bcc VS zfs

On Linux, this can be done using BPF (Berkley Packet Filter). In fact there is a tool in BCC[0] called filetop, which lists reads/writes by process and file[1].

0. https://github.com/iovisor/bcc

1. https://github.com/iovisor/bcc/blob/master/tools/filetop.py

I recommend you investigate using eBPF hooks in the linux kernel for this allocation tracking, it is measurable much faster. E.g. https://github.com/iovisor/bcc/blob/master/tools/memleak.py

Hey there! Of course. There are a few good examples here and here. Yes, they're specific tools (which I, by the way, do recommend), but you can have a look at the BPF code here as well.

Thanks for the link :) this one seems pretty much like what I was looking for, with Vxlan encapsulation, will bother you if i face issues on running the script they have there. Looks like couple of years old since the last contribution to that git page: https://github.com/iovisor/bcc/tree/master/examples/networking/tunnel_monitor

What we saw was stalling in the kernel run queue. One of the other people on our team did some debugging with eBPF. I think they may have used runqlen.py. But as soon as they evicted the one damaging pod from the machine, all the other workloads were no longer starved.

We set out to look for an additional probe that can help us complete the picture. Even then, we find eBPF useful to explore the linux TCP stack. Tools like stacksnoop or stackcount can be used to understand the flow a network packet is going through when it’s processed and compare different functions to see how “noisy” each function is. Searching for data probing locations consists of a constant trade-off between being too nosy and being blind, and we’re looking for the sweet spot in the middle.

If you're interested specifically in bash, you can look into bcc's bashreadline to output user commands. If you're interested in applying security policies to potential user commands, you can also take a look at Tracee although other open source solutions exist here as well.

That's right! There's no "traditional" loops as programs have to be proved to terminate at some point.

That being said, very recently support for bounded loops landed [0]. It's very exciting and useful, and I've seen it reduce verification times significantly, but we can't use this yet as it requires kernel 5.3 or greater, and we would like to support as many users as possible!

[0]: https://lwn.net/Articles/877062/

[1]: https://github.com/iovisor/bcc/commit/38304256c49a02aecbf78f...

Yup, this popped up for me about a month ago on a Raspberruy Pi system, fortunately not on the root, but still. On ARM the kernel_neon_begin / kernel_neon_end symbols were made GPL only[0] which is quite puzzling on many levels.

The discussion of the reported issue on the OpenZFS GitHub is very educational, both on the maintainers side (useful, sensible), and the other commenters syde (kinda "thermonuclear")[1]

In the meantime I just downgraded the kernel as well, fortunately still had it in the pacman cache; then pinned it for now.

Sent an email to to the patch creator / signer-off too, to feed back on that "this is not expected to be disruptive to existing users." part of the commit (as in [0]).

[0]: https://lore.kernel.org/all/[email protected]

[1]: https://github.com/openzfs/zfs/issues/14555

See: https://github.com/openzfs/zfs/releases/tag/zfs-2.1.9

Recompiling kernel modules does not fix them if they are not compatible with the kernel, interfaces can change. Take a look at the descriptions here: https://github.com/openzfs/zfs/releases These are code releases, not compiled modules.

Phase 1 first step involves compiling the DKMS but using zfs 2.1.8 to avoid the non-deterministic send-stream produced if Embedded Blocks feature is enabled, as I want to send streams using fifo-split

This is a fast-released bugfix release. Major impact is resolving an issue with native encryption: https://github.com/openzfs/zfs/issues/14413

Just use 2.1.8, it's out now. Officially supports 6.1.x https://github.com/openzfs/zfs/releases/tag/zfs-2.1.8

I am also using encryption btw. Though I was just told today that you apparently really don't want to use encryption with 2.1.8. Pretty nasty bug