Can corruption happen by powering off the computer (not properly shutting it down) during balance or scrub of a btrfs volume?
Can corruption happen by powering off the computer (not properly shutting it down) during balance or scrub of a btrfs volume?
Edit: including corruption of superblocks
Yes there's always a chance corruption can happen from a hard power off, always keep reliable backups.
Everyone saying it'll be fine is speaking theoretically. Practically I can attest to full and total file system corruption under this scenario.
Especially if you're using raid5 for multi disk.
Corruption should be automatically detected and fixed. The operations will resume on bootup.
However I wouldn't push your luck.
Storage devices can fail at any time for any reason. Always have a backup.
Fwiw, I think BTRFS is better than ext4 and friends at actually detecting whether a block is corrupted or not.
"Better" in the sense that it actually has the ability to check for corruption at all, as all metadata and data are checksummed.
As long as the hardware functions as it should (e.g. respects barriers) and there is no software bug in the stack, no.
That's a highly unlikely scenario though. Make backups.
As long as there are no problems with the btrfs code? Hahahahaha!! There are.
And also in any other filesystem's code or the block layers below the filesystem. As I said, unlikely scenario.
I had some kernel panics here and there... but the last one panic was fatal. Suddenly a lot of /usr/lib/lib<name>.so files were empty and also X11 stopped working...
Theoretically, no.
In reality, possibly/yes!
What do you have?
As a proud Orange pi zero 3 owner (which I'm using it as a "lab rat" by testing several things, including shutting it down like its a router)....? Nah.
How's it been? I see mention of Orange Pi more frequently these days.
They're objectively better than the Raspberry Pi in every way and are much more standard ARM devices than the weird boot process of the Raspberry Pi, so generally speaking, more OSes just work.
My Orange Pi 5 actually supports an open source EDK2 port so it can run any Aarch64 operating system that supports UEFI and ACPI or Device Tree which means almost every Linux distro, all the BSDs, Windows, and even more exotic and up and coming options.
I actually bought it to test my own OS development project specifically because it's one of the few ARM boards that supports the common boot and firmware standards.
On the Raspberry Pi 5 which I also have if you want to use anything other than their own officially supported Linux distributions (so far only Pi OS and Ubuntu) then you have to modify your kernel or bootloader to work with its wacky boot ROM, lack of UEFI or U-Boot, and somewhat non-standard Device Tree along with tons of undocumented peripherals.
Oh, and the Orange Pi has twice the number of cores. The RPi 5 has four Cortex A76 cores while the Orange has four Cortex A76 cores and four Cortex A55 cores in a big.LITTLE configuration.
Honestly, any of the Rockchip RK3588 or RK3588S boards are way better than a Raspberry Pi. At this point, the only thing Raspberry Pi has going for it over its major competitors is the fact that the brand itself isn't Chinese (although many of the boards are made in China).
It's pretty good. Even better than the raspberry pi 4 considering its performance and comically low power draw (It doesn't go over the 2W mark even with both cpu and gpu at 100%.). Other than that, it's perfectly usable, despise its lack of (out of the box) support.
Since btrfs uses Copy on Write, as long as the data makes it onto disk in tact, any further btrfs operations on the data will be safe against sudden power loss. It might need the opportunity to repair some stuff once power is restored (scrub), but the data (and metadata) should still be there and recoverable, not left in some partial state that can't be resolved.
*with tact
Hah. Tactfully copied to disk intact*
No.
No