fredcooke | On my ZFS mirror machine I'm having to go from 23.04 to 23.10 before I can go 24.04 - how annoying - lunar to mantic in progress, mantic to noble next up if it's not a total fail, fingers crossed but I used tmux so there's that at least | 10:45 |
---|---|---|
fredcooke | pretty bare machine - only interesting things on it are zfs and syncthing - so should go quickly and smoothly - have also been progressing the main lubuntu machine and tidied up the sources.list.d of the lubuntu canary machine until dist-upgrade showed nothing lurking. Got rid of i386 from the main box too. Making progress, bit by bit. | 10:49 |
eyeoh | you have to go from 23.04 to 23.10 because you made the mistake of going to a non-LTS release | 10:53 |
eyeoh | where as if you had stayed on 22.04, that will allow you to go to the next LTS, which is 24.04 | 10:53 |
eyeoh | even years for LTS. Which I think is the only sensible way to run Ubuntu; because who has time to upgrade every 6 months and deal with regressions | 10:54 |
eyeoh | whereas every Debian Stable release can be considered "LTS", but even better - because there's extra testing that goes on before each Debian Stable release | 10:54 |
eyeoh | all those issues people have in Fedora, Arch, and even Ubuntu - the Debian testing team will know about them too, and do their best to ensure such bad bugs don't make it to Stable | 10:55 |
eyeoh | so I'm very appreciative of those who run bleeding edge distros ;) | 10:55 |
fredcooke | done and done - she's on noble now and the zpool seems healthy | 11:34 |
eyeoh | nice | 11:35 |
fredcooke | eyeoh: it wasn't a mistake, it was the lowest and most stable version with a high enough zfs version - 22.10 and 22.04 did not have decent ZFS nor did 20.04 or 20.10 or 18.04 or 18.10 - 23.04 was the first with it and 23.04 is more stable than 23.10 | 11:35 |
fredcooke | I'm using a dubious ppa by a deceased legend for the 18.04 zfs that is under this machine - and I think I have to keep that version as I move though until I reach noble on it - which is a bit worrying as the vastly different kernel versions could cause chaos and make me sad | 11:37 |
fredcooke | not touching this one upgrade wise until I have all data mirrored on the noble one - there are some minor issues I haven't figured out in send/recv that I need to sort ASAP | 11:37 |
fredcooke | lol @ appreciative of bleeding edge users, me too - and I was one for ages, but not anymore - a friend used to run gentoo - he's on ubu - another uses arch - another uses nixos | 11:38 |
fredcooke | lots of hard work, but getting there :-) | 11:39 |
eyeoh | ah, that vaguely reminds me of how the closed nvidia drivers are only happy with a certain range of kernel versions | 11:39 |
eyeoh | I started diving into the rabbit hole of filesystems a few months ago and ended up going with btrfs since my use case is quite simple, and I didn't want to deal with extra steps especially when it came to live USBs | 11:40 |
eyeoh | and for the first time in my life, I feel I have a much more proper backup system in place; since snapshots make it so much easier to send off to external storage, compared to rsync | 11:41 |
fredcooke | good luck with that - if it's working for you, great - I researched it before I started buying hardware and decided no thanks | 11:46 |
fredcooke | I had the same feeling when I got most of my data off garbage media (used to have an old iomega 2 bay nas, and various HDDs etc) and had it all on ZFS - including finding and fixing some corruption along the way by doing copies from multiple sources where possible and doing recursive diffs then examining the variants and deciding which one was bad and good etc | 11:47 |
fredcooke | the drives attached to my not-yet-remote mirror machine (mid grade NUC) are 16TB - 1x Ironwolf Pro, 1x Exos X18, 1x Skyhawk AI - low chance of parallel failures | 11:49 |
fredcooke | similar deal with the real time use sets - except each set has an SSD in there - 4TB triple mirror 1x SSD 2x disparate HDD and 8TB triple mirrow same style | 11:50 |
fredcooke | the 4TB is actually 2 2 way mirrors with local sync - but that was probably a mistake - will be fine once I automate the snapshots | 11:50 |
eyeoh | I went with btrfs in part because I didn't believe everything that naysayers had said is relevant to me; and wanted to see for myself. But also because I'm influenced by Fedora's decision to go with it in 2020 | 12:02 |
eyeoh | so at least I'm in good company as far as newbie users go | 12:03 |
eyeoh | again, no raid for me | 12:03 |
eyeoh | I have at least two computers and backup; I can't really see anything going too wrong that I can't recover from. Which is much better than previously when I only had one not-up-to-date backup | 12:03 |
fredcooke | fair enough - zfs has not been without flaws - though I have not hit any of them - and did my homework on which versions to trust and avoid before beginning - so far so good including the fact that I have 8 HDDs on usb cables on two machines - and even that doesn't cause issues - provides you use an acceptable identifier when you add the disk and not something that isn't repeatable | 12:23 |
fredcooke | I would like a way of saying "don't mount" when receiving a snapshot since that config comes with it | 12:24 |
fredcooke | but it's manageable | 12:24 |
eyeoh | I mostly rely on the second pair of eyes that is the Debian Stable and Debian Stable Security team, and also Backports on my non-nvidia ThinkPad | 12:55 |
fredcooke | 2.1.11 in deb stable, 2.2.2 on noble 24.04 - 2.1.6 on 18.04 bionic with the ppa, 23.04 had 2.1.9 or something, bionic without ppa was 0.8.3 or lower or not there? and focal without is 0.8.3 and jammy 22.04 is 2.1.5 so I'm stuck on the PPA until I get to noble unless I build my own or something or find another ppa that can supply fresher builds that match jammy or focal | 13:02 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!