zstd 15 compressed btrfs on-top of dm_crypt/Luks2 on-top of dm_raid 5 on 4 slow 2.5" HDDs may be quite great for storage and ok-ish for accessing individual files.

But man is it slow in copying 3TB of files off of this thing onto another zstd 9 compressed btrfs on-top of dm_crypt/Luks2 on-top of dm_raid 5 on 4 also slow 3.5" HDDs.

#btrfs #Linux #LVM

What the fuck is #LVM doing when you issue the "unmount" command after you did a "sync" and when it reports being "100% synced"? (it's an LVM RAID5 just in case that changes anything)

Why is it taking that long to unmount?

WTF?!?

#Linux #NixOS

T-40h until my disks are full and until then I need to have found a solution (Need about 10TB of additional usable storage).

Either:
* By then I need to have bought and received multiple ≥5TB disks (it's a LVM-RAID5, so 2-3 additional ≥5TB drives)
* Create a 2nd LVM with 1TB (or 4TB) disks and do silly things with overlays
* find something better.

#homelab #datahorder #LVM

Best time to need new HDDs (not). And why? Cause my #LVM is stupidly sized (ok not really, but for the current storage prices it kinda is).

Would need 2-3x 5TB 2.5" HDDs.

Anyone know how reliable using 10x 1TB drives, grouped into 2x 5TB RAID0s for using them as pvs within another LVM-RAID5 would be?

#Linux #NixOS #Storage #HomeLab

Anyone know if there is something equally simplistic and universal than LVM that allows for storage policies?

Aka. instead of needing equally sized disks to get something like RAID-5/6 but with an arbitrary amount of drives in arbitrary sizes? (Without the capacity capping).

For now say like I'd have something silly like this:
4x 5 TB
2x 20 TB
20x 1 TB
1x 500 GB
+ change

Goal:
* Encryption at rest
* Tolerates 2 drive failures without any dataloss at all (by more partial)

#LVM #Linux #Storage

How to best organize various different-sized drives for data, apps and games #partitioning #mount #harddrive #filesystem #lvm

https://askubuntu.com/q/1564681/612

Oh "rclone copy" apparently doesn't force a sync to disk. And if you're running it a server with e.g. 512GB of RAM than this can result in a ":wq" in vim to appear as if it froze.

Is vim forcing a generic sync to disk when it exists? Or is this some kind of kernel bug that it doesn't just force that one file to sync to disk and instead waits for everything currently pending to be written?
What's going on here exactly?!?

#vim #rclone #linux #lvm