| *** igordc has quit IRC | 00:51 | |
| *** sarob has joined #kata-dev | 03:56 | |
| *** sameo has joined #kata-dev | 04:43 | |
| kata-irc-bot2 | <eric.ernst> @bergwolf @zhangwei555 PTAL @ https://github.com/kata-containers/runtime/pull/1648 | 05:10 |
|---|---|---|
| kata-irc-bot2 | <eric.ernst> and https://github.com/kata-containers/packaging/pull/497 | 05:10 |
| kata-irc-bot2 | <eric.ernst> Want to get these in for the RC that *will* happen tomorrow! | 05:10 |
| *** sarob has quit IRC | 05:17 | |
| *** tmhoang has joined #kata-dev | 05:28 | |
| *** gnawux has joined #kata-dev | 06:12 | |
| *** sarob has joined #kata-dev | 06:13 | |
| *** sarob has quit IRC | 06:19 | |
| *** gnawux has quit IRC | 06:22 | |
| *** sarob has joined #kata-dev | 06:46 | |
| *** sameo has quit IRC | 06:49 | |
| *** sgarzare has joined #kata-dev | 07:08 | |
| *** sarob has quit IRC | 07:19 | |
| *** gwhaley has joined #kata-dev | 07:23 | |
| *** jodh has joined #kata-dev | 07:25 | |
| *** sgarzare has quit IRC | 07:42 | |
| *** sameo has joined #kata-dev | 07:42 | |
| *** sgarzare has joined #kata-dev | 07:44 | |
| *** sarob has joined #kata-dev | 07:48 | |
| *** davidgiluk has joined #kata-dev | 08:04 | |
| brtknr | yankcrime: sorry it wasnt synced properly it seems, looking okay now? | 08:21 |
| *** sarob has quit IRC | 08:22 | |
| yankcrime | brtknr: yeah looks good | 08:34 |
| gwhaley | brtknr: is updated to Intel here now ;-) | 08:34 |
| brtknr | it was rsync playing up... | 08:35 |
| yankcrime | rsync never plays up | 08:36 |
| yankcrime | suspect pebcak | 08:36 |
| brtknr | yankcrime: thats a true believer | 08:42 |
| yankcrime | :D | 08:43 |
| brtknr | stefanha: I am updating the blog now with the fio_jobfile used | 09:17 |
| brtknr | stefanha: Please refresh | 09:18 |
| stefanha | brtknr: thanks! | 09:32 |
| *** lpetrut has joined #kata-dev | 10:35 | |
| *** tmhoang has quit IRC | 10:48 | |
| *** tmhoang has joined #kata-dev | 10:49 | |
| *** tmhoang has quit IRC | 10:56 | |
| *** tmhoang has joined #kata-dev | 10:56 | |
| *** gwhaley has quit IRC | 11:00 | |
| *** gwhaley has joined #kata-dev | 12:06 | |
| *** eernst has joined #kata-dev | 14:20 | |
| *** lpetrut has quit IRC | 14:43 | |
| *** lpetrut has joined #kata-dev | 14:47 | |
| *** kgz has quit IRC | 14:50 | |
| *** kragniz has joined #kata-dev | 14:51 | |
| *** kragniz is now known as kgz | 14:51 | |
| *** sameo has quit IRC | 14:51 | |
| *** lpetrut has quit IRC | 14:57 | |
| kata-irc-bot2 | <james.o.hunt> whoa...! We're down to 48 PRs non-do-not-merge PRs in the backlog! Can we get it down to 40 by COB folks? :-) | 15:01 |
| kata-irc-bot2 | <eric.ernst> how do you query this, @james.o.hunt | 15:02 |
| gwhaley | hah, I think we just got a bunch of new release ones jodh... prolly bumped it again! | 15:02 |
| gwhaley | the primary link to the search is on the herders page | 15:02 |
| kata-irc-bot2 | <eric.ernst> I just merged most of those, so we may be ok Graham | 15:02 |
| kata-irc-bot2 | <eric.ernst> oh, shame. | 15:03 |
| gwhaley | then you tweak the search a little to only look at the DNM and do the head math - is what I do.. | 15:03 |
| gwhaley | so, here is the 'find all DNM pr's' link... https://github.com/search?q=org%3Akata-containers+is%3Apr+is%3Aopen+%21label%3Ado-not-merge | 15:04 |
| gwhaley | I wonder if you can negate that label bit... | 15:04 |
| gwhaley | (I know github doesn't support multi label searches there - grrrr) | 15:04 |
| gwhaley | oops, drop the not... https://github.com/search?q=org%3Akata-containers+is%3Apr+is%3Aopen+label%3Ado-not-merge | 15:05 |
| kata-irc-bot2 | <james.o.hunt> @graham.whaley - you can negate a label by adding a leading dash: https://github.com/pulls?utf8=%E2%9C%93&q=is%3Aopen+is%3Apr+org%3Akata-containers+-label%3Ado-not-merge | 15:05 |
| kata-irc-bot2 | <james.o.hunt> 47 y'all! | 15:06 |
| kata-irc-bot2 | <james.o.hunt> 45 if we exclude wip too: https://github.com/pulls?utf8=%E2%9C%93&q=is%3Aopen+is%3Apr+org%3Akata-containers+-label%3Ado-not-merge+-label%3Awip | 15:07 |
| kata-irc-bot2 | <james.o.hunt> 44! | 15:08 |
| gwhaley | yep, got it - the magic is the '-' in front of the search item. thx | 15:08 |
| kata-irc-bot2 | <james.o.hunt> low-hanging fruit @jose.carlos.venegas.m / @eric.ernst - https://github.com/kata-containers/osbuilder/pull/276 | 15:09 |
| kata-irc-bot2 | <james.o.hunt> 42 after adding missing wip labels to two wip PRs! | 15:11 |
| kata-irc-bot2 | <james.o.hunt> down to 41 as another PR was missing a label | 15:12 |
| stefanha | eric.ernst: Daniel Berrange <berrange@redhat.com> ("danpb" on IRC) from libvirt is happy to review code and answer questions about libvirt and the Go bindings. API: https://godoc.org/github.com/libvirt/libvirt-go, XML API: https://godoc.org/github.com/libvirt/libvirt-go-xml | 15:13 |
| kata-irc-bot2 | <eric.ernst> Thanks Stefan! | 15:15 |
| kata-irc-bot2 | <james.o.hunt> Down to 40!: https://github.com/pulls?page=1&q=is%3Aopen+is%3Apr+org%3Akata-containers+-label%3Ado-not-merge+-label%3Awip&utf8=%E2%9C%93 | 15:19 |
| kata-irc-bot2 | <eric.ernst> 39! | 15:19 |
| kata-irc-bot2 | <james.o.hunt> That isn't 40 factorial btw... ;) | 15:19 |
| kata-irc-bot2 | <james.o.hunt> \o/ | 15:19 |
| kata-irc-bot2 | <james.o.hunt> :fireworks: :fire: :fire: :fire: | 15:19 |
| kata-irc-bot2 | <xu> :,) | 15:20 |
| kata-irc-bot2 | <eric.ernst> now if we could just merge https://github.com/kata-containers/osbuilder/pull/200 | 15:20 |
| kata-irc-bot2 | <eric.ernst> :) | 15:20 |
| kata-irc-bot2 | Action: james.o.hunt whistles... and then goes to look at that ol' :chestnut: ... | 15:21 |
| kata-irc-bot2 | <james.o.hunt> 38! | 15:22 |
| kata-irc-bot2 | Action: graham.whaley wonders if james.o.hunt is just going around slapping DNM on everything :-) | 15:23 |
| kata-irc-bot2 | <eric.ernst> hehe | 15:23 |
| kata-irc-bot2 | <graham.whaley> so... fine - when we write our weekly reports, we can report both numbers - but, let's not completely ignore the big pile of DNMs eh :slightly_smiling_face: | 15:23 |
| kata-irc-bot2 | <eric.ernst> @archana.m.shinde around? | 15:23 |
| kata-irc-bot2 | <eric.ernst> I think we need someone to take over https://github.com/kata-containers/runtime/pull/1233, https://github.com/kata-containers/runtime/pull/1208 | 15:24 |
| kata-irc-bot2 | <eric.ernst> stale-bot ftw? | 15:24 |
| kata-irc-bot2 | <eric.ernst> GH is supposed to announce new feature today.. .maybe GH actions gA? | 15:25 |
| kata-irc-bot2 | <james.o.hunt> @graham.whaley now there's an idea. That, or `hub-merge --all --jfdi` ;) | 15:26 |
| kata-irc-bot2 | <graham.whaley> jodh, if you happen to add a link to show all non-dnm non-wip PRs to the herders wiki page... ;) Otherwise, I'll probably do it on Monday when I do the next herding email | 15:27 |
| kata-irc-bot2 | <graham.whaley> what we need is a banner/ticker that shows the counts....hehe | 15:28 |
| kata-irc-bot2 | <james.o.hunt> @graham.whaley - I've tweaked https://github.com/kata-containers/community/wiki/Review-Team-Rota | 15:40 |
| kata-irc-bot2 | <gmmaharaj> stefanha: have you ever tested virtio-fs with vm templating &/or initrd images? | 15:41 |
| kata-irc-bot2 | <gmmaharaj> also i am thinking we will reduce the default cache size to 1024/2048 given that all our Ci will happen on VMs which are barely either 4G or 8G RAM size. and if i am right today, if the total ram is smaller than the cache size, virtio-fs fails to start. | 15:42 |
| davidgiluk | hmm why should it fail to start | 15:45 |
| davidgiluk | I'd have to check but I think I mmap a /dev/zero and give it no perms, but I'd hoped that wouldn't consume anything useful | 15:45 |
| kata-irc-bot2 | Action: gmmaharaj goes to get the exact error message | 15:46 |
| gmmaha | davidgiluk: you will hit this error if the cache size is bigger than the total ram in the machine. | 15:48 |
| gmmaha | docker: Error response from daemon: OCI runtime create failed: qemu-system-x86_64: /home/ganeshma/qemu/memory.c:2155: memory_region_get_ram_ptr: Assertion `mr->ram_block' failed.: unknown. | 15:48 |
| davidgiluk | gmmaha: My challenge is that I need to reserve virtual address space for the cache; it should never really get any memory | 15:49 |
| davidgiluk | but I can see depending on the ulimit perhaps ulimit is doing that | 15:49 |
| kata-irc-bot2 | <eric.ernst> @james.o.hunt question on yq.... what's the rationale for checking for yq only in the GOPATH/bin directory? | 15:51 |
| kata-irc-bot2 | <graham.whaley> iirc, there are two `yq`s in the wild, and our scripts always pull the one we want vi go/github. but, I could be confusing with jq... | 15:54 |
| gmmaha | davidgiluk: these are the ulimit settings. | 15:54 |
| gmmaha | core file size (blocks, -c) 0 | 15:54 |
| gmmaha | data seg size (kbytes, -d) unlimited | 15:54 |
| gmmaha | scheduling priority (-e) 0 | 15:54 |
| gmmaha | file size (blocks, -f) unlimited | 15:54 |
| gmmaha | pending signals (-i) 19708 | 15:54 |
| gmmaha | max locked memory (kbytes, -l) 16384 | 15:54 |
| gmmaha | max memory size (kbytes, -m) unlimited | 15:54 |
| gmmaha | open files (-n) 1024 | 15:54 |
| gmmaha | pipe size (512 bytes, -p) 8 | 15:54 |
| gmmaha | POSIX message queues (bytes, -q) 819200 | 15:54 |
| gmmaha | real-time priority (-r) 0 | 15:54 |
| gmmaha | stack size (kbytes, -s) 8192 | 15:54 |
| gmmaha | cpu time (seconds, -t) unlimited | 15:54 |
| gmmaha | max user processes (-u) 19708 | 15:54 |
| gmmaha | virtual memory (kbytes, -v) unlimited | 15:54 |
| gmmaha | file locks (-x) unlimited | 15:54 |
| davidgiluk | gmmaha: So this is just running it inside a small VM with no overcommit/swap? | 15:55 |
| gmmaha | davidgiluk: yup. | 15:56 |
| davidgiluk | gmmaha: OK, can you file a github issue against our qemu/github tree and that will keep me honest so I can't forget about it | 15:56 |
| gmmaha | davidgiluk: :) you got it. will do! | 15:57 |
| davidgiluk | gmmaha: You should be OK running with say 256MB or 512MB cache if you're a bit tight for space | 15:57 |
| gmmaha | davidgiluk: right. anything smaller than the total ram and i am fine. given that we know kata's CI runs on VMs that are not really big, i think dropping the default to 1024 or 2048 in kata's config might be good for the mostpart | 15:58 |
| davidgiluk | gmmaha: It should at least fail more cleanly than that | 16:02 |
| kata-irc-bot2 | <james.o.hunt> @eric.ernst - yep, @graham.whaley's correct - we're trying to avoid using the python version (https://github.com/kislyuk/yq) by mistake. | 16:03 |
| *** lpetrut has joined #kata-dev | 16:06 | |
| kata-irc-bot2 | <eric.ernst> sigh. OK - cool. | 16:06 |
| *** iamweswilson has joined #kata-dev | 16:22 | |
| *** sameo has joined #kata-dev | 16:23 | |
| *** sgarzare has quit IRC | 16:56 | |
| *** gwhaley has quit IRC | 17:05 | |
| *** jodh has quit IRC | 17:06 | |
| *** sarob has joined #kata-dev | 17:10 | |
| *** lpetrut has quit IRC | 17:17 | |
| *** tmhoang has quit IRC | 17:28 | |
| kata-irc-bot2 | <manohar.r.castelino> davidgiluk we have a second issue with virtio-fs... when running it with kubernetes we hotplug memory as we start with a small VM. So the hotplugged DIMMs are not visible to the the daemon... so we will need to support hotplug for us leverage virtio-fs in k8s | 17:42 |
| davidgiluk | manohar: Interesting, I'd not tried hotplugging, the theory is that we should get a new 'set_mem_table' message over the vhost user to tell us about the new mapping; how upset the daemon gets about it's mapping changing is an interesting question | 17:44 |
| *** psyton has joined #kata-dev | 17:51 | |
| *** tmhoang has joined #kata-dev | 17:53 | |
| kata-irc-bot2 | <eric.ernst> davidgiluk one thing I notice in my testing is that if I use vsock, things fail. | 18:14 |
| kata-irc-bot2 | <eric.ernst> "things fail" -- virtiofsd is defunct and timeout/failure for creating a container. | 18:14 |
| kata-irc-bot2 | <eric.ernst> if I disable vsock, it starts up fine. | 18:14 |
| *** sarob has quit IRC | 18:25 | |
| davidgiluk | manohar: I had an idea; can you give me details about how you're hot plugging the dimm? It has to be shared in the right way for the daemon to see it | 18:28 |
| davidgiluk | eernst: That's fun, the two *should* be entirely separate | 18:28 |
| davidgiluk | eernst: Again, can you file a github bug against one of our components please | 18:29 |
| *** sarob has joined #kata-dev | 18:56 | |
| *** dklyle_ has joined #kata-dev | 19:02 | |
| *** david-lyle has quit IRC | 19:04 | |
| kata-irc-bot2 | <eric.ernst> @gmmaharaj PTAL: https://github.com/kata-containers/runtime/pull/1663 | 19:22 |
| *** davidgiluk has quit IRC | 19:31 | |
| *** sarob has quit IRC | 19:43 | |
| kata-irc-bot2 | <gmmaharaj> @eric.ernst quick small feedback on the change. otherwise it's good. | 19:44 |
| kata-irc-bot2 | <gmmaharaj> quick check and i can approve it. | 19:44 |
| kata-irc-bot2 | <eric.ernst> that's just matching how it is done in qemu's toml | 19:45 |
| kata-irc-bot2 | <gmmaharaj> alright. | 19:50 |
| kata-irc-bot2 | <gmmaharaj> that ought to work | 19:50 |
| kata-irc-bot2 | <gmmaharaj> ohh @eric.ernst you need to fix the commit msg. ``` Found 1 commit between commit c8c167caab5144409d75a6e24d45ff9dad40b2e1 and branch master ERROR: No "Fixes" found + rc=1 + '[' 1 -ne 0 ']' + cat ERROR: checkcommits failed. See the document below for help on formatting commits for the project. https://github.com/kata-containers/community/blob/master/CONTRIBUTING.md#patch-format ``` | 19:50 |
| kata-irc-bot2 | <eric.ernst> repushed | 19:51 |
| *** iamweswilson has left #kata-dev | 20:25 | |
| *** iamweswilson has joined #kata-dev | 20:25 | |
| *** sameo has quit IRC | 20:41 | |
| *** tmhoang has quit IRC | 20:48 | |
| *** gmmaharaj has joined #kata-dev | 21:04 | |
| *** gmmaha has quit IRC | 21:05 | |
| *** gmmaharaj is now known as gmmaha | 21:05 | |
| *** gmmaha has quit IRC | 21:34 | |
| *** gmmaha has joined #kata-dev | 21:36 | |
| *** sarob has joined #kata-dev | 21:45 | |
| *** sarob has quit IRC | 22:31 | |
| *** eernst has quit IRC | 22:32 | |
| *** eernst has joined #kata-dev | 22:36 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!