| *** gatuus has joined #openstack-swift | 00:02 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Remove some cruft in test_proxy_logging https://review.openstack.org/426978 | 00:05 |
|---|---|---|
| *** wasmum has joined #openstack-swift | 00:07 | |
| *** _JZ_ has quit IRC | 00:11 | |
| *** csmart has quit IRC | 00:20 | |
| *** vint_bra has quit IRC | 00:21 | |
| *** ukaynar has joined #openstack-swift | 00:44 | |
| *** jamielennox is now known as jamielennox|away | 00:58 | |
| *** catintheroof has quit IRC | 01:02 | |
| *** gatuus has quit IRC | 01:07 | |
| *** jamielennox|away is now known as jamielennox | 01:12 | |
| kota_ | good morning | 01:19 |
| *** tqtran has quit IRC | 01:25 | |
| *** tqtran has joined #openstack-swift | 01:27 | |
| *** tqtran has quit IRC | 01:32 | |
| *** dims has quit IRC | 01:35 | |
| *** dims_ has joined #openstack-swift | 01:35 | |
| *** ukaynar has quit IRC | 01:37 | |
| *** catintheroof has joined #openstack-swift | 01:37 | |
| *** catintheroof has quit IRC | 01:37 | |
| *** catintheroof has joined #openstack-swift | 01:37 | |
| *** dims_ has quit IRC | 01:42 | |
| *** dims has joined #openstack-swift | 01:42 | |
| *** newmember has joined #openstack-swift | 01:45 | |
| mattoliverau | kota_: morning, how you feeling today? | 01:45 |
| kota_ | mattoliverau: better than tommorow thanks | 01:46 |
| kota_ | no | 01:47 |
| kota_ | yesterday | 01:47 |
| kota_ | I don't know why I said tommorow :\ | 01:47 |
| kota_ | it's time machine | 01:47 |
| JimCheung | Kota_: glad you're better now ;) | 01:47 |
| kota_ | my head seems still sick :/ | 01:47 |
| JimCheung | Flu? | 01:48 |
| mattoliverau | kota_: lol, or you let slip that you are kota_ from 2 days in the future, sent back to finish some work.. you ntt guys have some fun tech ;P | 01:48 |
| kota_ | JimCheung: thanks, yeah, I was a flu patient. | 01:49 |
| JimCheung | Take it easy! | 01:49 |
| kota_ | that's the hardest one once ever | 01:49 |
| JimCheung | It knocked you out for almost a week! | 01:50 |
| kota_ | JimCheung: exactly and it also make my family down :,( | 01:51 |
| JimCheung | Kota_: Sorry to hear! Don't stress to much! | 01:52 |
| kota_ | JimCheung: np | 01:53 |
| JimCheung | Kota_: I have a quick question tho. Did you put together a test suite for SHSS? We've submitted the code for our product Phazr and would like some tips on how to submit a test suite for verification. If not, no worries. | 01:54 |
| JimCheung | Kota_: This is the code review. https://review.openstack.org/#/c/424353/ | 01:55 |
| patchbot | patch 424353 - liberasurecode - Add Phazr.IO libphazr backend to liberasurecode | 01:55 |
| kota_ | JimCheung: will look at | 02:02 |
| kota_ | JimCheung: actually, for now we don't have online gate for our shss | 02:02 |
| JimCheung | Kota_: No problem. | 02:04 |
| kota_ | JimCheung: is your backend available for any users? | 02:04 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Better optimistic lock in get_hashes https://review.openstack.org/419787 | 02:04 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Fix race when consolidating new partition https://review.openstack.org/426336 | 02:04 |
| JimCheung | Eventually, yes | 02:04 |
| clayg | timburke: sorry I was wrong, i'll respin the EC stuff tomorrow | 02:04 |
| timburke | clayg: huh? which part? | 02:05 |
| clayg | kota_: looking forward to going back over ec_duplication - sorry it's taking so long! | 02:05 |
| clayg | timburke: the reconstructor changes | 02:05 |
| timburke | yeah, the zero-weight thing makes me really nervous the more i think about it | 02:06 |
| clayg | I was moving real slow on the suffix hashing stuff - so I didn't get to your comments | 02:06 |
| kota_ | clayg: thanks! | 02:06 |
| clayg | timburke: why? what does the replicator do in this case? | 02:06 |
| clayg | timburke: the existing code fucking *sucks* operationally | 02:06 |
| kota_ | JimCheung: ok, great. if it's available for gate, i have an idea to set up it to the gate. | 02:06 |
| clayg | timburke: I think you maybe hope it was a good idea because me and pluse wrote it that way - but it sucks - it just sucks balls hard - it seemed like a good idea at the time - but nope - it sucks ;) | 02:07 |
| JimCheung | Kota_: Cool! | 02:07 |
| kota_ | clayg: do you have recomendation (or what you want) for my eyes to re-start my upstream review work? | 02:07 |
| kota_ | clayg: in recent activities. | 02:07 |
| kota_ | it looks optimize get_hashes or so on... | 02:08 |
| clayg | kota_: the only things I'm tracking are the suffix hashing and EC reconstructor optimization | 02:08 |
| kota_ | is going | 02:08 |
| kota_ | kk | 02:08 |
| clayg | Maybe ask acoles_ if there's anything you can do to help with https://review.openstack.org/#/c/419787 when he gets on | 02:09 |
| patchbot | patch 419787 - swift - Better optimistic lock in get_hashes | 02:09 |
| kota_ | clayg: ok, will do | 02:09 |
| clayg | kota_: you worked on the EC disk/progress/partition reporting - maybe you can help with patch 425468 | 02:11 |
| patchbot | https://review.openstack.org/#/c/425468/ - swift - Shuffle disks and parts in reconstructor | 02:11 |
| timburke | clayg: but like, it's at least going to move *somewhere*, and that should be somewhere *earlier in the get_more_nodes iter*. there's something like progress, even if it sucks operationally. i don't disagree that it should be smarter about waiting to get it there in fewer hops, but as the patch stands now, i *can't decommission drives* because i have to hit *every primary* | 02:11 |
| *** klrmn has quit IRC | 02:12 | |
| clayg | timburke: revert jobs only go to the one priarmy - the frag_index means there is only ever one spot for any give part-fragment - the node_index in the primary in the iter | 02:12 |
| timburke | haven't looked at the replicator yet. any way we can look at the responses, see that hey, a bunch of these are telling me to cool my heels, and think maybe we should listen? | 02:12 |
| kota_ | clayg: it sounds nice patch, the title is what I'd like to get in the next from my global ec :P | 02:13 |
| clayg | and if you have a part that needs to go to diskA and it's on diskB (draiing) and diskA is offline you should fail diskA | 02:13 |
| clayg | in replicated land you could just unmount it - but reconstructor fail in place (507) handling is crap | 02:13 |
| kota_ | i know, the reconstruct_fa gather nodes by node index order. | 02:13 |
| timburke | so sync_to will only ever have one node? | 02:14 |
| timburke | maybe i'm misunderstanding | 02:14 |
| clayg | timburke: maybe down the road - but the existing revert to handoff behavior is so wrong it is *crippling* to EC rebalance - like that shit doesn't work - it just spins it's wheels - it's broken - it was a mistake - the replicator doesn't do this | 02:14 |
| clayg | the only time the replicator will ship a part-replica somewhere else is if the remote 507's - but *mostly* if it can't get to the primaries it just holds it - and that works pretty well when there's contention during rebalance at scale | 02:15 |
| clayg | for revert jobs - yeah sync_to just has the one primary - the one where that frag_index goes | 02:15 |
| clayg | timburke: so step 1 - do NOT revert to handoff ; step 2 - find out if it is *ever* a good idea to revert to a handoff and only do it under those very specific and limited circumstances recognizing no action is better than mis-action when you have 100K parts that need to revert | 02:17 |
| clayg | at least - that's how I'm selling it | 02:17 |
| timburke | https://github.com/openstack/swift/blob/2.12.0/swift/obj/reconstructor.py#L783-L791 ? | 02:17 |
| clayg | I don't think you're thinking about it *wrong* - I think you're just not as freaked/out/pissed-off as I am about the terrible terrible side effect of revert to handoff as written | 02:17 |
| clayg | timburke: it's a rare case of a part that only has tombstones - i shouldn't have said "only ever has one* | 02:18 |
| patchbot | Error: No closing quotation | 02:18 |
| clayg | it is very rare to have a part with just tombstones - mostly I don't care that much what it does in that case - i guess i'd like it to talk to sync tombstones to some subset of the primaries > ~3 and then delete the part | 02:19 |
| *** JimCheung has quit IRC | 02:23 | |
| timburke | aight, check: replicator only ever tries to send handoffs to primaries, never handoffs to handoffs | 02:25 |
| kota_ | timburke: IIRC, it's correct. handoff should not make more handoffs | 02:26 |
| kota_ | timburke: update_deleted only has a for-loop with job['nodes'] and not call get_more_nodes | 02:27 |
| *** dmorita has joined #openstack-swift | 02:28 | |
| *** ukaynar has joined #openstack-swift | 02:29 | |
| clayg | i thought the HTTP_INSUFFICIENT_STORAGE stuff in update was mirroed in update_deleted - but I guess it never really looks at the results of the REPLICATE response | 02:30 |
| clayg | i guess at some poin you just gotta remove those failed devices form the ring :\ | 02:31 |
| clayg | i gotta run | 02:31 |
| kota_ | bah :/ | 02:32 |
| clayg | timburke: thanks for think about those changes - i hope it will all be more clear in the morning | 02:32 |
| openstackgerrit | Tim Burke proposed openstack/swift: Do not revert fragments to handoffs https://review.openstack.org/425441 | 02:37 |
| *** ukaynar_ has joined #openstack-swift | 02:43 | |
| *** ukaynar has quit IRC | 02:43 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Do not revert fragments to handoffs https://review.openstack.org/425441 | 02:46 |
| *** winggundamth_ has joined #openstack-swift | 03:02 | |
| *** ukaynar_ has quit IRC | 03:03 | |
| *** ukaynar has joined #openstack-swift | 03:03 | |
| *** newmember has quit IRC | 03:05 | |
| *** newmember has joined #openstack-swift | 03:05 | |
| *** catintheroof has quit IRC | 03:09 | |
| *** catintheroof has joined #openstack-swift | 03:11 | |
| *** catintheroof has quit IRC | 03:15 | |
| *** chosafine has joined #openstack-swift | 03:20 | |
| *** chosafine has joined #openstack-swift | 03:20 | |
| *** bkopilov has quit IRC | 03:24 | |
| *** klrmn has joined #openstack-swift | 03:52 | |
| *** newmember has quit IRC | 03:55 | |
| *** psachin has joined #openstack-swift | 03:55 | |
| *** newmember has joined #openstack-swift | 03:56 | |
| *** takashi has joined #openstack-swift | 03:59 | |
| *** klrmn has quit IRC | 04:07 | |
| *** klrmn has joined #openstack-swift | 04:26 | |
| *** dmorita has quit IRC | 04:28 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:37 | |
| *** JimCheung has joined #openstack-swift | 04:42 | |
| *** bkopilov has joined #openstack-swift | 04:47 | |
| *** JimCheung has quit IRC | 04:47 | |
| *** sams-gleb has joined #openstack-swift | 04:49 | |
| *** sams-gleb has quit IRC | 04:54 | |
| *** ppai has joined #openstack-swift | 05:06 | |
| *** m_kazuhiro has joined #openstack-swift | 05:26 | |
| *** m_kazuhiro has quit IRC | 05:27 | |
| *** takashi has quit IRC | 05:30 | |
| *** klrmn has quit IRC | 05:35 | |
| *** sams-gleb has joined #openstack-swift | 05:51 | |
| *** sams-gleb has quit IRC | 05:56 | |
| *** janonymous has joined #openstack-swift | 06:02 | |
| *** dmorita has joined #openstack-swift | 06:28 | |
| *** chosafine has quit IRC | 06:31 | |
| *** dmorita has quit IRC | 06:33 | |
| *** sams-gleb has joined #openstack-swift | 06:36 | |
| *** takashi has joined #openstack-swift | 06:48 | |
| *** ukaynar has quit IRC | 06:52 | |
| *** ukaynar has joined #openstack-swift | 06:52 | |
| *** ukaynar has quit IRC | 06:53 | |
| *** silor has joined #openstack-swift | 07:10 | |
| *** rcernin has joined #openstack-swift | 07:12 | |
| *** silor1 has joined #openstack-swift | 07:23 | |
| *** silor has quit IRC | 07:24 | |
| *** silor1 is now known as silor | 07:24 | |
| *** takashi has quit IRC | 07:26 | |
| *** ChubYann has quit IRC | 07:36 | |
| *** tesseract has joined #openstack-swift | 07:54 | |
| *** takashi has joined #openstack-swift | 07:59 | |
| *** oshritf has joined #openstack-swift | 07:59 | |
| *** rledisez has joined #openstack-swift | 08:09 | |
| *** newmember has quit IRC | 08:27 | |
| *** geaaru has joined #openstack-swift | 08:40 | |
| *** foexle has joined #openstack-swift | 08:47 | |
| *** oshritf has quit IRC | 09:03 | |
| *** kei_yama has quit IRC | 09:16 | |
| *** jordanP has joined #openstack-swift | 09:21 | |
| *** vint_bra has joined #openstack-swift | 09:24 | |
| *** mvk has quit IRC | 09:24 | |
| *** vint_bra has quit IRC | 09:26 | |
| *** cbartz has joined #openstack-swift | 09:38 | |
| *** acoles_ is now known as acoles | 09:50 | |
| *** takashi has quit IRC | 09:53 | |
| *** mvk has joined #openstack-swift | 09:55 | |
| *** JimCheung has joined #openstack-swift | 09:58 | |
| *** JimCheung has quit IRC | 10:02 | |
| *** oshritf has joined #openstack-swift | 10:12 | |
| *** oshritf has quit IRC | 10:13 | |
| *** sams-gle_ has joined #openstack-swift | 10:26 | |
| *** sams-gleb has quit IRC | 10:26 | |
| *** sams-gle_ has quit IRC | 10:52 | |
| *** sams-gleb has joined #openstack-swift | 10:52 | |
| *** sams-gleb has quit IRC | 10:57 | |
| *** dmellado has quit IRC | 11:00 | |
| *** dmellado has joined #openstack-swift | 11:03 | |
| *** mvk has quit IRC | 11:10 | |
| *** mvk has joined #openstack-swift | 11:23 | |
| *** newmember has joined #openstack-swift | 11:24 | |
| *** newmember has quit IRC | 11:30 | |
| *** sams-gleb has joined #openstack-swift | 11:34 | |
| openstackgerrit | Merged openstack/swift: Remove some cruft in test_proxy_logging https://review.openstack.org/426978 | 11:39 |
| *** ganders has joined #openstack-swift | 11:57 | |
| *** bkopilov has quit IRC | 12:11 | |
| *** NM has joined #openstack-swift | 12:12 | |
| *** catintheroof has joined #openstack-swift | 12:22 | |
| *** SkyRocknRoll has quit IRC | 12:46 | |
| *** Jeffrey4l__ is now known as Jeffrey4l | 12:47 | |
| *** oshritf has joined #openstack-swift | 12:58 | |
| *** winggundamth_ has quit IRC | 13:07 | |
| openstackgerrit | Merged openstack/swift: Clean up EC overview docs a bit https://review.openstack.org/426971 | 13:08 |
| openstackgerrit | Merged openstack/swift: Do not revert fragments to handoffs https://review.openstack.org/425441 | 13:08 |
| *** dmorita has joined #openstack-swift | 13:29 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Fix misleading hash invalidations test comments https://review.openstack.org/427205 | 13:33 |
| *** dmorita has quit IRC | 13:33 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 13:35 |
| *** catinthe_ has joined #openstack-swift | 13:36 | |
| *** catintheroof has quit IRC | 13:40 | |
| *** sgundur has quit IRC | 13:48 | |
| *** vint_bra has joined #openstack-swift | 13:49 | |
| *** sgundur has joined #openstack-swift | 13:49 | |
| openstackgerrit | Merged openstack/swift: Warn about using EC with isa_l_rs_vand and nparity >= 5 https://review.openstack.org/425496 | 14:05 |
| *** oshritf has quit IRC | 14:16 | |
| *** oshritf has joined #openstack-swift | 14:22 | |
| *** JimCheung has joined #openstack-swift | 14:25 | |
| *** JimCheung has quit IRC | 14:29 | |
| *** oshritf has quit IRC | 14:34 | |
| *** bkopilov has joined #openstack-swift | 15:03 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Fix race when consolidating new partition https://review.openstack.org/426336 | 15:07 |
| *** sams-gleb has quit IRC | 15:12 | |
| *** sams-gleb has joined #openstack-swift | 15:13 | |
| *** oshritf has joined #openstack-swift | 15:14 | |
| *** cdelatte has joined #openstack-swift | 15:14 | |
| *** sams-gleb has quit IRC | 15:17 | |
| *** psachin has quit IRC | 15:27 | |
| *** gatuus has joined #openstack-swift | 15:27 | |
| *** caiobrentano has quit IRC | 15:28 | |
| *** sams-gleb has joined #openstack-swift | 15:28 | |
| *** caiobrentano has joined #openstack-swift | 15:30 | |
| notmyname | good morning | 15:32 |
| *** ppai has quit IRC | 15:38 | |
| notmyname | nice! two of the priority patches have landed https://wiki.openstack.org/wiki/Swift/PriorityReviews | 15:38 |
| notmyname | acoles: you were called out by zaitcev on https://review.openstack.org/#/c/337297/ | 15:38 |
| patchbot | patch 337297 - swift - Add support to increase object ring partition power | 15:38 |
| zaitcev | yes | 15:39 |
| zaitcev | Well... | 15:39 |
| notmyname | it's got 2 +2s (from mattoliverau and zaitcev) | 15:39 |
| notmyname | /cc cschwede | 15:39 |
| zaitcev | I would mind Clay or Sam either | 15:39 |
| *** winggundamth_ has joined #openstack-swift | 15:39 | |
| zaitcev | Givin my history of poor quality reviews... | 15:39 |
| zaitcev | er. would NOT | 15:39 |
| zaitcev | case in point | 15:40 |
| cschwede | push the button, push the button, push the button... ;) | 15:40 |
| notmyname | I'll talk to torgomatic today, see what he's up to. I know he's been working on something else internally | 15:44 |
| *** glange_ has left #openstack-swift | 15:45 | |
| *** catinthe_ has quit IRC | 15:45 | |
| notmyname | looks like the global EC patch needs rebasing | 15:48 |
| *** _JZ_ has joined #openstack-swift | 15:57 | |
| *** foexle has quit IRC | 16:00 | |
| *** rcernin has quit IRC | 16:04 | |
| notmyname | dims: I'm trying to find out *why* a particular version of a package is used in upper-constraints. I found the relevant patch, but it's from a bot and "updated from generate-constraints" | 16:16 |
| notmyname | dims: how that work? anything further I can look for? | 16:16 |
| notmyname | (I'm asking you since you're the one who +2/+A the patch) | 16:16 |
| notmyname | https://review.openstack.org/#/c/309273/ | 16:16 |
| patchbot | patch 309273 - requirements - Updated from generate-constraints (MERGED) | 16:16 |
| *** d0ugal has quit IRC | 16:16 | |
| *** d0ugal has joined #openstack-swift | 16:17 | |
| *** d0ugal has quit IRC | 16:17 | |
| *** d0ugal has joined #openstack-swift | 16:17 | |
| *** oshritf has quit IRC | 16:19 | |
| notmyname | dims: what I'm looking for is any reason behind why the versions in that patch are the versions that are there. taking the first one, why did django move from 1.8.12 to 1.8.13? is there anything that references that | 16:19 |
| notmyname | I didn't find a `generate-constraints` script in the openstack requirements repo | 16:20 |
| notmyname | dims: and my second question is this: if I wanted to test a swift in the gate with a version of a dependency that's newer than what's in upper-constraints, is that possible to do? | 16:21 |
| notmyname | (specifically the dependency I'm curious about and what started me on this path is eventlet. I'm interested in the 0.20.1 release) | 16:22 |
| *** newmember has joined #openstack-swift | 16:26 | |
| *** oshritf has joined #openstack-swift | 16:26 | |
| *** winggundamth_ has quit IRC | 16:34 | |
| *** chsc has joined #openstack-swift | 16:38 | |
| *** oshritf has quit IRC | 16:50 | |
| *** oshritf has joined #openstack-swift | 16:50 | |
| *** oshritf has quit IRC | 16:51 | |
| *** tesseract has quit IRC | 16:56 | |
| dims | notmyname : where do i start :) the bot runs once a day and looks in pypi if there are newer versions of stuff that are ok with g-r specified values. then proposes a review with all those changes | 17:08 |
| *** cbartz has quit IRC | 17:08 | |
| notmyname | ok | 17:08 |
| dims | then we try to run a bunch of tests to see if we would break anything and allow those changes to merge | 17:09 |
| dims | for any bot proposed update, one core is enough, we have noted that in the readme | 17:09 |
| dims | for eventlet, for 0.20.1 we ran into issues - https://review.openstack.org/#/c/417590/ | 17:09 |
| patchbot | patch 417590 - requirements - Update eventlet uc to 0.20.1 | 17:09 |
| dims | if you see that review dirk has mentioned some issues while testing | 17:10 |
| dims | you can also see gate-cross-* jobs which test that review against master of some of the projects | 17:10 |
| notmyname | interesting. ok | 17:11 |
| dims | so you could add a gate-cross job for swift | 17:11 |
| dims | so we would know even before things merge into u-c | 17:11 |
| notmyname | actually I was just told about an issue with recent eventlet and swift, so there's something with that too | 17:13 |
| dims | you can get more input from those the newer folks like dirk, prometheanfire, tonyb who are actively doing this kind of stuff nowadays | 17:13 |
| dims | right, not surprised | 17:13 |
| notmyname | however, recent eventlet also includes a patch that's required for us to enable pypy gate testing | 17:13 |
| notmyname | so obviously that can't happen yet for the other reasons, but that's why I was interested | 17:13 |
| dims | y oslo.messaging needs it too for py35 support | 17:14 |
| notmyname | so help me with this... | 17:14 |
| dims | i see 3 attempts so far for that bump https://review.openstack.org/#/q/eventlet+project:%255E.*requirements.*+branch:master | 17:14 |
| * dims listening | 17:15 | |
| notmyname | suppose I wanted to patch swift to work around changes in recent eventlet. how to I get that tested against recent eventlet in the gate? | 17:15 |
| notmyname | if the upper constraints is lower, then it won't hapen, right? | 17:16 |
| notmyname | substitute "eventlet" for any dependency, but eventlet seems to be appropriate today | 17:16 |
| dims | typically i file or look for a review in requirements repo with that bump, then file a review in say glance or swift with Depends-On | 17:16 |
| dims | yep | 17:16 |
| notmyname | so in the likely case, you'd submit with a depends-on, which will apply the requirements bump to the test runner node, ensure that tests pass, then likely update the patch to *remove* the depends on, just to make sure we aren't breaking older stuff or necessarily requiring the min version bump, then when both rounds have passed, land the patch | 17:18 |
| notmyname | or, test locally, comment that it "works on my machine with newer dependency", land it, and hope for the best when/if the gate finally allows the newer version | 17:19 |
| dims | y there are many variations. that is one | 17:19 |
| dims | notmyname : typically i try to tell people that things should work with both old and newer version...unless they want to block the old version with a range in g-r | 17:20 |
| *** garyj has joined #openstack-swift | 17:20 | |
| notmyname | right, of course. it's a question of how many contortions the author and reviewers are willing to go through in order to get the openstack robots to say +1 | 17:21 |
| notmyname | I mean, theoretically, I can just delete the tests and therefore remove any blockers to landing code. but that's probably a terrible idea ;-) | 17:22 |
| dims | notmyname : i prefer to file my actual changes in one review and then throw a WIP that depends on my actual changes and the upper-constraints change. that way the WIP is just throw away | 17:23 |
| dims | and used only for testing | 17:23 |
| dims | we have to work across a whole lot of projects, examples of things stuck this week were oslo.context and stevedore where we did not break anyone, but we had to make changes in glance/neutron etc to accomodate the new version and then land the u-c change | 17:24 |
| dims | requirements team is always looking to make things easier so any concrete suggestions would be very welcome. more ideas to do this better is awesome | 17:25 |
| dims | all i ask is spend some time there to see what problems they face when you try to go beyond a single project | 17:26 |
| dims | i really don't think of it as making bots happy. its trying to make sure we don't break anyone as far as we can tell (as obviously you can't add CI jobs for every combination) | 17:28 |
| *** JimCheung has joined #openstack-swift | 17:28 | |
| notmyname | oh yeah, definitely. I wasn't trying to say anything opposed to that :-) | 17:29 |
| notmyname | dims: thanks for the info and insight this morning | 17:29 |
| dims | :) | 17:29 |
| *** dmorita has joined #openstack-swift | 17:33 | |
| *** arch-nemesis has joined #openstack-swift | 17:34 | |
| *** dmorita has quit IRC | 17:38 | |
| *** caiobrentano has quit IRC | 17:41 | |
| *** caiobrentano has joined #openstack-swift | 17:42 | |
| *** dmorita has joined #openstack-swift | 17:42 | |
| *** rledisez has quit IRC | 17:48 | |
| *** ppai has joined #openstack-swift | 17:52 | |
| *** gatuus has quit IRC | 17:53 | |
| *** mvk has quit IRC | 17:54 | |
| *** catintheroof has joined #openstack-swift | 17:55 | |
| *** jamielennox is now known as jamielennox|away | 17:58 | |
| *** geaaru has quit IRC | 17:59 | |
| *** dmorita_ has joined #openstack-swift | 18:00 | |
| *** dmorita has quit IRC | 18:00 | |
| timburke | notmyname: on the part-power increase, i started looking at it late yesterday. i'll try to get my thoughts together and post them today | 18:03 |
| timburke | so far, i think most of it's just going to be messaging | 18:04 |
| notmyname | timburke: great, thanks | 18:08 |
| *** manous has joined #openstack-swift | 18:09 | |
| *** newmember has quit IRC | 18:10 | |
| *** ukaynar has joined #openstack-swift | 18:12 | |
| *** jordanP has quit IRC | 18:14 | |
| *** garyj has quit IRC | 18:23 | |
| *** ppai has quit IRC | 18:23 | |
| *** tqtran has joined #openstack-swift | 18:23 | |
| clayg | morning | 18:26 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 18:28 |
| *** mvk has joined #openstack-swift | 18:28 | |
| *** klrmn has joined #openstack-swift | 18:28 | |
| *** silor has quit IRC | 18:30 | |
| clayg | dims: notmyname: interesting convo on g-r - thanks for the information! | 18:31 |
| *** silor has joined #openstack-swift | 18:31 | |
| notmyname | backblaze 2016 hard drive report https://www.backblaze.com/blog/hard-drive-benchmark-stats-2016/ | 18:36 |
| *** ChubYann has joined #openstack-swift | 19:00 | |
| *** manous has quit IRC | 19:04 | |
| *** garyj has joined #openstack-swift | 19:04 | |
| *** foexle has joined #openstack-swift | 19:10 | |
| clayg | already!? feels like we just had one of those - about smart metrics... | 19:11 |
| *** manous has joined #openstack-swift | 19:17 | |
| clayg | kota_: oh no! you weren't able to rebase patch 219165 last night!? | 19:27 |
| patchbot | https://review.openstack.org/#/c/219165/ - swift - EC Fragment Duplication - Foundational Global EC C... | 19:27 |
| *** silor has quit IRC | 19:28 | |
| *** pcaruana has quit IRC | 19:31 | |
| *** manous has quit IRC | 19:34 | |
| *** ganders has quit IRC | 19:51 | |
| *** foexle has quit IRC | 19:54 | |
| *** garyj has quit IRC | 20:14 | |
| *** acoles is now known as acoles_ | 20:23 | |
| *** esnyder has joined #openstack-swift | 20:31 | |
| esnyder | hi all, newbie here. i'm writing a new swift piece of swift middleware, and trying to figure out unit testing log messages. i was hoping to use the testfixtures LogCapture facility, but can't seem to get it to work. anyone have advice or tips? | 20:34 |
| *** NM has quit IRC | 20:40 | |
| *** newmember has joined #openstack-swift | 20:45 | |
| *** cdelatte has quit IRC | 20:51 | |
| *** newmember has quit IRC | 21:05 | |
| *** vinsh has quit IRC | 21:17 | |
| tonyb | notmyname: I'm hoping to use the PTG to gather a few interested parties to help verify the new eventlet release. | 21:23 |
| tonyb | notmyname: we've been stuck on an old version for a while as we hit a series of issues the last time we tried to update | 21:23 |
| tonyb | notmyname: so I'd say it'd be very disruptive to try now | 21:24 |
| clayg | esnyder: a *bunch* of tests use test.unit.debug_logger to get a logger object that has a method "get_lines_for_level" | 21:24 |
| clayg | maybe grep around some other middleware tests for one of those strings and try to crib it? | 21:24 |
| esnyder | @clayg, ah, thanks for the tip! | 21:25 |
| *** Jeffrey4l_ has joined #openstack-swift | 21:34 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| notmyname | tonyb: cool. yeah, I wasn't expecting to try anything right now. after the release is fine. | 21:35 |
| notmyname | tonyb: is there a list of associated bugs right now? or a tag in LP to use for stuff that breaks with new eventlet | 21:35 |
| tonyb | notmyname: nope. I'd have to go looking. | 21:36 |
| notmyname | k | 21:36 |
| tonyb | notmyname: and add a little science to my gut feel | 21:36 |
| tonyb | notmyname: "NOTE: New versions of eventlet should not be accepted lightly as they have earned a reputation of frequently breaking things." | 21:37 |
| tonyb | notmyname: from global-requirements.txt | 21:37 |
| notmyname | noted | 21:37 |
| tonyb | notmyname: gotta scoot | 21:38 |
| *** catintheroof has quit IRC | 21:38 | |
| *** catintheroof has joined #openstack-swift | 21:38 | |
| clayg | wow! anecdata has been a word since the 1980's! | 21:38 |
| *** catintheroof has quit IRC | 21:39 | |
| MooingLemur | clayg: that's the first time I've seen that word but I can guess what it means. :P | 21:43 |
| *** vinsh has joined #openstack-swift | 21:49 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Remove per-device reconstruction stats https://review.openstack.org/427476 | 21:53 |
| clayg | oh gross! | 21:56 |
| timburke | clayg: ? | 21:57 |
| clayg | so my hashes.pkl has the new keys in them from testing the new branch - trying to rollback to another branch freaks out the reconstrcutor | 21:57 |
| timburke | well that seems unfortunate :-/ | 21:57 |
| clayg | the whole reason we cover _get_hashes with the pop's is because both reconstructor/replicator iterate over the suffix, hash returned from that method | 21:57 |
| clayg | i wonder if it's only until the next conslidate call | 21:58 |
| clayg | whoever fixed swift-object-info to work on tombstone metadata is a %^&*ing SAINT | 22:03 |
| *** vinsh has quit IRC | 22:04 | |
| *** catintheroof has joined #openstack-swift | 22:05 | |
| *** catintheroof has quit IRC | 22:05 | |
| *** catintheroof has joined #openstack-swift | 22:06 | |
| clayg | so, not it doesn't really seem that anyone is going out of their way to kill the keys that aren't suffixes | 22:08 |
| clayg | no obvious backwards compatible way to add the keys I need really jumps out | 22:09 |
| *** catintheroof has quit IRC | 22:11 | |
| *** caiobrentano_ has joined #openstack-swift | 22:12 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Better optimistic lock in get_hashes https://review.openstack.org/419787 | 22:14 |
| *** caiobrentano has quit IRC | 22:15 | |
| *** jamielennox|away is now known as jamielennox | 22:15 | |
| *** jeblair has quit IRC | 22:16 | |
| *** caiobrentano_ has quit IRC | 22:17 | |
| clayg | seriously - data is so gross - I don't know how we manage to ever keep anything working! | 22:20 |
| *** vint_bra has quit IRC | 22:22 | |
| *** jeblair has joined #openstack-swift | 22:22 | |
| *** vinsh has joined #openstack-swift | 22:29 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Make the reconstructor handoffs_first work (and useful) https://review.openstack.org/425493 | 22:36 |
| *** sams-gleb has quit IRC | 22:41 | |
| *** sams-gleb has joined #openstack-swift | 22:41 | |
| *** sams-gleb has quit IRC | 22:45 | |
| *** caiobrentano has joined #openstack-swift | 23:07 | |
| *** caiobrentano has quit IRC | 23:22 | |
| *** chsc has quit IRC | 23:35 | |
| *** kei_yama has joined #openstack-swift | 23:35 | |
| clayg | notmyname: does https://wiki.openstack.org/wiki/Swift/PriorityReviews automatically do the strike-through when something lands (that would be surprising right?) | 23:37 |
| notmyname | nope. I added those this morning | 23:38 |
| notmyname | <strike> </strike> tags | 23:38 |
| clayg | notmyname: you're so on top of it! | 23:38 |
| notmyname | not if everyone keeps being so productive and reviewing/landing things! ;-) | 23:39 |
| notmyname | updated with another stike | 23:40 |
| notmyname | *strike | 23:40 |
| notmyname | .gitignore accepts regext right? | 23:42 |
| notmyname | regex | 23:42 |
| notmyname | so if I have a dir that has one good file in it and I want to ignore everything else that's there... | 23:42 |
| notmyname | [^(myfile.txt)] | 23:43 |
| notmyname | ? | 23:43 |
| notmyname | not sure | 23:43 |
| notmyname | anyone know? | 23:43 |
| notmyname | sorry, you were too slow. google was faster ;-) | 23:45 |
| notmyname | for the curious: "!/path/to/known/file\n/path/to/known/*\n" | 23:45 |
| notmyname | so you tell it to ignore everything in the folder. except that one file (via the !) | 23:46 |
| clayg | ah yeah - i've encountered that exclude syntax before | 23:46 |
| clayg | notmyname: I think there's another trick where you can put a .gitignore in that /path/to/known that sometimes makes it easier to read... - i forget how that works too | 23:47 |
| notmyname | oh, interesting | 23:47 |
| clayg | but then you have to remember that *other* command to make git tell you *which* ignore rule is keepig it from detecting a file | 23:48 |
| clayg | ... if you start using a bunch of them | 23:48 |
| clayg | it's a loose loose - just check everything in | 23:48 |
| notmyname | lol @ timburke (typical in swift-landia). "oh hey can we get one more review" "sure! <leaves review with 42 inline comments>" | 23:49 |
| notmyname | (laughing at timburke isn't typical. big reviews are typical) | 23:49 |
| *** tdasilva has quit IRC | 23:49 | |
| timburke | hey, if you didn't want comments, don't ask for comments :-) | 23:50 |
| timburke | but like half of those are docs | 23:51 |
| * notmyname is packing up to drive back to SF, so offline for a while | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!