pyznap VS zfs

Sanoid/syncoids been mentioned but honestly for once a week learning by doing . Pyznap also excellent when you want to automate. https://github.com/yboetz/pyznap

In contrast, zfs snapshots are immutable, and thus anything short of a hardware failure can be addressed with a simple zfs rollback command. This includes deliberate, accidental, and malicious actions. They can also be automated (I personally use pyznap but syncoid is also quite popular), creating what is effectively an incremental backup. I maintain - for each dataset - 24 hourly, 7 daily, 6 monthly, and 1 yearly snapshot. Additionally, I have a wholly separate server that wakes up once a day to ingest these snapshots via zfs send/recv, so even if I made a horrible mistake or suffered a catastrophic hardware failure, I could completely restore from the other server. This last point brings snapshots firmly into the realm of backups, IMO.

Another favorite option is Pyznap, which is python based and originally created to have have a few features and changes compared to sanoid. The author is also active here on reddit. I and not sure what the differences are anymore, it'll come down to trying them and preference.

I have two Supermicro X9 2Us, each with Proxmox. One has allegedly existed solely as a backup target, which wakes up daily to ingest ZFS snapshots using pyznap. Unfortunately, for reasons which are unclear, this particular node doesn't always like to see its boot device, which is an NVMe drive. It's the exact same board as my primary, with the exact same modified BIOS to allow booting from NVMe. It usually takes 2-3 cycles before it'll see it and boot.

I have two nearly identical systems, both running Proxmox, with Debian VMs. One is a backup, which (once this is worked out) will wake up daily to ingest incremental backups. I'm using pyznap to handle the backup strategy.

https://github.com/yboetz/pyznap/issues/1#issuecomment-351015432

Be careful if you use ZFS-on-root, make sure not to snapshot bpool or it will brick your system and require a complete reinstall.

https://github.com/openzfs/zfs/issues/13873

> The only non-junk PCIe3 option that's even advertised here recently is the overpriced WD Red SN700.

Those WD drives seem to have some real issues, at least with ZFS and btrfs. :(

https://github.com/openzfs/zfs/discussions/14793

IMO the "next fs" is just zfs. They somewhat recently merged RAIDZ expansion feature https://github.com/openzfs/zfs/pull/12225 and make regular improvements. If no file system has what you need today, zfs will probably be the first one to have it "tomorrow," imo.

https://github.com/openzfs/zfs/issues/15526#issuecomment-181...

> zpool get all tank | grep bclone

> kc3000 bcloneused 442M

> kc3000 bclonesaved 1.42G

> kc3000 bcloneratio 4.30x

> My understanding is this: If the result is 0 for both bcloneused and bclonesaved then it's safe to say that you don't have silent corruption.

A couple years ago, I had an idea for convincing a filesystem to go faster using 2 compression steps instead of one. I couldn't see why it wouldn't work, and I also couldn't convince myself it should.

It seems to have worked out. [1]

[1] - https://github.com/openzfs/zfs/commit/f375b23c026aec00cc9527...

https://github.com/openzfs/zfs/issues/7631

This is a long-standing issue with zvols which affects overall system stability, and has no real solution as of yet.