Klara Systems

There is the testimonial on Klara Systems’ website: "The developers at Klara Inc. have been outstanding in helping Wasabi resolve ZFS issues and improve performance. Their understanding of ZFS internals is unmatched anywhere in the industry" - Jeff Flowers, CTO, Wasabi Technologies https://klarasystems.com/ You could also search the OpenZFS repository for commits with the word Wasabi in them.

Any of the enterprise customers of klara Systems are likely ZFS production folks. https://klarasystems.com/?amp

> Assigning individual NICs or running a browser with extra isolation is one of the classic Linux namespace tutorials. With FreeBSD you can set up a standalone virtualized network stack: * https://klarasystems.com/articles/virtualize-your-network-on-freebsd-w… * https://man.freebsd.org/cgi/man.cgi?vnet * https://wiki.freebsd.org/Jails/VNET

Huh!? There was some amazing drama here, during the first ports of the early 90s. https://www.unixmen.com/freebsd-vs-openbsd/ and then the lawsuits: https://www.freebsd.org/doc/en_US.ISO8859-1/articles/explaining-bsd/wh… https://klarasystems.com/articles/history-of-freebsd-part-2-bsdi-and-u…

For movies/pictures and sound (well everything that is already compressed) you can deactivate it, zle too. EDIT: On the other hand leave it on, zfs tests the file if it's compress-able..if not it makes nothing: https://klarasystems.com/articles/openzfs1-understanding-transparent-c… >>ZFS Compression, Incompressible Data and Performance >>You may be tempted to set “compression=off” on datasets which primarily have incompressible data on them, such as folders full of video or audio files. We generally recommend against this – for one thing, ZFS is smart enough not to keep trying to compress incompressible data, and to never store data compressed, if doing so wouldn’t save any on-disk blocks.

Klara Systems | https://klarasystems.com | Technical Marketing Engineer | Remote | Full-time / Contract Klara Systems is a FreeBSD and ZFS engineering company serving enterprise customers across North America and Europe. Our team is senior, globally distributed, and operates with a high degree of autonomy. We’re looking for a Technical Marketing Engineer who turns engineering work into technical content that helps buyers evaluate Klara’s services and provides engineers with useful, technically accurate resources. can find. You will interview engineers and consultants, mine incident write-ups, support cases, upstream PRs, mailing lists, and delivery lessons for source material, then turn that material into technical articles, pages, comparisons, FAQs, and migration guides. You will also build scripts and LLM-powered workflows to automate knowledge extraction, reporting, and first drafts. You should have: - A background in software development, systems administration, or a related technical field. - The ability to hold a technical conversation with an operating systems or storage expert and turn it into something publishable. - Daily experience using AI tools to produce technically accurate and high-quality assets. - An understanding of technical buyer personas and how to build content they actually want to read. Apply: https://klarasystems.com/careers/technical-marketing-engineer/

When setting up root-on-ZFS on FreeBSD, it's worth knowing about boot environments (a concept originally from Solaris): * https://klarasystems.com/articles/managing-boot-environments/ * https://wiki.freebsd.org/BootEnvironments * https://man.freebsd.org/cgi/man.cgi?query=bectl * https://dan.langille.org/category/open-source/freebsd/bectl/ * https://vermaden.wordpress.com/2022/03/14/zfs-boot-environments-revolu… It lets you patch/upgrade an isolated environment without touching the running bits, reboot into that environment, and if things aren't working well boot back into the last known-good one.

No big fight, you just have to exclude the devices at boot so the vm can take them over [1] and if you have a AMD-System add that [2] and use the nvme virtual harddisk [3] [1] https://dflund.se/~getz/Notes/2024/freebsd-gpu/#bhyve-passthrough [2] https://wiki.freebsd.org/bhyve/pci_passthru#Additional_Notes [3] Creating a Windows Server 2019 VM using vm-bhyve : https://klarasystems.com/articles/from-0-to-bhyve-on-freebsd-13-1/

I use FreeBSD on my machines because it has netgraph: https://klarasystems.com/articles/inside-freebsd-netgraph-advanced-net…

> 3. Rudimentary resource controls with no default enforcement (important!). Seems pretty extensive to me, including R/W bytes/s and R/W ops/s: * https://docs.freebsd.org/en/books/handbook/jails/#jail-resource-limits * https://klarasystems.com/articles/controlling-resource-limits-with-rct… * https://man.freebsd.org/cgi/man.cgi?query=rctl

> - [0] https://klarasystems.com/articles/ I personally wrote several of them :)

> Are there any fun lab projects, posts, educational series targeted at FreeBSD? Klara Systems [0], Vermaden [1] and IT Notes [2] seems to be the most active and popular. - [0] https://klarasystems.com/articles/ - [1] https://vermaden.wordpress.com/posts/ - [2] https://it-notes.dragas.net/categories/freebsd/

You don't have to use ports, FreeBSD has supported pre-compiled binary packages since 2014. https://klarasystems.com/articles/a-quick-look-at-the-history-of-packa…

I use it for simple tasks where spotting a mistake is easy. Like writing language binding for a REST API. It's a bunch of methods that look very similar, simple bodies. But it saves quite some work Or getting keywords to read about from a field I know nothing about, like caching with zfs. Now I know what things to put in google to learn more to get to articles like this one https://klarasystems.com/articles/openzfs-all-about-l2arc/ which for some reason doesn't appear in top google results for "zfs caching" for me

Do you know if you're still using if_bridge? I remembered this article from klara that goes a bit more into the details. https://klarasystems.com/articles/using-netgraph-for-freebsds-bhyve-ne…

Klara Systems | ZFS Engineering Manager | US Remote | Full-time | $170k - $210k We are a bootstrapped, high growth company looking for a Engineering Manager to help us scale more efficiently. About us: We focus on software development services for open source and community-driven software. We are a dynamic, high growth startup with marquee customers whose names you will recognize. Customers look to us because of our intense focus on innovation and deep experience with open-source technologies. You will be excited by the chance to build structure and define process to facilitate our growth. Job details and applications: https://klarasystems.com/careers/zfs-engineering-manager/

See special-vdev in this article. https://klarasystems.com/articles/openzfs-understanding-zfs-vdev-types…

Massive IO performance over KVM due to the emulated NVME driver. Bhyve may be juvenile compared to KVM but it seems to have a better design and a better license. https://klarasystems.com/articles/virtualization-showdown-freebsd-bhyv… edit: also bhyve runs on SmartOS too, and I think Bryan is going to be more comfortable with a Solaris OS under the hood

With rctl [0][1] or I believe another hack was that you can with the tcsh shell using it as a man in the middle. [0] https://klarasystems.com/articles/controlling-resource-limits-with-rct… [1] https://wiki.freebsd.org/Hierarchical_Resource_Limits

> FreeBSD jails don't have a one-command way to install a preconfigured jail for a specific service FreeBSD does have that tool, its BastilleBSD: https://bastillebsd.org/ For example, this is the Bastillefile for running consul: https://gitlab.com/bastillebsd-templates/consul/-/blob/master/Bastille… > And they don't have a way (that I know of) to write a config file describing a bunch of jails and services, then run that file and install all of those services on a new machine Sounds like you want to use pot with nomad: https://klarasystems.com/articles/cluster-provisioning-with-nomad-and-…

Not according to the computer I'm typing this on, 31 years later :) FWIW, Klara has a good series on the history of FreeBSD: https://klarasystems.com/articles/history-of-freebsd-unix-and-bsd/ https://klarasystems.com/articles/history-of-freebsd-part-2-bsdi-and-u… https://klarasystems.com/articles/history-of-freebsd-part-3-early-days… https://klarasystems.com/articles/history-of-freebsd-part-4-bsd-and-tc…

> if you have 4 KB-sector disks then you should probably use a physical block size granularity of 4 KB for your ZFS pool rather than 512 bytes According to Allan Jude (OpenZFS & FreeBSD developer), you should never use 512-byte recordsize even on an HDD with 512-byte sectors. Always start from 4 KiBs, probably even from 8. https://klarasystems.com/articles/tuning-recordsize-in-openzfs

> What companies have actually found is that collaborating on OSS infrastructure projects is extremely beneficial, essentially outsourcing some of the work to focus more on their unique selling points. I think there are examples of this in FreeBSD: after a company effectively forked the project, FreeBSD kept rolling forward—with bug fixes, updated drivers, etc. The company(s) had to maintain larger and larger internal diffs to get all of these improvements. At some point they tended to just start giving back whatever wasn't their secret sauce and keep their own code as self-contained as possible. I think a popular workflow nowadays is to just track -CURRENT [1], and maybe branch when they cut a release of their own product. [1] https://klarasystems.com/articles/evaluating-freebsd-current-for-produ…

Klara Inc. | Multiple roles | Full-time (Contractor) | Remote | https://klarasystems.com/careers/ As a provider of open-source services, Klara focuses on the art of open software and community-driven development. We believe that in order to foster creativity and advance technology in an ethical fashion we must continue fueling the spirit of open source. At Klara, we rely on a mix of specialists and community to deliver FreeBSD, ZFS and Arm development services. We help customers standardize their environments running on FreeBSD and ZFS, and accelerate their platforms based on FreeBSD, ZFS and Arm. We are currently looking for: - FreeBSD Kernel Hackers - FreeBSD Operating System Developers - OpenZFS Hackers - Sys admins with FreeBSD/ZFS experience

> Huh? What are you talking about? ZFS has had a write cache from day 1: ZIL. The ZIL is specifically not a cache. It can have cache-like behavior in certain cases, particularly when paired with fast SLOG devices, but that's incidental and not its purpose. Its purpose is to ensure integrity and consistency. Specifically, the writes to disk are served from RAM[1], the ZIL is only read from in case of recovery from an unclean shutdown. ZFS won't store more writes in the SLOG than what it can store in RAM, unlike a write-back cache device (which ZFS does not support yet). [1]: https://klarasystems.com/articles/what-makes-a-good-time-to-use-openzf…

This is actually Part2. Part 1: https://klarasystems.com/articles/history-of-zfs-part-1-the-birth-of-z…

I don't think it was just license : https://klarasystems.com/articles/deep-diving-into-the-strengths-of-fr… Otherwise they could have taken NetBSD, OpenBSD or other BSD

Probably not, but ZFS is actually smart enough to store the uncompressed version if compression doesn't save space. In other words: zfs will try lz4 compression on those video files, notice that it doesn't gain anything and store it uncompressed. [0] https://klarasystems.com/articles/openzfs1-understanding-transparent-c…

That guide is for non-vnet jails. If you want vnet, use this one: https://blog.uidrafter.com/engineering/freebsd-jails-network-setup Why VNET? "VNET jails give each jail its own isolated copy of the network stack" https://klarasystems.com/articles/virtualize-your-network-on-freebsd-w…

> I doubt a highly tuned Linux box will be slower than the equivalent FreeBSD one. That tracks but there's been a lot of work being done on FreeBSD's stack as well. Most people are aware about the contributions Netflix has made to FreeBSD; they say their FreeBSD-based CDN streams terabytes per second [1]. In addition to the basic network stack, FreeBSD can load additional stacks like RACK TCP [2] stack (using Recent ACK, PRR, etc.) that can be loaded as a module and selected per-socket or as the system default. This is the re-written stack that Netflix uses after they helped develop it. It seems that FreeBSD is kind of a testbed for new network implementations like Netgraph [3], a graph-based networking subsystem that enables modular, real-time packet processing inside the kernel and Network Offload and Socket Splicing (SO_SPLICE) [4] that enables kernel-level TCP proxying, resulting in reduced CPU overhead and eliminating unnecessary data copies. I'm not usually into networking but this stuff sounds very interesting. [1]: "Case Study: Maintaining the World’s Fastest Content Delivery Network at Netflix on FreeBSD" - https://freebsdfoundation.org/netflix-case-study/ (PDF). [2]: "The RACK-TLP Loss Detection Algorithm for TCP" - https://datatracker.ietf.org/doc/html/rfc8985 [3]: "Inside FreeBSD Netgraph: Behind the Curtain of Advanced Networking" - https://klarasystems.com/articles/inside-freebsd-netgraph-advanced-net… [4]:…