I have been using OpenSUSE Tumbleweed for a while but have decided to come back to EOS after looking at the new release. I have a few questions/issues that I would like some suggestions/recommendations on please.
Should I use BTRFS and Snapper (with the BTRFS assistant and GRUB) or the SystemD boot. The main reason I was using OpenSUSE was because it offered the option to rollback if anything broke but I’m not sure it’s that needed as I haven’t seen anyone have any trouble with updates on EOS for ages. Is EOS/Arch stable enough that the need to rollback is so rare, it isn’t needed?
I have been trying to work out whether to use EXT4 or XFS for storing data, mainly on my two other internal drives - an SSD for Steam games and a Data drive for storing my music, etc. Sort of like a backup in case the main drive has issues (broken OS updates, etc - see 1.) Is there any advantage to one over the other in terms of speed or reliability? I’ve read through loads of information on both but can’t really decide which to use. I also have a NAS which I backup data from via a USB dock to external drives so would like the most reliable filesystem on them so I don’t lose any data.
That’s pretty much it, as you can see there are two issues but they are both very related; they are essentially the same thing - I ideally want the fastest possible performance but with the best reliability/disaster-recovery.
Thanks in advance for your help everyone, I’m really excited to try the new release, especially as I love the Plasma desktop and think this could well be my daily-driver permanently. Also a massive thank you and well done to the maintainers, this release looks amazing and if it is all I expect it to be, I may have found my Linux home. I will be supporting the project by the way.
I happen to use both in parallel. Both openSUSE and EOS are quite stable. Maybe the EndeavorOS community is better. I agree that a well-maintained system doesn’t need much to restore, EOS is like that, but I also have good experiences with openSUSE.
I have had snapshots for years and have never once needed to roll back an Arch-based system.
xfs does have some advantages but if you are just putting a filesystem on a disk and using it, I doubt you would see much difference. Both are highly mature filesystems that work fine.
These are all rare/specific-to-user cases, and furthermore the kinds of things I experienced in “my experimental, not sure what I’m doing days” of Linux. You really shouldn’t sound the alarm for things like these.
Arch (and Linux in general) is very stable - especially if you use one of the rollback/system restore tools the OP mentioned. And even more so if you learn what does what, whether a feature is actually needed, etc. before trying to install/implement it.
I’m new to EOS and Arch in general. Have been using EOS now for about 2-3 months and no issues whatsoever.
Generally speaking, most issues tend to occur when users tweak their systems with who knows what kinds of customizations or configurations.
I make very few changes to the defaults installs other than to add a couple of my favorite apps.
Good backups are the most important thing to avoid issues.
On a rolling release, changing dependencies, new features as well as larger changes in kernel and drivers can, once in a while, lead to quite “experimental” stadiums, don’t they?
That was not my intent. Not at all, not remotely.
I’m not challenging that. It’s why I am trying hard to get that engulfed EOS partition back to functioning.
.
Same for me. It took almost 6 months before my EOS was sunk by yay.
Which is, to some degree, the very heart of using Linux on a personal machine or personal network component (in contrast to a public or an enterprise server). Right?
By the way, your statement is the same established explanation for why Windows OSes do not work reliably (“It’s CCleaner/Classic Shell/Open-Shell-Menu/alternative file manager/[…]'s fault!”).
Actually, in most cases, it’s the user’s fault. Not the additional/alternative app they added, but the fact that sometimes we get excited to customise and fail to pay proper attention to and make fail safes for any warnings prior to customising.
This happened quite often for people using BURG to customise their GRUB (in place of GRUB really), and equally for those trying to add speciality kernels like RT, Hardened, etc.
As far as I can tell, I did not use BURG in any way. I did not try to replace GRUB or to add kernels or such. I did not even try to customise GRUB, I only wanted to have a tool at hand for repairing eventual boot failures.
In the PC session leading to EOS’s boot failure, I did not do anything prior to or after yay system update. yay was really the breakpoint here, whatever the deeper, maybe longstanding root of the problem might be.
.
Even if so, that is not @mjl’s question. @mjl just wants to know whether and which precautionary measures should be taken – really aside from question of fault. I only wanted to contribute to the former topic.
I just thought I would throw my choices (so far) into the discussion here. I set up my systems with all data on a separate drive (with redirected ‘standard’ directories such as Music and Downloads), and on that data drive I tend to run either XFS or ZFS - but another system still on ext4 has never given me problems. I suspect hardware would cause troubles before ANY of the file systems would.
As for btrfs, I gather that it now has achieved reliable status - but I skip it as I have never seen a need for snapshots. IF an update gives trouble (the most likely cause for trouble for me - I don’t ‘play’ as much as I used to!) there exist 2 solutions with EnOS for handling it - downgrade and eos-shifttime - which can back things up to a situation before the upgrade difficulties began (thanks Arch!). I have never regretted NOT having snapshots, but as I mentioned I don’t ‘play’ the way I used to back when a reinstall seemed reasonable!
Not directly an answer to your questions, but my experiences with stability and recovery in the couple of years I’ve been on EnOS…
When I switched from 'Buntu to Manjaro I heard all those horror stories about updates “bricking” ( that word was even censored from the Manjaro forum back in the day) your computer and that if you are not using Timeshift you’re just playing with fire, because rolling release distros are scary.
I’ve used Timeshift for a couple of years, even when I switched to EndeavourOS, out of habit, but I haven’t needed to use it. In fact, it was counterproductive – I only ever restored snapshots when I did something stupid, instead of fixing my mess myself. So it actually slowed down my learning, by providing me with an unnecessary crutch. So, one day, I just deleted it and deleted all the snapshots. It’s been a couple of years and I don’t miss it at all. I never actually needed it.
Btrfs snapshots are a bit different than that, but I haven’t yet found the need or motivation to leave ext4 behind.
Yay did NOT break your installation. YOU did. EndeavourOS did not break your installation. YOU did. You could have reinstalled the mess you made many times over at this point. It’s also considered rude to enter other peoples threads and link back to yours.
You appear to be diverting the focus of this discussion by consistently including hyperlinks to your previous threads. It’s worth noting that digital forums have established etiquette to maintain a civil and organized conversation.
Like @dalto already mentioned, #1 is purely personal preference. I used BTRFS and snapshots for a while … until I needed the drive space. Now I just store my important stuff on a NAS with USB backup (same as you); and rebuild if I actually (or eventually) screw anything up. I find it less stressful and sometimes Zen-like.
I use EXT4 for all my other drives, which perform great for what I run on them.
that word was even censored from the Manjaro forum back in the day) your computer and that if you are not using Timeshift you’re just playing with fire, because rolling release distros are scary.