| *** wuchunyang has joined #zuul | 00:26 | |
| *** y2kenny has quit IRC | 00:29 | |
| *** wuchunyang has quit IRC | 00:30 | |
| *** y2kenny has joined #zuul | 00:46 | |
| *** wuchunyang has joined #zuul | 00:58 | |
| *** openstackgerrit has joined #zuul | 01:24 | |
| openstackgerrit | Merged zuul/nodepool master: nodepool-functional-k8s: don't run on Xenial https://review.opendev.org/c/zuul/nodepool/+/763921 | 01:24 |
|---|---|---|
| *** holser has quit IRC | 01:40 | |
| *** Goneri has quit IRC | 02:03 | |
| mnaser | pabelanger: happy to see oyu're progressing there, look forward to seeing you pushing updates | 02:41 |
| *** vishalmanchanda has joined #zuul | 02:44 | |
| *** hamalq has joined #zuul | 02:45 | |
| *** hamalq has quit IRC | 02:49 | |
| *** bhavikdbavishi has joined #zuul | 02:52 | |
| *** hamalq has joined #zuul | 03:00 | |
| *** bhavikdbavishi1 has joined #zuul | 03:03 | |
| *** bhavikdbavishi has quit IRC | 03:04 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:04 | |
| *** hamalq has quit IRC | 03:05 | |
| *** piotrowskim has quit IRC | 03:08 | |
| openstackgerrit | Merged zuul/nodepool master: Bump diskimage-builder requirement to 3.4.0 https://review.opendev.org/c/zuul/nodepool/+/763894 | 03:16 |
| *** bhavikdbavishi has quit IRC | 04:23 | |
| *** bhavikdbavishi has joined #zuul | 04:24 | |
| *** jfoufas1 has joined #zuul | 05:45 | |
| *** evrardjp has quit IRC | 05:47 | |
| *** evrardjp has joined #zuul | 05:48 | |
| *** zenkuro has joined #zuul | 06:00 | |
| *** bhavikdbavishi1 has joined #zuul | 06:04 | |
| *** bhavikdbavishi has quit IRC | 06:05 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 06:05 | |
| *** saneax has joined #zuul | 07:01 | |
| *** bhavikdbavishi has quit IRC | 07:08 | |
| *** zenkuro has quit IRC | 07:19 | |
| *** zenkuro has joined #zuul | 07:20 | |
| *** y2kenny has quit IRC | 07:24 | |
| *** mach1na has joined #zuul | 07:29 | |
| *** rpittau|afk is now known as rpittau | 07:37 | |
| *** bhavikdbavishi has joined #zuul | 07:39 | |
| *** iurygregory is now known as iurygregory|afk | 07:48 | |
| *** jcapitao has joined #zuul | 07:55 | |
| *** hashar has joined #zuul | 08:06 | |
| *** tosky has joined #zuul | 08:50 | |
| *** jpena|off is now known as jpena | 08:55 | |
| *** iurygregory|afk is now known as iurygregory | 09:10 | |
| *** sean-k-mooney1 has joined #zuul | 09:16 | |
| *** sean-k-mooney has quit IRC | 09:17 | |
| *** bhavikdbavishi has quit IRC | 10:04 | |
| openstackgerrit | Daniel Blixt proposed zuul/zuul-jobs master: Use script to populate test file tree fixtures https://review.opendev.org/c/zuul/zuul-jobs/+/764062 | 10:06 |
| *** wuchunyang has quit IRC | 10:09 | |
| openstackgerrit | Daniel Blixt proposed zuul/zuul-jobs master: Allow mirror push to delete current branch https://review.opendev.org/c/zuul/zuul-jobs/+/764152 | 10:27 |
| *** hashar has quit IRC | 10:50 | |
| *** bhavikdbavishi has joined #zuul | 11:05 | |
| *** anxolerd has joined #zuul | 11:20 | |
| *** bhavikdbavishi1 has joined #zuul | 11:22 | |
| anxolerd | Hello, I've heard a lot of cool things about zuul and now decided to evaluate, how it fits my needs. I've set up server for CI and configured and started zuul and nodepool services there but have no idea how to configure basic testing pipelines for the repo. I failed to find examples in documentation. Can anybody help me with that? | 11:22 |
| *** bhavikdbavishi has quit IRC | 11:23 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 11:23 | |
| avass | anxolerd: this might help you https://www.youtube.com/watch?v=vb0Iuf-6wHs :) | 11:25 |
| avass | there also examples in the repo but I'm not sure if it's part of the web page | 11:26 |
| avass | anxolerd: https://opendev.org/zuul/zuul/src/branch/master/doc/source/examples | 11:26 |
| *** jcapitao is now known as jcapitao_lunch | 11:35 | |
| *** rfolco|ruck has joined #zuul | 11:36 | |
| *** bhavikdbavishi has quit IRC | 11:50 | |
| *** bhavikdbavishi has joined #zuul | 11:51 | |
| *** anxolerd has quit IRC | 11:51 | |
| Phoenikzz | Hi, I've been working a little while with zuul (in same team as avass) and we recently encountered a problem where a checked out branch on a node had been deleted at origin. While testing the solution we also encountered that the test fixture file trees in zuul-jobs wont checkout properly on Windows. I've proposed fixes to both problems, very happy to get feedback, their referenced as Daniel Blixt's changes above | 11:54 |
| *** mach1na has quit IRC | 12:08 | |
| zbr | avass: can you please add a comment on https://review.opendev.org/c/zuul/zuul/+/763333 ? i am trying to figure out what else needs to be done to unblock it. | 12:16 |
| zbr | apparently there are lots of ideas for improvement when someone else has to do them. | 12:17 |
| zbr | i fix one request, two days later someone say it was better before... | 12:17 |
| *** jpena is now known as jpena|lunch | 12:35 | |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul master: Improve summary build layout https://review.opendev.org/c/zuul/zuul/+/763333 | 12:40 |
| *** jcapitao_lunch is now known as jcapitao | 12:44 | |
| avass | linking to a specific logline doesn't seem to work: https://zuul.opendev.org/t/zuul/build/23bac03b379d4617aa3c2574acc65a55/log/job-output.txt#134 | 12:47 |
| avass | zbr: in a meeting I'll check in a bit :) | 12:48 |
| *** rlandy has joined #zuul | 12:52 | |
| *** mach1na has joined #zuul | 12:52 | |
| *** vishalmanchanda has quit IRC | 13:26 | |
| *** bhavikdbavishi has quit IRC | 13:33 | |
| *** jpena|lunch is now known as jpena | 13:35 | |
| *** zenkuro has quit IRC | 13:37 | |
| *** zenkuro has joined #zuul | 13:38 | |
| *** hashar has joined #zuul | 13:38 | |
| *** sshnaidm is now known as sshnaidm|afk | 13:56 | |
| *** rpittau is now known as rpittau|bbl | 14:29 | |
| *** sshnaidm|afk is now known as sshnaidm | 14:34 | |
| *** bhavikdbavishi has joined #zuul | 14:41 | |
| *** bhavikdbavishi1 has joined #zuul | 14:44 | |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul-jobs master: Remove ansible-lint path exclusions https://review.opendev.org/c/zuul/zuul-jobs/+/731471 | 14:44 |
| *** bhavikdbavishi has quit IRC | 14:45 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 14:45 | |
| *** vishalmanchanda has joined #zuul | 14:53 | |
| *** Goneri has joined #zuul | 14:56 | |
| *** saneax has quit IRC | 15:03 | |
| *** hashar has quit IRC | 15:14 | |
| *** Phoenikzz has quit IRC | 15:14 | |
| *** ikhan has quit IRC | 15:39 | |
| *** rpittau|bbl is now known as rpittau | 15:44 | |
| zbr | does anyone need a proof regarding why not wrapping logged line is bad? here is the challange, find docker package version installed by tox on this job: https://review.rdoproject.org/zuul/build/f580fab3cc3444269556d9fdf2481e8e/console | 15:47 |
| *** sean-k-mooney1 is now known as sean-k-mooney | 15:47 | |
| zbr | not wrapping lines is a nonsense imho, the only way i was able to extract the info out was to copy the line in an external editor to read it | 15:48 |
| fungi | that looks to me like proof of why really long loglines are a problem | 16:06 |
| fungi | granted, `ctrl-f docker` in my browser took me right to "docker==4.4.0" | 16:07 |
| zbr | fungi: true, but find does not help much if you want to look at all installed packages, for me that line is one that often need to inspect. | 16:13 |
| zbr | not challenge level 2: try to do the same from your mobile, or tablet. ;) | 16:14 |
| zbr | the who concept of iframe and page in page was debunked as an anti-pattern many years ago, that one is almost the same. | 16:16 |
| zbr | imho, we should have soft-wrapping default and a min-width of 79 chars, to avoid some undesired behavior for extremely narrow cases. | 16:18 |
| zbr | fungi: btw, it would be nice if you can review the #763333 (summary), i implemented the requests received from ian and others. | 16:20 |
| zbr | these UI changes seem to be particularly frustrating to ship, every new reviewer comes with his own set of preferences | 16:21 |
| *** jfoufas1 has quit IRC | 16:25 | |
| fungi | i'm probably the last one you want weighing in on that sort of thing... i still miss the days when most interfaces were designed for monochrome text-only serial terminals, and avoid touching a web browser at all if i can help it | 16:33 |
| fungi | also i stick to keybindings for navigation and rarely touch a pointing device | 16:34 |
| avass | fungi: having reliable keybindings and not having to touch a mouse is great | 16:37 |
| avass | zbr: I think corvus is the one with the strongest opinion. I don't think he was against it as long as it's configurable | 16:38 |
| avass | zbr: also it seems more to me like people want to have a discussion and come to an agreement before merging something instead of just approving something | 16:39 |
| zbr | yep, but configurable comes with costs, development, testing and maintenance, also conflicting with other changes in the same are. | 16:39 |
| zbr | imho, it should not be configurable. | 16:40 |
| zbr | there are at least two other features that would explode the complexity if we want it configurable: ANSI and display of line numbers. | 16:41 |
| avass | I don't have a strong opinion either way | 16:41 |
| *** irclogbot_1 has quit IRC | 16:41 | |
| zbr | i will probably rebase the old change for making it configurable after the other summary changes go in. | 16:42 |
| zbr | i cannot drive too many changes in parallel | 16:42 |
| zbr | meanwhile other CI systems are busy implementing cool stuff like collapsable sections on console output | 16:43 |
| zbr | if anyone is curious about section, already supported by gh-actins, gitlab and travis, i discovered about them while reviewing https://github.com/ansible-community/molecule/pull/2967 | 16:46 |
| avass | fungi: you might want to take a look at browsh: https://www.brow.sh/ ;) | 16:46 |
| *** irclogbot_3 has joined #zuul | 16:47 | |
| zbr | avass: lol, i tried loading gerrit on it, did something funny, not useful. | 16:50 |
| avass | zuul barely works in it | 16:51 |
| zbr | if it am correct it tried to convert the browser into ANSI, not really a browser. It could also try to load youtube of 2fps. | 16:52 |
| avass | pretty much | 16:52 |
| zbr | ouch pypi degraded | 16:53 |
| avass | but it's neat, runs in the terminal and has a monochrome mode :) | 16:55 |
| zbr | fungi: would you mind if I send the deprecation message around replacing python2 with python3 as defaults on zuul-jobs? | 17:00 |
| zbr | changes like https://review.opendev.org/c/zuul/zuul-jobs/+/742799 are blocked due to lack of announce. | 17:00 |
| *** anxolerd has joined #zuul | 17:00 | |
| zbr | i guess a notification to zuul-discuss saying that during the following weeks we would change defaults to make use of python3 and this may impact some users that would need... | 17:01 |
| anxolerd | I'd like to contribute to nodepool project, but it's the first time I use opendev and gerrit. Can you please help me? I've added SSH key to my gerrit account, but when I do git-review I receive "access denied (pubkey)" error. | 17:03 |
| zbr | anxolerd: probably you missed to load your key (agent), as it should work right away. | 17:04 |
| avass | yep that should work if you've set up your ssh key correctly | 17:04 |
| *** mach1na has quit IRC | 17:05 | |
| anxolerd | I am trying ssh command with that key explicitly (ssh -i anxolerd@review.opendev.org -p ....), and receive the same error | 17:05 |
| avass | maybe try with 'ssh -v' to get more verbose logs, maybe it's rejecting the key because of bad permissions or similar | 17:07 |
| anxolerd | ```debug1: Offering public key: /home/oleksandr/.ssh/id_rsa RSA SHA256:p3v4FwWMRJBNoXrYNBQRP/XHCXaUvRX45mwpOJWgsjI explicit agentdebug1: send_pubkey_test: no mutual signature algorithm``` | 17:08 |
| avass | oh is this the problem with fedora 33 someone mention a couple of days back? | 17:09 |
| avass | anxolerd: could it be this? https://www.reddit.com/r/Fedora/comments/jh9iyi/f33_openssh_no_mutual_signature_algorithm/ | 17:10 |
| anxolerd | yes, exactly! | 17:10 |
| anxolerd | Updated to f33 few days ago | 17:11 |
| anxolerd | Thanks | 17:11 |
| avass | anxolerd: here's more info in the release notes: https://fedoraproject.org/wiki/Releases/33/ChangeSet#Strong_crypto_settings:_phase_2 | 17:11 |
| avass | :) | 17:11 |
| avass | and opendev is even mentioned here: https://bugzilla.redhat.com/show_bug.cgi?id=1884920#c2 :) | 17:12 |
| openstack | bugzilla.redhat.com bug 1884920 in openssh "Cannot ssh into CentOS 6 using ssh key authentication" [Low,Closed: notabug] - Assigned to jjelen | 17:12 |
| clarkb | however the upgrade we just did should have fixed that | 17:13 |
| pabelanger | anxolerd: that is cloudnull :) | 17:13 |
| clarkb | I dont have a fedora 33 install to confirm but I did test it on my local host enabling only the sha2 options and it worked | 17:13 |
| clarkb | possible that fedora33 is complaining about sha1 being offered at all? | 17:14 |
| clarkb | fungi: ^ you probably grok this better than I do | 17:14 |
| openstackgerrit | Oleksandr Kovalchuk proposed zuul/nodepool master: Fix bindep dependencies for ubuntu-focal https://review.opendev.org/c/zuul/nodepool/+/764220 | 17:17 |
| anxolerd | I believe, I should change my SSH key for the longer one) | 17:27 |
| cloudnull | What? | 17:39 |
| pabelanger | cloudnull: you hit f33 ssh key issue on review.opendev.org | 17:41 |
| pabelanger | or atleast bugzilla says you did | 17:41 |
| cloudnull | OIC. | 17:41 |
| cloudnull | Yup. That was a problem. Super easy to fix in local ssh config | 17:42 |
| cloudnull | kinda an annoyance. But meh. | 17:42 |
| clarkb | well it should be fixed now | 17:42 |
| clarkb | and if it isnt more ifno would be appreciated | 17:42 |
| anxolerd | what kind of information can I provide? | 17:44 |
| clarkb | I think we need a comparison of what things f33 allows vs what the server offers | 17:46 |
| clarkb | we upgraded the server whoch should have modernized the hashes to those f33 allows but maybe f33 is very particular | 17:46 |
| openstackgerrit | Merged zuul/zuul-jobs master: Refresh intermediate TLS certs for testing https://review.opendev.org/c/zuul/zuul-jobs/+/764023 | 17:52 |
| *** jpena is now known as jpena|off | 18:03 | |
| fungi | it might help to run `ssh-keyscan -p 29418 review.opendev.org` on f33 (ideally with no workarounds in place) and see what order the host keys are showing up in | 18:17 |
| fungi | for me (on debian/unstable) it's seeing the following: ssh-rsa, ecdsa-sha2-nistp256, ssh-ed25519 | 18:18 |
| fungi | since f33 dislikes ssh-rsa and gerrit offers no other stronger rsa options, you might need to create an ecdsa or ed25519 keypair and use that instead? | 18:20 |
| clarkb | hrm I swear I tested a couple weeks back against atest server allowing only rsa + sha2 and it worked | 18:21 |
| clarkb | maybe it doesnt ofder it by default but if you request it explicitly you get it | 18:21 |
| *** vishalmanchanda has quit IRC | 18:21 | |
| fungi | lemme check review-test for comparison | 18:22 |
| fungi | same set but different order: ecdsa-sha2-nistp256, ssh-rsa, ssh-ed25519 | 18:23 |
| *** sshnaidm has quit IRC | 18:23 | |
| clarkb | I think it may be configurable now in the server too | 18:23 |
| clarkb | something we can look at after turkeying | 18:23 |
| *** jcapitao has quit IRC | 18:23 | |
| fungi | oh, now review.o.o ir returning them in a different order too. so i guess the order it serves them in is nondeterministic. that's ovely | 18:23 |
| fungi | lovely | 18:23 |
| *** sshnaidm has joined #zuul | 18:24 | |
| fungi | yeah, testing repeatedly i get them back in a random order | 18:24 |
| clarkb | weird. iirc the protocol says they are supposed to usethe first one on the client list that is in the server list | 18:24 |
| clarkb | maybe keyscam is doing them in different orders? | 18:25 |
| *** hashar has joined #zuul | 18:25 | |
| clarkb | since ssh doesnt actually have a proper keyscan in the protocol you have to send as client a single option and iterate through those aiui | 18:25 |
| clarkb | (this is how the nodepool keyscan works now) | 18:25 |
| anxolerd | `ssh-keyscan -p 29418 review.opendev.org# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)[review.opendev.org]:29418 ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQCfsIj/jqpI+2CFdjCL6kOiqdORWvxQ2sQbCzSzzmLXic8yVhCCbwarkvEpfUOHG4eyB0vqVZfMffxf0Yy3qjURrsroBCiuJ8GdiAcGdfYwHNfBI0cR6kydBZL537YDasIk0Z3ILzhwf7474Lm | 18:26 |
| anxolerd | kVzS7V2tMTb4ZiBS/jUeiHsVp88FZhIBkyhlb/awAGcUxT5U4QBXCAmerYXeB47FPuz9JFOVyF08LzH9JRe9tfXtqaCNhlSdRe/2pPRvn2EIhn5uHWwATACG9MBdrK8xv8LqPOik2w1JkgLWyBj11vDd5I3IjrmREGw8dqImqp0r6MD8rxqADlc1elfDIXYsy+TVH# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)[review.opendev.org]:29418 ecdsa-sha2-nistp256 | 18:26 |
| anxolerd | AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBD70+y/Y7dJQBNboK6k+hq0DEGgFRzW0GRRVneYZHZ0fJMTeEDU04HziYky8fhV4eJWeqQmh3pbdPJHoqDWIGEA=# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)[review.opendev.org]:29418 ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIncSZ8k+c0YJTO6Wux/Kpu2jnktFkepoU3jBkbIzXve# | 18:26 |
| anxolerd | review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)` | 18:26 |
| anxolerd | `ssh-keyscan -p 29418 review.opendev.org# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)[review.opendev.org]:29418 ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQCfsIj/jqpI+2CFdjCL6kOiqdORWvxQ2sQbCzSzzmLXic8yVhCCbwarkvEpfUOHG4eyB0vqVZfMffxf0Yy3qjURrsroBCiuJ8GdiAcGdfYwHNfBI0cR6kydBZL537YDasIk0Z3ILzhwf7474Lm | 18:26 |
| anxolerd | kVzS7V2tMTb4ZiBS/jUeiHsVp88FZhIBkyhlb/awAGcUxT5U4QBXCAmerYXeB47FPuz9JFOVyF08LzH9JRe9tfXtqaCNhlSdRe/2pPRvn2EIhn5uHWwATACG9MBdrK8xv8LqPOik2w1JkgLWyBj11vDd5I3IjrmREGw8dqImqp0r6MD8rxqADlc1elfDIXYsy+TVH# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)[review.opendev.org]:29418 ecdsa-sha2-nistp256 | 18:26 |
| anxolerd | AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBD70+y/Y7dJQBNboK6k+hq0DEGgFRzW0GRRVneYZHZ0fJMTeEDU04HziYky8fhV4eJWeqQmh3pbdPJHoqDWIGEA=# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)[review.opendev.org]:29418 ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIncSZ8k+c0YJTO6Wux/Kpu2jnktFkepoU3jBkbIzXve# | 18:26 |
| anxolerd | review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)# review.opendev.org:29418 SSH-2.0-GerritCodeReview_3.2.5.1-60-gbb9ea229fb-dirty (APACHE-SSHD-2.4.0)` | 18:26 |
| anxolerd | ssh-rsaecdsa-sha2-nistp256ssh-ed25519 | 18:27 |
| anxolerd | So I receive ssh-rsa as well | 18:27 |
| *** sshnaidm has quit IRC | 18:28 | |
| fungi | clarkb: oh, i think you're right, ssh-keyscan is probably randomizing the order in which it asks for different key types | 18:28 |
| *** sshnaidm has joined #zuul | 18:28 | |
| *** hashar has quit IRC | 18:30 | |
| clarkb | fungi: sshd.kex fwiw but the docs say the sha256 should be in there too | 18:32 |
| *** rpittau is now known as rpittau|afk | 18:52 | |
| *** holser has joined #zuul | 18:57 | |
| *** avass has quit IRC | 18:59 | |
| pabelanger | is there an example of ZUUL_SIBLINGS and how it works for container image builds? | 18:59 |
| *** yourname_ has joined #zuul | 19:00 | |
| clarkb | pabelanger: nodepool and dib are set up that way in integration tedting for dob | 19:00 |
| *** hashar has joined #zuul | 19:03 | |
| pabelanger | k, thanks | 19:04 |
| *** holser has quit IRC | 19:05 | |
| *** holser has joined #zuul | 19:05 | |
| *** holser has quit IRC | 19:06 | |
| *** holser has joined #zuul | 19:07 | |
| *** holser has quit IRC | 19:08 | |
| *** holser has joined #zuul | 19:09 | |
| *** holser has quit IRC | 19:10 | |
| *** holser has joined #zuul | 19:11 | |
| *** yourname_ is now known as avass | 19:11 | |
| *** holser has quit IRC | 19:11 | |
| *** avass has quit IRC | 19:12 | |
| *** holser has joined #zuul | 19:13 | |
| *** anxolerd has quit IRC | 19:13 | |
| *** bhavikdbavishi has quit IRC | 19:16 | |
| *** yourname_ has joined #zuul | 19:18 | |
| *** yourname_ is now known as avass | 19:25 | |
| pabelanger | how often should zuul-registry prune be run? | 19:27 |
| pabelanger | every hour? | 19:27 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Fix typo with container_images siblings logic https://review.opendev.org/c/zuul/zuul-jobs/+/764230 | 19:31 |
| clarkb | I think we do it daily | 19:33 |
| clarkb | if you do it too aggressively then depends on wont work | 19:33 |
| pabelanger | k | 19:33 |
| *** wuchunyang has joined #zuul | 19:33 | |
| pabelanger | oh, I guess I need to redo http password in gerrit | 19:33 |
| pabelanger | depends-on stopped working in zuul.a.c | 19:34 |
| clarkb | you need to use bawic auth instead of digest. generating a new secret is a good idea though | 19:35 |
| *** wuchunyang has quit IRC | 19:38 | |
| pabelanger | heh, think I need a restart of the scheduler for that change | 19:39 |
| pabelanger | heh | 19:44 |
| pabelanger | rate-limited by dockerhub | 19:44 |
| pabelanger | clarkb: so for zuul_siblings, diskimage-builder is in requirements.txt, which we install into the container. Then, because .zuul-sibling folder is found, we install DIB from that, and uninstall previous version, does that sound right? | 19:56 |
| pabelanger | trying to set this up for ansible/ansible image | 19:56 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Switch to quay.io for container-build jobs https://review.opendev.org/c/zuul/zuul-jobs/+/764237 | 20:02 |
| *** armstrongs has joined #zuul | 20:59 | |
| clarkb | pabelanger: ya | 21:01 |
| clarkb | siblings always runs after the normal install | 21:01 |
| *** rfolco|ruck has quit IRC | 21:08 | |
| *** armstrongs has quit IRC | 21:08 | |
| *** hashar has quit IRC | 21:32 | |
| ianw | switching to quay.io seems ok; but i guess still leaves us with the same issue really | 21:45 |
| ianw | i've only seen rate limiting issues with limestone atm, due to i'm guessing the nat situation | 21:45 |
| fungi | currently switching to quay.io leaves you with nothing because they're down (thanks amazon?) | 21:49 |
| pabelanger | well, i think because of https://www.docker.com/increase-rate-limits you hit it more in docker | 21:50 |
| pabelanger | I don't believe quay.io has limits right now | 21:50 |
| fungi | (when they're up anyway!) | 21:50 |
| fungi | quay.io doesn't have much of anything right now, unless the outage is over | 21:51 |
| pabelanger | yah, not sure how good it is honestly | 21:51 |
| pabelanger | just starting into container world | 21:51 |
| pabelanger | just know podman > docker now it seems | 21:52 |
| pabelanger | on another note, I have speculative container working | 21:52 |
| pabelanger | and it is nice | 21:52 |
| fungi | i'm mostly going by reports of the tripleo team's current scramble to move from quay.io to the rdo registry, shortly after they moved to quay.io to avoid dockerhub rate limits | 21:53 |
| pabelanger | ah, that is surprising | 21:54 |
| pabelanger | given quay.io is redhat thing | 21:54 |
| fungi | i guess it doesn't matter much whose frontend it is, if the backend they've decided to outsource it to falls offline | 21:55 |
| *** zenkuro has quit IRC | 22:18 | |
| *** zenkuro has joined #zuul | 22:18 | |
| ianw | how about instead of pulling the entire image of an operating system, we split it up into smaller parts. call them, say, packages | 22:20 |
| ianw | and then we write a small tool to download those packages. they can even express dependencies between each other | 22:20 |
| ianw | that way, we could just mirror those "packages", and you'd only have to download exactly what you need? | 22:21 |
| clarkb | ianw: madness | 22:23 |
| fungi | this idea seems somehow familiar | 22:27 |
| tristanC | ianw: but how to you install incompatible packages on the same system? | 22:32 |
| fungi | use a distribution which allows them to be compatible | 22:36 |
| fungi | (alternative names and pathing, support for coinstalling different versions of the same software) | 22:37 |
| fungi | the software packaging aspects of modern "containers" aren't especially original, nor are the system isolation technologies on which they rely. it's really the combination of the two which has proven to be a recent innovation | 22:38 |
| tristanC | fungi: then how to install packages from different distribution or distribution version? | 22:38 |
| fungi | now you're just reiterating gödel's completeness theorem | 22:39 |
| fungi | to make a distribution of all distributions it must contain itself | 22:40 |
| tristanC | fungi: not necessarly, debootstrap or dnf --installroot may install foreign packages | 22:42 |
| fungi | but is that then a singular distribution, or multiple distributions? | 22:44 |
| fungi | taking debian as an example, it's not uncommon to package multiple versions of tools or libraries as long as the package names for them can be kept distinct. that seems like it would count as one distribution | 22:45 |
| fungi | various approaches are taken to make them coinstallable, which tends to depend a lot on the language and build tooling for the software in question | 22:47 |
| fungi | anaconda is probably a more extreme (though relatively python-specific) example of keeping multiple versions of applications and libraries installed in parallel | 22:48 |
| fungi | nixos is an even more extreme and language-agnostic approach | 22:49 |
| tristanC | right, but i guess the issue is that you are bound to a single distribution packages, and its trick to fix eventual conflicts | 22:50 |
| tristanC | nixos (or guix) is indeed more flexible to manage conflicts, but you would still bound to their distrubtion | 22:51 |
| fungi | it's still an answer for how you might install otherwise "incompatible" packages on the same system, to answer your original question. if you're looking for a single distribution to provide them, then worrying you're beholden to that single distribution is a bit of a false dichotomy | 22:55 |
| fungi | also anaconda is probably a good parallel to e.g. dockerhub, a way for independent package managers to publish their work so it can be consumed by a consistent toolset but not conflict with the packages managed by others even if they depend on different versions of common libraries | 23:02 |
| tristanC | fungi: well with container, you are not beholden to a single distribution, and it may be inefficient, but it works for any languages or package system | 23:03 |
| fungi | sure, again anaconda isn't limited to one programming language either, and provides distinct separation between package installation paths | 23:04 |
| fungi | (it just happens to get used mostly for python) | 23:04 |
| fungi | and to be clear, i'm not really a fan of anaconda, for the same reasons i'm skeptical of docker containers | 23:05 |
| fungi | your security and vulnerability management goes from "this is the distribution's policy on handling security-critical updates" to "well it's all over the place but whoever they are i'm sure they've got this" | 23:07 |
| tristanC | heh, i'm also skeptical of docker containers, or more generally the container image format, but i am not aware of a better solution | 23:07 |
| *** rlandy has quit IRC | 23:07 | |
| ianw | ultimately, a micro-kernel, components talking via rpc channels and capabilities | 23:23 |
| fungi | bring back the hurd | 23:25 |
| fungi | well, okay, *finish* the hurd ;) | 23:26 |
| tristanC | there is also https://mirage.io/ | 23:30 |
| fungi | unikernels seem like the exact opposite of microkernels | 23:31 |
| fungi | "build all your userspace into the kernel" vs "build as much of your kernel as possible into composable userspace" | 23:32 |
| fungi | and their usages are similarly polarized: reboot with a new kernel to replaces the entire system" as opposed to "replace most/all of your kernel bit at a time without rebooting" | 23:34 |
| *** tosky has quit IRC | 23:51 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!