and you do not need to worry about getting the swift team in trouble | 00:14 |
| *** itlinux has quit IRC | 00:14 | |
| notmyname | placing slogging under the swift umbrella will work just fine, but it will add extra work for both you and me (and others). however if you have it as a separate project, like storlets or liberasurecode, then you will be able to have more control over what happens in the slogging project | 00:15 |
| notmyname | kota_ did this with both storlets and liberasurecode, so he should be able to help you (and in Japanese, if that helps) | 00:16 |
| notmyname | even if you add slogging as a separate project, you are welcome to use this IRC channel for discussing patches and asking questions | 00:17 |
| notmyname | I'm suggesting that you add slogging as a separate project so that you will be able to have a separate group of core reviewers to land patches. you will not have to wait for swift-core to review patches, and you will not have to wait on me as the swift ptl to do a release | 00:18 |
| notmyname | kei-ichi: I need to travel home now, but feel free to keep asking questions. maybe someone else can answer, or I will see them later when I'm online again | 00:19 |
| kei-ichi | Thank you for warm and kind advise for me! | 00:20 |
| kei-ichi | (I didn't kono liberasurecode and storlets is done by kota_ ) | 00:21 |
| kei-ichi | notmyname I think I can mange this ! I'll try, thanks! | 00:21 |
| *** gerardgine has quit IRC | 00:35 | |
| *** gerardgine has joined #openstack-swift | 00:36 | |
| openstackgerrit | Merged openstack/swift feature/deep: Add EXPANDING state to ShardRange https://review.openstack.org/551331 | 00:43 |
| *** gerardgine has quit IRC | 00:56 | |
| *** cshastri_ has joined #openstack-swift | 01:01 | |
| *** germs has joined #openstack-swift | 01:15 | |
| *** germs has quit IRC | 01:19 | |
| kota_ | good morning | 01:23 |
| mattoliverau | kota_: morning | 01:23 |
| kota_ | timburke: good work on https://review.openstack.org/#/c/552174/. one question, is it able to port to other in-process func tests? | 01:24 |
| patchbot | patch 552174 - swift - Capture probe test logs | 01:24 |
| kota_ | my try is at https://review.openstack.org/#/c/548226/ that uses another way from yours but your way looks smarter than mime. | 01:25 |
| patchbot | patch 548226 - swift - Pull execution logs of functests | 01:25 |
| kota_ | mattoliverau: o/ | 01:25 |
| kota_ | oh mattoliverau. do you have a chance to look at https://review.openstack.org/#/c/551956/ ? | 01:26 |
| patchbot | patch 551956 - python-swiftclient - Add force auth retry mode in swiftclient | 01:26 |
| kota_ | that is for a client issue when using sort of reverse proxy or load balancer. | 01:27 |
| kota_ | I'm sure it's NOT complete fix but it can mitigate the issue until fixing everything... | 01:27 |
| * kota_ is reading the back log... kei-ichi looks doing something... | 01:28 | |
| kei-ichi | kota_ good morning o/ | 01:29 |
| kota_ | kei-ichi: o/ | 01:29 |
| kei-ichi | I'm now try to create top-level project for slogging according to swift member's advice. | 01:29 |
| kei-ichi | I'm now try to create top-level project for slogging according to swift member's advice. | 01:29 |
| *** gerardgine has joined #openstack-swift | 01:30 | |
| kota_ | kei-ichi: got it. it's the reason why you abandon the first proposed patch set. | 01:30 |
| kei-ichi | The patch is for setting slogging under swift, so abandoned for resetting state. | 01:31 |
| kei-ichi | But I think the way to send patch is not so different. | 01:32 |
| *** kei_yama has quit IRC | 01:32 | |
| kota_ | okay | 01:33 |
| *** Sukhdev has quit IRC | 01:36 | |
| *** gerardgine has quit IRC | 01:42 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Try to set post logs in func tests https://review.openstack.org/552240 | 01:47 |
| *** zhongjun has joined #openstack-swift | 01:54 | |
| *** rcernin has quit IRC | 02:04 | |
| *** gerardgine has joined #openstack-swift | 02:19 | |
| *** kei_yama has joined #openstack-swift | 02:23 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Try to set post logs in func tests, too https://review.openstack.org/552240 | 02:27 |
| *** itlinux has joined #openstack-swift | 02:30 | |
| *** gerardgine has quit IRC | 02:45 | |
| *** psachin has joined #openstack-swift | 02:47 | |
| *** bkopilov has quit IRC | 03:05 | |
| *** rcernin has joined #openstack-swift | 03:07 | |
| *** germs has joined #openstack-swift | 03:15 | |
| *** germs has quit IRC | 03:15 | |
| *** germs has joined #openstack-swift | 03:15 | |
| *** germs has quit IRC | 03:20 | |
| mattoliverau | kota_: looks good, want to test, but there also seems to be a missing link in the commit message. number [4]. | 03:23 |
| kota_ | oh that was missing!? sorry. | 03:23 |
| kota_ | let me check. | 03:23 |
| kota_ | oh | 03:24 |
| kota_ | [4] should be https://review.openstack.org/#/c/69187/ | 03:25 |
| patchbot | patch 69187 - python-swiftclient - Port to python-requests (MERGED) | 03:25 |
| kota_ | mattoliverau: thx! I'll update soon | 03:25 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/python-swiftclient master: Add force auth retry mode in swiftclient https://review.openstack.org/551956 | 03:30 |
| kota_ | mattoliverau: ;) | 03:31 |
| mattoliverau | cool | 03:31 |
| openstackgerrit | Merged openstack/swift master: Optionally drop common prefixes in swift-*-info output https://review.openstack.org/342267 | 03:39 |
| openstackgerrit | Merged openstack/swift master: Capture probe test logs https://review.openstack.org/552174 | 03:39 |
| *** gerardgine has joined #openstack-swift | 03:44 | |
| *** links has joined #openstack-swift | 03:50 | |
| *** gyee has quit IRC | 03:59 | |
| *** bkopilov has joined #openstack-swift | 04:36 | |
| kei-ichi | I re-sent the review request to openstack-infra/project-config as https://review.openstack.org/#/c/552282/ . | 04:48 |
| patchbot | patch 552282 - openstack-infra/project-config - Add slogging projects | 04:48 |
| kei-ichi | Now slogging is defined as separated project from swift. | 04:48 |
| kei-ichi | Let me add as review of this. notmyname mattoliverau kota_ . If you have comments please let me know again. | 04:48 |
| kei-ichi | Thank you for your advise! o/ | 04:49 |
| *** gerardgine has quit IRC | 04:49 | |
| *** gerardgine has joined #openstack-swift | 04:50 | |
| kota_ | kei-ich: i think you can maintain the slogging-core team via gerrit's Web UI but not sure who are included in default in the list. probably ask at #openstack-infra channel is good way to make sure it. | 05:10 |
| kota_ | unfortunately every project i can see was created by another person. | 05:11 |
| *** germs has joined #openstack-swift | 05:16 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Try to set post logs in func tests, too https://review.openstack.org/552240 | 05:19 |
| *** germs has quit IRC | 05:21 | |
| kei-ichi | I heard from kazsh about that point. infra-team just create this group so I need to contact to infra-team to add myself as a member of this. | 05:22 |
| *** itlinux has quit IRC | 05:26 | |
| *** arunman has joined #openstack-swift | 05:32 | |
| kota_ | kei-ichi: good to know. sounds reasonable. | 05:35 |
| *** gerardgine has quit IRC | 06:16 | |
| *** gerardgine has joined #openstack-swift | 06:17 | |
| *** stewie925 has joined #openstack-swift | 06:21 | |
| *** germs has joined #openstack-swift | 06:29 | |
| *** germs has quit IRC | 06:29 | |
| *** germs has joined #openstack-swift | 06:29 | |
| *** tovin07_ has joined #openstack-swift | 06:48 | |
| *** d0ugal has joined #openstack-swift | 06:48 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: py3: port cli form_signature and unit tests https://review.openstack.org/552315 | 06:50 |
| *** germs has quit IRC | 06:52 | |
| mattoliverau | ^ that was an easy one. HMAC function in py3 expects binary strings. After that it just works.. though it's probably one of the most basic scripts we have :) | 06:54 |
| *** gerardgine has quit IRC | 06:56 | |
| *** neonpastor has quit IRC | 07:00 | |
| *** neonpastor has joined #openstack-swift | 07:01 | |
| kota_ | hmmm... I don't know why zuul doesn't like me to do post.yaml script with become: true even the master timburke wrote can be run in the gerrit. | 07:06 |
| kota_ | omg, centos environ for probe is under ansible 2.3.3 so that at least the import_playbook won't work anyway. | 07:21 |
| *** rcernin has quit IRC | 07:24 | |
| kota_ | my 4th attempt to get functional test logs... | 07:27 |
| kota_ | i really want the stuff in the master because sometimes s3api can fail with the gate but not in my local :/ | 07:27 |
| *** armaan has quit IRC | 07:28 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 07:34 |
| *** pcaruana has joined #openstack-swift | 07:42 | |
| *** armaan has joined #openstack-swift | 07:44 | |
| *** hseipp has joined #openstack-swift | 07:46 | |
| *** links has quit IRC | 07:52 | |
| *** links has joined #openstack-swift | 07:59 | |
| *** joeljwright has joined #openstack-swift | 08:10 | |
| *** ChanServ sets mode: +v joeljwright | 08:10 | |
| *** gkadam has joined #openstack-swift | 08:11 | |
| *** tesseract has joined #openstack-swift | 08:19 | |
| *** tesseract-RH has joined #openstack-swift | 08:20 | |
| *** tesseract-RH has quit IRC | 08:20 | |
| openstackgerrit | Thomas Goirand proposed openstack/swift master: Fix tests using O_TMPFILE https://review.openstack.org/552093 | 08:39 |
| *** gkadam is now known as gkadam-brb | 08:44 | |
| *** gkadam-brb is now known as gkdam | 08:52 | |
| *** gkdam is now known as gkadam | 08:52 | |
| acoles | good morning | 09:16 |
| *** mvk has quit IRC | 09:32 | |
| arunman | timburke: I'm trying to find a way on how to delete the files under /srv/node/partition1/objects/*. We've hit an issue where we run out of inodes and not able to store any new objects. I tried to check if container-reconciler is cleaning this up based on the reclaim_age property. Looks like it's not. Wanted to check if there's something that I could be missing! | 09:51 |
| *** tovin07_ has quit IRC | 09:59 | |
| *** kei_yama has quit IRC | 10:33 | |
| openstackgerrit | Alistair Coles proposed openstack/swift feature/deep: Rename dummy_shard_range to own_shard_range https://review.openstack.org/552472 | 10:37 |
| *** cshastri_ has quit IRC | 10:38 | |
| *** pcaruana has quit IRC | 10:45 | |
| *** mvk has joined #openstack-swift | 11:00 | |
| *** rcernin has joined #openstack-swift | 11:02 | |
| openstackgerrit | Thomas Goirand proposed openstack/swift master: Fix tests using O_TMPFILE https://review.openstack.org/552093 | 11:04 |
| *** geaaru has joined #openstack-swift | 11:17 | |
| *** arunman has quit IRC | 11:18 | |
| *** bkopilov has quit IRC | 11:30 | |
| *** rcernin has quit IRC | 12:00 | |
| openstackgerrit | Ondřej Nový proposed openstack/swift master: Fix tests using O_TMPFILE https://review.openstack.org/552093 | 12:06 |
| *** arunman has joined #openstack-swift | 12:15 | |
| *** arunman has quit IRC | 12:31 | |
| *** zhurong has joined #openstack-swift | 12:35 | |
| *** zhurong has quit IRC | 12:42 | |
| *** zhurong has joined #openstack-swift | 12:45 | |
| *** arunman has joined #openstack-swift | 12:55 | |
| *** hseipp has quit IRC | 13:03 | |
| *** psachin has quit IRC | 13:07 | |
| *** HW-Peter has joined #openstack-swift | 13:08 | |
| *** HW-Peter has quit IRC | 13:10 | |
| *** HW-Peter has joined #openstack-swift | 13:11 | |
| *** zhurong has quit IRC | 13:15 | |
| *** spacial_being has joined #openstack-swift | 13:26 | |
| *** spacial_being has quit IRC | 13:28 | |
| openstackgerrit | Merged openstack/swift feature/deep: Rename dummy_shard_range to own_shard_range https://review.openstack.org/552472 | 13:32 |
| *** arunman has quit IRC | 13:36 | |
| *** armaan has quit IRC | 13:38 | |
| *** armaan has joined #openstack-swift | 13:38 | |
| *** arunman has joined #openstack-swift | 13:39 | |
| DHE | kinda crazy question. has anybody tried swift on top of ZFS? either in a RAID configuration or not... | 13:42 |
| *** bkopilov has joined #openstack-swift | 13:42 | |
| *** armaan has quit IRC | 13:43 | |
| *** mingyu has joined #openstack-swift | 14:03 | |
| *** arunman has quit IRC | 14:27 | |
| zigo | notmyname: timburke: My patch passes zuul tests: https://review.openstack.org/#/c/552093/ | 14:37 |
| patchbot | patch 552093 - swift - Fix tests using O_TMPFILE | 14:37 |
| *** gerardgine has joined #openstack-swift | 14:42 | |
| *** armaan has joined #openstack-swift | 14:43 | |
| *** gerardgine has quit IRC | 14:43 | |
| openstackgerrit | Alexandre Lécuyer proposed openstack/swift master: Add optional parameters to common/utils fallocate() https://review.openstack.org/548034 | 14:53 |
| *** hseipp has joined #openstack-swift | 15:24 | |
| *** cshastri_ has joined #openstack-swift | 15:28 | |
| root_743 | hello, is there any way to rebuild my rings? all my data got quarantined because we added swift_hash_path_prefix after some objects where already in place and now I had to delete that parameter but still no luck with the data, I tried to move the objects from /srv/node/d1/quarantined/objects to /srv/node/d1/objects but nothing :( | 15:29 |
| *** itlinux has joined #openstack-swift | 15:45 | |
| *** arunman has joined #openstack-swift | 15:45 | |
| *** gyee has joined #openstack-swift | 15:51 | |
| *** links has quit IRC | 15:55 | |
| notmyname | DHE: yes, swift on ZFS has been tried, but it's been a while. IIRC last time it was tried, ZFS performance wasn't as good as XFS, especially as the drive fills up (inodes more than bytes) | 15:55 |
| *** links has joined #openstack-swift | 16:08 | |
| *** armaan has quit IRC | 16:14 | |
| *** armaan has joined #openstack-swift | 16:15 | |
| timburke | kota_: good question -- i'm not sure. i did it that way because of how the probe test job sets up its logging. in particular, it grabs https://github.com/openstack/swift/blob/master/doc/saio/rsyslog.d/10-swift.conf | 16:18 |
| arunman | timburke: Any thoughts on my query above? Not sure if you missed it! | 16:26 |
| *** links has quit IRC | 16:30 | |
| timburke | arunman: i *had* missed it! i'm maybe not the best for fielding out-of-inodes questions (torgomatic or clayg might be better) but i wouldn't expect the container-reconciler to do much to clean up reclaimable tombstones -- woulda thought the object-replicator/reconstructor (or even object-auditor) would be more the place for it | 16:36 |
| arunman | ah.. sure timburke | 16:38 |
| arunman | torgomatic, clayg : do you have any suggestions on handling inodes related issues from my query above? | 16:40 |
| timburke | fwiw, https://github.com/openstack/swift/blob/2.17.0/swift/obj/diskfile.py#L1002-L1034 does the cleaning up... | 16:40 |
| notmyname | out of inodes is hard. you want to delete stuff, obviously, but a delete in swift *creates* a new file (and thus an inode). replication processes are what clean up data and move it off of a drive, if necessary | 16:42 |
| arunman | ok. i thought it's the container-reconciler that will take care of it. | 16:44 |
| arunman | so in this case, i assume the reclaim_age value is getting picked up from *container-replicator* section in container-server.conf file | 16:44 |
| arunman | is my understanding correct? | 16:45 |
| notmyname | the container reconciler detects when a container has been created with 2 different storage policies and it reconciles any objects in the 2 copies and puts them all in the correct storage policy | 16:45 |
| arunman | alright | 16:46 |
| notmyname | arunman: yes. the reclaim age is used to determine when deleted data is removed from the drives. but are you just looking at container data? are the full drives only in the container rings? | 16:46 |
| arunman | my issue is with the objects data. i don't see this issue with container data | 16:48 |
| *** hseipp has quit IRC | 16:48 | |
| notmyname | ok. then take all that stuff about reclaim age and replicators and s/container/object :-) | 16:48 |
| *** hseipp has joined #openstack-swift | 16:48 | |
| notmyname | is your object data in a replicated policy or erasure coded policy? | 16:49 |
| notmyname | how many drives in your cluster are full? how many have available space? | 16:49 |
| DHE | notmyname: interesting.. I was hoping ZFS performance could offer advantages like no need for an object auditor, and of course compression | 16:51 |
| arunman | notmyname: I'm sorry... but, I don't think we have configured any of the policies that you are referring to. This is a basic all in one setup with a single drive config. | 16:51 |
| notmyname | DHE: compression is very dependent on what sort of data you're storing. if you're doing videos, pics, or backups (3 very very common swift use cases), then compression will hurt rather than help. the data's already compressed | 16:52 |
| DHE | yes I know... and ZFS does have algorithms which are intended to set and forget... | 16:53 |
| notmyname | DHE: the lack of auditing is nice, of course, but the real issue that I remember being reported is actually arunman's current one. if you take a zfs volume and fill it up (inode-wise) with a bunch of small files, what happens? when does it break. what I remember is that zfs breaks way before xfs | 16:53 |
| arunman | notmyname: now it turns out that my understanding about reconciler is wrong and i need to check further based on your suggestion about reclaim age and replicators | 16:54 |
| notmyname | arunman: all policies are either using replication or erasure coding. so if I understand what you're saying, you've got one machine with one drive? | 16:54 |
| arunman | yep.. that's right | 16:54 |
| *** gkadam has quit IRC | 16:54 | |
| notmyname | and your rings are set up with a single replica then? | 16:55 |
| arunman | yes | 16:55 |
| DHE | notmyname: ZFS breaks when it gets space full because the free space allocator ends up spending lots of time hunting for free space. in terms of inodes, ZFS has no fixed inode allocations. if you run out of inodes, it's because you're full anyway. | 16:55 |
| notmyname | so in your case, there's absolutely nothing that the replicator (or reconciler) processes can even do. you've got one replica of everything and in only one place | 16:55 |
| notmyname | arunman: just curious, but with that sort of setup you don't have any durability or availablity guarantees. why are you using swift instead of simply writing directly to your one hard drive? | 16:56 |
| notmyname | DHE: ah ok. like I said, it's been a while since I've looked at it. how does zfs handle xattrs? are there any limits? | 16:57 |
| notmyname | eg ext supports xattrs that fit in one inode (ie block). xfs supports more than that | 16:58 |
| DHE | there are no practical limits. small xattrs can be stored in the inode itself if they fit and the feature is enabled, otherwise a sort of "xattr inode" is allocated for the file | 16:58 |
| notmyname | cool | 16:58 |
| DHE | from what I read xfs does something similar | 16:58 |
| arunman | notmyname: we had set this up to store vm's metadata to handle certain use cases and this metadata is a bare minimal data. that's the reason we decided to do it this way so that this can be read from compute services by talking to swift. | 17:02 |
| arunman | with what you say, looks like i may have to revisit the current configuration | 17:03 |
| notmyname | arunman: ok. like I said, just curious. | 17:04 |
| notmyname | arunman: the simplest thing to "our cluster is out of space" is to add more hard drives. in your case, adding a few and going to at least 2 replicas would end up giving you a lot of benefit (durability, extra space, availability when hardware fails) | 17:06 |
| *** d0ugal has quit IRC | 17:09 | |
| arunman | notmyname: sure. i'll check that out | 17:10 |
| *** cshastri_ has quit IRC | 17:14 | |
| *** gerardgine has joined #openstack-swift | 17:20 | |
| *** hseipp has quit IRC | 17:26 | |
| notmyname | tdasilva: when I do something like `yum install epel-release` is it actually installing anything? or `yum install centos-release-openstack-queens` | 17:31 |
| notmyname | I'm looking for the equivalent of `apt-add-repository` | 17:32 |
| *** geaaru has quit IRC | 17:33 | |
| *** arunman has quit IRC | 17:42 | |
| *** mvk has quit IRC | 17:43 | |
| openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Merge remote-tracking branch 'gerrit/master' into feature/deep https://review.openstack.org/552644 | 17:58 |
| tdasilva | notmyname: sorry, stuck in meetings...don't know apt-add-repository, but it sounds exactly like the same thing... | 18:04 |
| tdasilva | yum install epel-release is just setting up the epel repo | 18:04 |
| notmyname | ok | 18:04 |
| notmyname | but it's the same command as installing some package right? | 18:05 |
| notmyname | that's what's confusing | 18:05 |
| notmyname | (to me) | 18:05 |
| notmyname | current struggle is that `yum install python-pip` works, but then `pip install ...` doesn't (pip:command not found) | 18:05 |
| tdasilva | yeah, you need to do `yum install python-pip` first before using pip itself... | 18:06 |
| notmyname | tdasilva: you're not making fun of me, are you? | 18:08 |
| notmyname | ;-) | 18:08 |
| tdasilva | lol | 18:08 |
| notmyname | I mean, yes, I get that I need to install it before using it | 18:09 |
| *** hseipp has joined #openstack-swift | 18:09 | |
| tdasilva | oh, you are saying that after install python-pip you are not able to run pip? | 18:09 |
| tdasilva | that would be weird | 18:09 |
| notmyname | https://gist.github.com/notmyname/53a2e559b50ad76f74822af2ad754064 | 18:10 |
| tdasilva | one sec | 18:12 |
| *** gerardgine has quit IRC | 18:12 | |
| tdasilva | notmyname: trying to run your example here... | 18:23 |
| notmyname | I fully expect that I'm missing something basic | 18:23 |
| tdasilva | notmyname: do you need to install from source? can't install rdo's swift? | 18:24 |
| notmyname | that's correct | 18:24 |
| notmyname | tdasilva: FWIW, here's my full (current) dockerfile https://gist.github.com/notmyname/d200071e6338794d7cc39fd3972713a9 | 18:24 |
| *** mvk has joined #openstack-swift | 18:28 | |
| tdasilva | notmyname: this worked for me: https://paste.fedoraproject.org/paste/0sZ5nVVwd6~wyzV~GTeTZw | 18:35 |
| notmyname | ok, thanks. I'll tray that | 18:35 |
| *** d0ugal has joined #openstack-swift | 18:39 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift master: Don't double-filter replication jobs https://review.openstack.org/551005 | 18:47 |
| openstackgerrit | Samuel Merritt proposed openstack/swift master: Remove object replicator's lockup detector/mitigator. https://review.openstack.org/552659 | 18:47 |
| clayg | swift/playbooks !? | 18:48 |
| tdasilva | ansible everything! | 18:48 |
| tdasilva | and according to notmyname use that to deploy in containers | 18:49 |
| tdasilva | just missing k8s | 18:49 |
| clayg | blockchain | 18:49 |
| notmyname | dockerchain.ai | 18:50 |
| tdasilva | heh, need to work that in there ;) | 18:50 |
| notmyname | (unfortunately taken) | 18:50 |
| tdasilva | but talking about blockchain, saw this yesterday: https://storj.io | 18:50 |
| *** gerardgine has joined #openstack-swift | 18:51 | |
| notmyname | is `lsb_release` a thing on rhel distros? | 18:55 |
| notmyname | hmm.. yes it is. but "Unable to execute lsb_release. Is it installed?" | 18:58 |
| notmyname | I need to install swift dependencies. to do that I need bindep. to do that I need pip. to do that I need easy_install. to run bindep I need lsb_release. I guess to get bindep's dependencies I need ... bindep? I dunno | 18:59 |
| notmyname | have I ever mentioned that computers are terrible? | 18:59 |
| tdasilva | lol | 18:59 |
| *** itlinux has quit IRC | 18:59 | |
| tdasilva | you might need to install redhat-lsb ?? | 19:00 |
| notmyname | yeah, that's what google just told me :/ | 19:00 |
| tdasilva | https://github.com/thiagodasilva/ansible-saio/blob/master/roles/pre_config/tasks/main.yml | 19:00 |
| notmyname | -lsb-core actually | 19:00 |
| notmyname | yeah, I think that's what should be in the playbook in the swift repo, TBH | 19:01 |
| notmyname | I should install git, checkout swift, install ansible, and run the playbooks in swift. done | 19:01 |
| notmyname | ideally doing some self-discovery with bindep.txt and requirements.txt | 19:02 |
| tdasilva | notmyname: lost you...fwiw, my understanding is that infra already takes care of doing the git checkout and bindep installation | 19:03 |
| tdasilva | but this is for gate jobs... | 19:03 |
| notmyname | yeah, I want to be able to take the repo and set it up without relying on -infra bootstrapping | 19:03 |
| notmyname | I'm not doing this for gate jobs | 19:03 |
| *** d0ugal has quit IRC | 19:04 | |
| notmyname | maybe eventually what I'm doing will be incorporated into some gates. actually, that's the plan. but it needs to run on my dev box | 19:04 |
| tdasilva | right...understood....for the playbooks in the swift repo, i started with https://github.com/thiagodasilva/ansible-saio/blob/master/roles/pre_config/tasks/main.yml and then realized that they do a bunch of bootstrapping so i just started removing things | 19:04 |
| notmyname | yeah, what you did makes sense, and I love it | 19:04 |
| *** itlinux has joined #openstack-swift | 19:05 | |
| tdasilva | so you want to run swift in a container with 1 replica? | 19:05 |
| notmyname | in my current case, I need to test a third-party thing that depends on a swift endpoint being available. it's currently doing this with a docker container, but there's a lot of hard-coding of stuff in it. I'm updating it to use whatever's current in swift, dynamically | 19:07 |
| notmyname | so it's part of some CI testing | 19:07 |
| tdasilva | got it | 19:07 |
| notmyname | and I do want to use the new zuul v3 third-party stuff to put it into our check queues | 19:07 |
| *** tesseract has quit IRC | 19:11 | |
| notmyname | more success, so far! | 19:11 |
| notmyname | but this is wierd: "Error mounting /mnt/sdb1: mount: /srv/swift-disk: failed to setup loop device: No such file or directory\n" | 19:15 |
| openstackgerrit | Alistair Coles proposed openstack/swift feature/deep: Use state rather than subtlety to make shards of shards authoritative https://review.openstack.org/551356 | 19:25 |
| openstackgerrit | Alistair Coles proposed openstack/swift feature/deep: Root leader decides when shard containers shard https://review.openstack.org/552120 | 19:25 |
| openstackgerrit | Alistair Coles proposed openstack/swift feature/deep: Store epoch and root container in shard ranges table https://review.openstack.org/552668 | 19:25 |
| *** d0ugal has joined #openstack-swift | 19:26 | |
| *** d0ugal has quit IRC | 19:26 | |
| *** d0ugal has joined #openstack-swift | 19:26 | |
| openstackgerrit | Alistair Coles proposed openstack/swift feature/deep: Remove unnecessary timestamp shenanigans when creating shard container https://review.openstack.org/552133 | 19:27 |
| *** hseipp has quit IRC | 19:58 | |
| tdasilva | notmyname: can you share your latest Dockerbuild? | 20:06 |
| tdasilva | although it sounds like you are running into a problem with the ansible playbook | 20:07 |
| notmyname | right. the dockerfile isn't really too different than what you saw (I just added a WORKDIR instead of RUN cd ... &&) | 20:08 |
| notmyname | my emergency valve is to rewrite the upstream playbooks to separate out the drive creation (and then not call it in my dockerfile) | 20:08 |
| notmyname | but I'm hoping not to have to do that | 20:09 |
| notmyname | surely there must be a way to enable a docker container to do a loopback mount on a local file | 20:09 |
| openstackgerrit | Merged openstack/pyeclib master: Switch from pep8 to flake8 for linting https://review.openstack.org/462304 | 20:15 |
| *** armaan has quit IRC | 20:38 | |
| *** armaan has joined #openstack-swift | 20:39 | |
| *** d0ugal has quit IRC | 20:45 | |
| *** gerardgine has quit IRC | 20:47 | |
| *** armaan has quit IRC | 20:57 | |
| *** mingyu has quit IRC | 20:58 | |
| *** armaan has joined #openstack-swift | 20:59 | |
| *** mingyu has joined #openstack-swift | 21:00 | |
| *** mingyu has quit IRC | 21:13 | |
| *** mingyu has joined #openstack-swift | 21:13 | |
| *** geaaru has joined #openstack-swift | 21:13 | |
| *** d0ugal has joined #openstack-swift | 21:18 | |
| *** d0ugal has quit IRC | 21:18 | |
| *** d0ugal has joined #openstack-swift | 21:18 | |
| *** gerardgine has joined #openstack-swift | 21:25 | |
| *** d0ugal has quit IRC | 21:37 | |
| *** rcernin has joined #openstack-swift | 21:59 | |
| *** geaaru has quit IRC | 22:00 | |
| *** armaan has quit IRC | 22:05 | |
| *** armaan has joined #openstack-swift | 22:06 | |
| *** armaan has quit IRC | 22:07 | |
| *** armaan has joined #openstack-swift | 22:08 | |
| *** gerardgine has quit IRC | 22:13 | |
| *** gerardgine has joined #openstack-swift | 22:14 | |
| clayg | notmyname: i don't think you can do the loopback in the container - that's the kernel's business :\ | 22:14 |
| clayg | you can expose a loopback *to* the container? | 22:15 |
| openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Fix up probetests check on feature/deep https://review.openstack.org/550291 | 22:28 |
| openstackgerrit | Samuel Merritt proposed openstack/swift master: Augment object-expirer's StatsD buckets. https://review.openstack.org/538016 | 22:29 |
| openstackgerrit | Merged openstack/swift feature/deep: Merge remote-tracking branch 'gerrit/master' into feature/deep https://review.openstack.org/552644 | 22:35 |
| *** gerardgine has quit IRC | 22:35 | |
| *** gerardgine has joined #openstack-swift | 22:40 | |
| *** zaitcev_ has joined #openstack-swift | 22:44 | |
| *** ChanServ sets mode: +v zaitcev_ | 22:44 | |
| *** zaitcev has quit IRC | 22:48 | |
| *** d0ugal has joined #openstack-swift | 23:03 | |
| *** gerardgine has quit IRC | 23:08 | |
| *** gerardgine has joined #openstack-swift | 23:12 | |
| openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Fix up probetests check on feature/deep https://review.openstack.org/550291 | 23:14 |
| *** kei_yama has joined #openstack-swift | 23:22 | |
| *** gerardgine has quit IRC | 23:37 | |
| *** gerardgine has joined #openstack-swift | 23:44 | |
| openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Fix up probetests check on feature/deep https://review.openstack.org/550291 | 23:50 |
| openstackgerrit | Merged openstack/swift master: Don't double-filter replication jobs https://review.openstack.org/551005 | 23:55 |
| *** germs has joined #openstack-swift | 23:55 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!