Hacker News Reader

Comment by barrkel | Hacker News Reader

I make snapshots every hour for 24 hours, every day for 7 days, every week for 4 weeks and every month for three months. This keeps the number of snapshots reasonable while keeping most stuff around should you accidentally delete or overwrite it. I can see a good argument for more fine grained snapshots (e.g. every few minutes) but I think I'd limit that to directories (and thus FSes) known to have frequent updates.

I sometimes blow away all snapshots if I have a big clear out or rewrite of data, or else that stuff sticks around for months.

View on HN

> I can see a good argument for more fine grained snapshots (e.g. every few minutes) but I think I'd limit that to directories (and thus FSes) known to have frequent updates.

Fun fact, the ZFS send-receive functionality was born with snapshots being created every few minutes in mind:

> ZFS send and receive wasn’t considered until late in the development cycle. The idea came to me in 2005. ZFS was nearing integration, and I was spending a few months working in Sun’s new office in Beijing, China. The network link between Beijing and Menlo Park was low-bandwidth and high-latency, and our NFS-based source code manager was painful to use. I needed a way to quickly ship incremental changes to a workspace across the Pacific. A POSIX-based utility (like rsync) would at best have to traverse all the files and directories to find the few that were modified since a specific date, and at worst it would compare the files on each side, incurring many high-latency round trips. I realized that the block pointers in ZFS already have all the information we need: the birth time allows us to quickly and precisely find the blocks that are changed since a given snapshot. It was easiest to implement ZFS send at the DMU layer, just below the ZPL. This allows the semantically-important changes to be transferred exactly, without any special code to handle features like NFSv4 style ACLs, case-insensitivity, and extended attributes. Storage-specific settings, like compression and RAID type, can be different on the sending and receiving sides. What began as a workaround for a crappy network link has become one of the pillars of ZFS, and the foundation of several remote replication products, including the one at Delphix.

* https://www.delphix.com/blog/delphix-engineering/zfs-10-year...

It remains a _delightfully_ straightforward way to do backups; I do incremental zfs send with hourly snapshots to my local NAS and also to an rsync.net ZFS account (https://www.rsync.net/products/zfsintro.html); it's incredibly resource un-intensive relative to your traditional scan-everything backup methods and I can trivially mount a past snapshot for browsing or spot restore or just do zfs send on the whole thing for catastrophic recovery.

As an added bonus, with (Open)ZFS encryption, you can send the 'raw' encrypted zfs send stream (over an encrypted SSH connection) to a destination where the file system is encrypted and where the zfs receive command does not have to decrypt it to save things.

That is, the remote destination does not need to have the key/passphrase to the file system to have an (encrypted) copy of the data.

If your production server goes down, you can restore the encrypted file system without the encryption key, and only when you try to mount the restored ZFS file system will you be prompted for the key/passphrase.

> This means that you can use ZFS replication to back up your data to an untrusted location, without concerns about your private data being read. With raw send, your data is replicated without ever being decrypted—and without the backup target ever being able to decrypt it at all. This means you can replicate your offsite backups to a friend's house or at a commercial service like rsync.net or zfs.rent without compromising your privacy, even if the service (or friend) is itself compromised.

* https://arstechnica.com/gadgets/2021/06/a-quick-start-guide-...

Do you happen to know if rsync.net supports this?

They support zfs send-recv and snapshots:

* https://www.rsync.net/products/zfsintro.html

* https://www.rsync.net/resources/howto/snapshots.html

Whether their ZFS implementation supports OpenZFS encryption is something you'll have to e-mail them about. They have an HN account:

* https://news.ycombinator.com/user?id=rsync

Edit: Yes, it seems that they do:

* https://news.ycombinator.com/item?id=26506263

"Do you happen to know if rsync.net supports this?"

Yes, we do.

If you get a zpool from us (which is a bit different than a "normal" rsync.net account) it is running the latest stable ZoL codebase and, thus, supports encryption and "raw send", etc.

However ...

As elegant and performant as 'zfs send' is, the benefits are realized in large datasets where efficiency and performance really matters. If you're just sending 200 or 400 or 800 GB of data to an untrusted destination (like rsync.net) you should probably just use borg[1][2].

[1] https://borgbackup.readthedocs.io/en/stable/

[2] https://www.stavros.io/posts/holy-grail-backups/

Hmm, well currently all my machines, phones etc backup to a nas (zol) which I then reclone up to an encrypted b2 bucket, but tbh the restores suck and they charge for traffic so I was contemplating just throwing em on rsync.net, you recon Borg->rs.n is the way to go instead of zfs send? Under 2tb..

Well, it's really up to you but let me put it this way ...

In order to give you a zpool of your own we need to give you a full blown VM (bhyve) with resource guarantees and your own IP address, etc.

So there is a 1TB minimum and no discounts.

Alternatively, if you just get a plain old rsync.net account the minimum account size is much smaller and if you're an expert and don't need any (borg specific) support there is a discounted plan[1].

[1] https://rsync.net/products/borg.html

Interesting I wasn't aware of the borg discount pricing, that may change my default position. I now know of three pricing schemes, basic, ZFS, and borg, the last two of which don't seem to be listed anywhere on the site. Got any other products hiding behind my ignorance? :)

If OpenZFS supported some kind of "virtual zpool" feature (I guess it doesn't exist yet) that you could provision out of your main ZFS infrastructure and hand off to the customer, would that be useful to reduce the resource overhead related to offering ZFS directly?

(Completely guessing, but I’ve managed a few pb of zfs) — It’s probably a zvol on the host exposed as a block device (or via iscsi or something) to the vm, zpool in a zvol is ok ;)

Oh neat. Composability is a helluva drug.

Do you have a sense of why dedicated/isolated infra (VM & IP) might be necessary to provide this service?

Isolation between users.

Are you offering a full IPv4-enabled VM or just IPv6?

We offer ipv4. Remember, it is ssh only so there is very little you can do with it other than backups.

We strongly encourage people to use ipv6, which we support everywhere ...

It's so handy! And I have personally tested a restore using it. It's a good idea to check out the read-only flags so you don't accidentally touch the file-system and invalidate an incremental backup.

> rsync ... incurring many high-latency round trips.

If I remember correctly, the rsync protocol is actually very nearly optimal in terms of minimising round trips.

If you have to actually compare the files on both ends, however, that's still many more round trips than necessary for zfs send, which is just sending all blocks with a birth time after the previous snapshot

Out of curiosity, what do you use to automate deleting of old snapshots? I’m aware of zfs-prune-snapshots, but what’s the best way to automate it (or something like it)? Cron job, or something else?

(For anyone wondering, https://github.com/bahamas10/zfs-prune-snapshots)

I use zfsnap.[1] It's been forever since I've worked on it (hence why it's now "archived" on GitHub), but it's stable for my use and I use it on all ZFS systems I manage.

[1] https://github.com/zfsnap/zfsnap

Checkout sandoid

sanoid/syncoid.

Can you take snapshots more often than hourly? For example can you take a snapshot every 15 minutes?

Yup; you could do every minute with no real problems if you like. I believe you start to run into performance issues with high (1000+) snapshot counts, and then mostly when actually doing snapshot operations, so you'd want to have something to trim snapshots so you only keep your minutely snapshots for an hour and so forth.

Reply on news.ycombinator.com