| 1 Nov 2021 |
Ryan Burns | It looks like most of that is due to the transient sphinx error though, what do we do about that? | 19:53:11 |
Vladimír Čunát | The queue for x86_64-darwin was even empty now. So now I restarted all staging-next failures (in the last evaluation). | 19:54:19 |
Vladimír Čunát | I'll hope that tomorrow it will be clearer that there are no large build regressions. | 20:03:19 |
Ryan Burns | I also noticed the rubyPackages.nokogiri failure breaking a couple dependers, but couldn't see an obvious cause and have resigned to bisecting it. It'll probably take a long time so I wouldn't hold up the merge for it | 20:04:09 |
Ryan Burns | * I also noticed the rubyPackages.nokogiri failure on x86_64-darwin breaking a couple dependers, but couldn't see an obvious cause and have resigned to bisecting it. It'll probably take a long time so I wouldn't hold up the merge for it | 20:04:24 |
hexa | we still have a ZHF upcoming | 21:59:46 |
hexa | from nov 5th | 22:01:14 |
das_j | That might be a good time to drop this PR here ;) https://github.com/NixOS/nixpkgs/pull/144061 | 22:11:36 |
das_j | Also what's the general opinion about the coreutils/ZFS thing? I'd really love to have a coreutils without that current patch on 21.11. The ZFS folks are working on a workaround (which consists of just disabling the broken functionality on their side). Do you think this can slip into the last staging(-next) cycle among the removal of the patchß0? | 22:13:58 |
das_j | * Also what's the general opinion about the coreutils/ZFS thing? I'd really love to have a coreutils without that current patch on 21.11. The ZFS folks are working on a workaround (which consists of just disabling the broken functionality on their side). Do you think this can slip into the last staging(-next) cycle among the removal of the patch? | 22:14:00 |
Vladimír Čunát | I find it a bit problematic that ZFS version on OS is (generally) quite decoupled from the coreutils version in nixpkgs. | 23:00:12 |
das_j | We could mention in the release notes that a reboot is highly recommended for ZFS users and only drop the patch after 21.11 is released? | 23:01:48 |
das_j | * We could mention in the release notes that a reboot is highly recommended for ZFS users and only drop the patch after 21.11 is released (with 21.11 backport)? | 23:02:24 |
Vladimír Čunát | I didn't mean just a reboot. The host may not even be NixOS. | 23:03:19 |
das_j | Oh, right | 23:03:30 |
das_j | So drop it on master after 21.11 is released and don't backport it? | 23:04:59 |
das_j | that gives ~6 months of time for users who are using the stable channels | 23:05:14 |
Vladimír Čunát | I think there are common use cases for nix(pkgs) to get flexible packages even on quite old/conservative systems. No idea how much of that might be ZFS... but hopefully the affected cases will be able to update their OS with the ZFS patch, too. | 23:05:57 |
Ryan Burns | Certainly doesn't help that red hat killed off centos 8 so we're going to have enterprise users on centos 7 until 2024 | 23:07:08 |
Vladimír Čunát | Do we know potential risks/down-sides from keeping the patch? | 23:08:50 |
Vladimír Čunát | If it's not bad, the non-backporting approach sounds nice to me. Though I don't really know anything about this stuff :-) | 23:09:59 |
Vladimír Čunát | It probably won't really help cases like CentOS 7, but it will give some buffer. | 23:10:42 |
das_j | I don't think RH supports ZFS on CentOS 7 anyway so users will use unofficial repos | 23:11:05 |
das_j | Probably:
# yum search zfs
Loaded plugins: fastestmirror, keys, rhnplugin
This system is receiving updates from RHN Classic or Red Hat Satellite.
Loading mirror speeds from cached hostfile
* centos-sclo-rh: mirror1.hs-esslingen.de
* centos-sclo-sclo: mirror.imt-systems.com
Warning: No matches found for: zfs
No matches found
| 23:11:21 |
Ryan Burns | heh, good pointt | 23:11:26 |
Vladimír Čunát | IIRC Red Hat officially supports XFS and maybe ext4, and experimentally btrfs. | 23:11:57 |
Ryan Burns | I think we'll be able to make a more informed decision once ZFS releases their fix and we start to see how other distros are adopting it | 23:12:18 |
Vladimír Čunát | They probably know better than us how their patches propagate. | 23:13:09 |
Vladimír Čunát | * They probably know better than us how their patches propagate to deployment. | 23:13:22 |
das_j | Alright, I'll ask on the workaround PR once it's merged | 23:16:10 |