diff options
| author | Paul Buetow <paul@buetow.org> | 2026-05-06 10:02:18 +0300 |
|---|---|---|
| committer | Paul Buetow <paul@buetow.org> | 2026-05-06 10:02:18 +0300 |
| commit | 138da68b60ea73e998ba1f450b1b202686c19dbb (patch) | |
| tree | 85448ab0b585929c9bf2c716bbc58e80377e2083 /gemfeed/DRAFT-unveiling-ior-ng-part-2.html | |
| parent | aeed324a599066b27e669b8bbb373fc592c3997e (diff) | |
Update content for html
Diffstat (limited to 'gemfeed/DRAFT-unveiling-ior-ng-part-2.html')
| -rw-r--r-- | gemfeed/DRAFT-unveiling-ior-ng-part-2.html | 245 |
1 files changed, 245 insertions, 0 deletions
diff --git a/gemfeed/DRAFT-unveiling-ior-ng-part-2.html b/gemfeed/DRAFT-unveiling-ior-ng-part-2.html new file mode 100644 index 00000000..73b67c75 --- /dev/null +++ b/gemfeed/DRAFT-unveiling-ior-ng-part-2.html @@ -0,0 +1,245 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> +<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"> +<head> +<meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> +<title>Unveiling I/O Riot NG — Part 2: install and compile once, run everywhere</title> +<link rel="shortcut icon" type="image/gif" href="/favicon.ico" /> +<link rel="stylesheet" href="../style.css" /> +<link rel="stylesheet" href="style-override.css" /> +</head> +<body> +<p class="header"> +<a href="https://foo.zone">Home</a> | <a href="https://codeberg.org/snonux/foo.zone/src/branch/content-md/gemfeed/DRAFT-unveiling-ior-ng-part-2.md">Markdown</a> | <a href="gemini://foo.zone/gemfeed/DRAFT-unveiling-ior-ng-part-2.gmi">Gemini</a> | <a href="https://snonux.foo">Microblog</a> | <a href="https://irregular.ninja">Street photography</a> +</p> +<h1 style='display: inline' id='unveiling-io-riot-ng--part-2-install-and-compile-once-run-everywhere'>Unveiling I/O Riot NG — Part 2: install and compile once, run everywhere</h1><br /> +<br /> +<span class='quote'>Draft — not in the gemfeed yet. Promote with the usual rename + index dance.</span><br /> +<br /> +<span>This is Part 2 of three. Part 1 is the demo-driven tour — what ior looks like, how the dashboard tabs work, how filtering and recording behave. This part is about the install dance for Rocky Linux 9 (with one annoying kernel-backport caveat) and, more interestingly, why you only have to do that dance on a single machine: the resulting binary is portable to every other Linux box thanks to CO-RE — Compile Once, Run Everywhere — plus full static linking. Part 3 is the under-the-hood companion (per-event schema, async-syscall caveats, the syscall-coverage probe generator, and post-mortem SQL on the parquet output).</span><br /> +<br /> +<span>If you came here for the dashboard tour, that's Part 1. If you want to know how the data pipeline is shaped after you've got ior running, that's Part 3. This one is for the moment between "I want to try this" and "OK, it's running on the box I care about."</span><br /> +<br /> +<a class='textlink' href='./DRAFT-unveiling-ior-ng-part-1.html'>Part 1: a guided tour</a><br /> +<a class='textlink' href='./DRAFT-unveiling-ior-ng-part-3.html'>Part 3: under the hood (schema, probe generator, ClickHouse)</a><br /> +<a class='textlink' href='https://codeberg.org/snonux/ior'>I/O Riot NG on Codeberg</a><br /> +<a class='textlink' href='./2018-06-01-realistic-load-testing-with-ioriot-for-linux.html'>the original I/O Riot post (2018)</a><br /> +<br /> +<a href='./unveiling-ior-ng/00-logo.png'><img alt='I/O Riot NG logo' title='I/O Riot NG logo' src='./unveiling-ior-ng/00-logo.png' /></a><br /> +<br /> +<br /> +<h2 style='display: inline' id='table-of-contents'>Table of Contents</h2><br /> +<br /> +<ul> +<li><a href='#unveiling-io-riot-ng--part-2-install-and-compile-once-run-everywhere'>Unveiling I/O Riot NG — Part 2: install and compile once, run everywhere</a></li> +<li>⇢ <a href='#installing-ior'>Installing ior</a></li> +<li>⇢ ⇢ <a href='#why-native-installation-is-a-mess'>Why native installation is a mess</a></li> +<li>⇢ ⇢ <a href='#what-the-docker-build-is-actually-doing'>What the Docker build is actually doing</a></li> +<li>⇢ <a href='#a-short-detour-ebpf-and-libbpfgo'>A short detour: eBPF and libbpfgo</a></li> +<li>⇢ <a href='#co-re--the-part-that-makes-the-binary-actually-portable'>CO-RE — the part that makes the binary actually portable</a></li> +<li>⇢ <a href='#a-note-on-cgo-overhead'>A note on cgo overhead</a></li> +<li>⇢ <a href='#if-you-want-to-go-deeper'>If you want to go deeper</a></li> +<li>⇢ <a href='#wrapping-up'>Wrapping up</a></li> +</ul><br /> +<h2 style='display: inline' id='installing-ior'>Installing ior</h2><br /> +<br /> +<span>The short answer: use Docker. One command, no toolchain setup, works from any Docker-capable Linux host with BTF available:</span><br /> +<br /> +<!-- Generator: GNU source-highlight 3.1.9 +by Lorenzo Bettini +http://www.lorenzobettini.it +http://www.gnu.org/software/src-highlite --> +<pre>git clone https://codeberg.org/snonux/ior ~/git/ior +cd ~/git/ior +mage buildDocker +</pre> +<br /> +<span>First run builds a Rocky Linux 9 builder image (~15–20 minutes). Subsequent runs reuse the cached image and finish in under a minute. The resulting static binary lands at <span class='inlinecode'>./ior</span>.</span><br /> +<br /> +<span>That's the officially supported install path, and it's the right one for anyone who just wants to run ior without living in its build system.</span><br /> +<br /> +<h3 style='display: inline' id='why-native-installation-is-a-mess'>Why native installation is a mess</h3><br /> +<br /> +<span>If you're curious why Docker became the answer, the native install on Rocky Linux 9 illustrates the problem well. Three separate things bite you before you even get to <span class='inlinecode'>mage build</span>:</span><br /> +<br /> +<span>Rocky 9 ships neither <span class='inlinecode'>libelf.a</span> nor <span class='inlinecode'>libzstd.a</span> — there are no <span class='inlinecode'>*-static</span> subpackages for either, only the dynamic <span class='inlinecode'>.so</span> files. Both have to be compiled from source. <span class='inlinecode'>libelf</span> from the elfutils source RPM, <span class='inlinecode'>libzstd</span> from the upstream GitHub release tarball.</span><br /> +<br /> +<span>Rocky 9 also only ships Go 1.25.x, but ior requires 1.26+. So Go itself has to be installed from go.dev in parallel with the library builds.</span><br /> +<br /> +<span>And there's a kernel quirk that used to make this section much longer. Pre-fix, ior would happily load on a stock 5.14 RHEL kernel and then die on the very first tracepoint attach with <span class='inlinecode'>BPF_LINK_CREATE</span>/<span class='inlinecode'>BPF_PERF_EVENT</span> returning <span class='inlinecode'>EACCES</span> — even as root, with SELinux permissive, with every BPF-related sysctl wide open. The cause is that RHEL 9 carries an <span class='inlinecode'>rt</span>-tree backport that adds <span class='inlinecode'>preempt_lazy_count</span> to <span class='inlinecode'>struct trace_entry</span>. That widens the BTF-emitted alias <span class='inlinecode'>trace_event_raw_sys_enter</span>/<span class='inlinecode'>_exit</span> by 8 bytes and shifts the <span class='inlinecode'>args</span>/<span class='inlinecode'>ret</span> offsets — but the actual context the kernel hands the BPF program is still <span class='inlinecode'>struct syscall_trace_enter</span>/<span class='inlinecode'>_exit</span>, where the offsets did not move. Programs written against <span class='inlinecode'>trace_event_raw_sys_*</span> (the conventional choice; bcc, libbpf-tools, and ior all used to do this) end up reading past <span class='inlinecode'>max_ctx_offset</span>, so the verifier rejects the attach. The fix — also what bcc shipped in [PR #4920](https://github.com/iovisor/bcc/pull/4920) and what inspektor-gadget did — is to type the BPF context as <span class='inlinecode'>syscall_trace_enter</span>/<span class='inlinecode'>_exit</span> directly. ior now generates its handlers that way, and stock 5.14 RHEL/Rocky/Alma works without an ElRepo kernel.</span><br /> +<br /> +<h3 style='display: inline' id='what-the-docker-build-is-actually-doing'>What the Docker build is actually doing</h3><br /> +<br /> +<span>The Dockerfile encodes exactly the same steps that a native install on Rocky 9 would require. Here is the full sequence so you have a mental model of what's inside the image, and so you could reproduce it on a bare host if you ever needed to:</span><br /> +<br /> +<!-- Generator: GNU source-highlight 3.1.9 +by Lorenzo Bettini +http://www.lorenzobettini.it +http://www.gnu.org/software/src-highlite --> +<pre><i><font color="silver"># 1) Enable repos and install build dependencies. CRB ships zlib-static / glibc-static.</font></i> +sudo dnf config-manager --set-enabled crb +sudo dnf install -y epel-release +sudo dnf install -y gcc clang bpftool elfutils-libelf-devel zlib-static \ + glibc-static libzstd-devel git make cmake wget rpmdevtools strace bpftrace +sudo dnf builddep -y elfutils + +<i><font color="silver"># 2) Install Go 1.26 from go.dev. Rocky 9 ships only Go 1.25.x, ior needs 1.26+.</font></i> +cd /tmp +wget -q https://go.dev/dl/go<font color="#000000">1.26</font>.<font color="#000000">2</font>.linux-amd<font color="#000000">64</font>.tar.gz +sudo tar -C /usr/local -xf go1.<font color="#000000">26.2</font>.linux-amd<font color="#000000">64</font>.tar.gz +echo <font color="#808080">'export PATH=/usr/local/go/bin:$HOME/go/bin:$PATH'</font> | sudo tee /etc/profile.d/go.sh +<b><u><font color="#000000">source</font></u></b> /etc/profile.d/go.sh + +<i><font color="silver"># 3) Build libelf.a from the elfutils source RPM.</font></i> +mkdir -p ~/src && cd ~ +dnf download --source elfutils-libelf +rpm -ivh elfutils-*.src.rpm +tar -C ~/src -xjf rpmbuild/SOURCES/elfutils-*.tar.bz2 +cd ~/src/elfutils-* +./configure --enable-deterministic-archives --disable-debuginfod --disable-libdebuginfod +make -C lib -j$(nproc) +make -C libelf -j$(nproc) +sudo cp -v libelf/libelf.a /usr/lib<font color="#000000">64</font>/ + +<i><font color="silver"># 4) Build libzstd.a from upstream (libzstd-devel doesn't ship the static archive).</font></i> +cd /tmp +wget -q https://github.com/facebook/zstd/releases/download/v<font color="#000000">1.5</font>.<font color="#000000">5</font>/zstd-<font color="#000000">1.5</font>.<font color="#000000">5</font>.tar.gz +tar xzf zstd-<font color="#000000">1.5</font>.<font color="#000000">5</font>.tar.gz +make -C zstd-<font color="#000000">1.5</font>.<font color="#000000">5</font>/lib -j$(nproc) libzstd.a +sudo cp -v zstd-<font color="#000000">1.5</font>.<font color="#000000">5</font>/lib/libzstd.a /usr/lib<font color="#000000">64</font>/ + +<i><font color="silver"># 5) Clone ior + libbpfgo, pin libbpfgo, build the static libbpf archive, install mage.</font></i> +mkdir -p ~/git +git clone https://codeberg.org/snonux/ior ~/git/ior +git clone https://github.com/aquasecurity/libbpfgo ~/git/libbpfgo +git -C ~/git/libbpfgo checkout v0.<font color="#000000">9.2</font>-libbpf-<font color="#000000">1.5</font>.<font color="#000000">1</font> +git -C ~/git/libbpfgo submodule update --init --recursive +make -C ~/git/libbpfgo libbpfgo-static +go install github.com/magefile/mage@latest + +<i><font color="silver"># 6) Generate the syscall-coverage handlers against THIS kernel and build.</font></i> +<i><font color="silver"># IOR_FORCE_GENERATE bypasses the strict diff against the committed audit file</font></i> +<i><font color="silver"># (the committed audit was generated against a different kernel build, and the</font></i> +<i><font color="silver"># generator's safeguard would otherwise refuse to overwrite it).</font></i> +cd ~/git/ior +env IOR_FORCE_GENERATE=<font color="#000000">1</font> GOTOOLCHAIN=auto mage generate +env GOTOOLCHAIN=auto mage all + +<i><font color="silver"># 7) Smoke test.</font></i> +sudo ./ior -plain -duration <font color="#000000">5</font> +</pre> +<br /> +<span>If you see <span class='inlinecode'>Probing for 5s</span> followed by CSV rows, the build is good. <span class='inlinecode'>mage buildDocker</span> runs all of this inside a container and hands you back just the final binary — the 15-minute first-run cost buys you never having to think about any of the above again.</span><br /> +<br /> +<h2 style='display: inline' id='a-short-detour-ebpf-and-libbpfgo'>A short detour: eBPF and libbpfgo</h2><br /> +<br /> +<span>If you haven't touched eBPF before: it's a small in-kernel bytecode VM. You compile a tiny C program, the kernel verifies it can't crash or loop forever, and then it runs every time some hook fires — a syscall enter/exit, a kprobe, a tracepoint, a network packet. The program writes events into a ring buffer that userspace mmaps and drains. No kernel module, no patched kernel, no debug symbols required.</span><br /> +<br /> +<span>ior plugs into the syscall tracepoints — <span class='inlinecode'>sys_enter_openat</span>, <span class='inlinecode'>sys_exit_read</span>, etc. — and the BPF side does the bare minimum: timestamp the event, copy a few fields, push to a perf ring buffer. All the heavy lifting (string interning, latency math, aggregation, the dashboard) is in Go on the userspace side.</span><br /> +<br /> +<span>The kernel ships a C library called libbpf that handles loading the program, attaching it to hooks, managing maps, and reading the ring buffer. There are two well-known ways to drive that from Go:</span><br /> +<br /> +<ul> +<li>libbpfgo (Aqua Security): a thin cgo wrapper around libbpf. You ship libbpf along with your binary and call into the same C API that <span class='inlinecode'>bpftool</span> and <span class='inlinecode'>perf</span> use.</li> +<li>cilium/ebpf: a from-scratch pure-Go reimplementation of everything libbpf does — ELF parser, BTF resolver, syscall layer, the lot.</li> +</ul><br /> +<span>I went with libbpfgo specifically because it's a wrapper, not a reimplementation. Whatever lands in libbpf upstream — new map types, new attach kinds, CO-RE fixes — I get for free the next kernel cycle. The pure-Go variant has to chase libbpf's feature set in parallel, and any divergence is on me to debug. For a tracer that's mostly value-add on the userspace side, "be a thin client of the kernel's own library" wins.</span><br /> +<br /> +<h2 style='display: inline' id='co-re--the-part-that-makes-the-binary-actually-portable'>CO-RE — the part that makes the binary actually portable</h2><br /> +<br /> +<span>The headline fact about ior's deployment story: build it once on one box, then <span class='inlinecode'>scp ior other-host:/usr/local/bin/</span> to anywhere else and it just runs. No recompile per kernel, no kernel-debuginfo dance, no DKMS hooks. Two mechanisms make that work, and they reinforce each other.</span><br /> +<br /> +<span>The first is plain old static linking on the userspace side. A quick refresher on what that means, since it's central to why "scp the binary anywhere" works: when you build a normal Linux executable, the linker has two ways to wire library code into your program. Dynamic linking ("shared library") leaves a placeholder in the binary that says "at run time, find <span class='inlinecode'>libfoo.so.6</span> somewhere on <span class='inlinecode'>LD_LIBRARY_PATH</span> and pull in its symbols." Static linking pastes the library's machine code directly into your binary at build time, so there's nothing to look up later. Dynamic is smaller on disk and lets distros patch shared libs without rebuilding everything; static is bigger but self-contained — no surprise about which version of the library the target box happens to have, no <span class='inlinecode'>error while loading shared libraries: libwhatever.so.6: cannot open shared object file</span> when the target ships a newer ABI.</span><br /> +<br /> +<span>For Go, this is mostly a non-issue. A pure-Go binary (no cgo) is statically linked by default — the Go toolchain produces a single self-contained ELF file with no <span class='inlinecode'>.dynamic</span> section and no <span class='inlinecode'>NEEDED</span> entries. You can <span class='inlinecode'>scp</span> it to any Linux box of the same architecture and it just runs. That's one of the quietly nice things about Go.</span><br /> +<br /> +<span>ior is the not-quite-pure case: it goes through cgo to call into libbpf, libelf, and libzstd, and each of those has its own .so on the build host. By default cgo links those C dependencies dynamically, which would defeat the "scp the binary anywhere" property — the target box would need to have matching <span class='inlinecode'>.so</span> files at matching versions, which is exactly the kind of dependency hell Go usually saves you from. The fix is the line <span class='inlinecode'>-extldflags "-static"</span> in ior's Magefile: it tells the external (C) linker to resolve <span class='inlinecode'>-lbpf -lelf -lzstd -lz</span> against the static archives (<span class='inlinecode'>.a</span> files) instead of the dynamic ones. That's why the install procedure above is so picky about having <span class='inlinecode'>libelf.a</span> and <span class='inlinecode'>libzstd.a</span> actually present on the build host — without them the C-side static link fails outright.</span><br /> +<br /> +<span>The result is a single ~23 MB binary with libbpf, libelf, libzstd, and zlib all baked in. None of them are looked up dynamically at runtime. The build host's library versions stay on the build host. (A couple of glibc resolver functions — <span class='inlinecode'>getpwnam_r</span> and friends — do still fall back to the target's libc, which is fine on any reasonable distro and is what the linker warnings during the build are about.)</span><br /> +<br /> +<span>The second, and the one that's actually unusual, is CO-RE — Compile Once, Run Everywhere. CO-RE is the eBPF feature that solves the "the kernel changed its struct layout between releases" problem.</span><br /> +<br /> +<span>The old I/O Riot was Systemtap. Systemtap programs are translated into a kernel module against the running kernel's exact headers, and that module then has to be loaded with <span class='inlinecode'>insmod</span>. That meant: the user has to install a kernel-debuginfo package matching their running kernel, and a fresh build per host (or per kernel update). On the BSD-style "you only run what you compiled here" laptop crowd that was tolerable; on a fleet of distros + kernel versions it was a recurring tax. Half of the original I/O Riot's README was about kernel-debuginfo dance steps.</span><br /> +<br /> +<span>CO-RE throws all of that out. The idea, in one paragraph: when you write a BPF program that reads <span class='inlinecode'>task->mm->start_stack</span>, you don't bake the offsets of those fields into the compiled program. Instead, the compiler emits relocation records ("at this instruction, fetch the offset of <span class='inlinecode'>mm</span> inside <span class='inlinecode'>task_struct</span>"). At load time, libbpf looks up the actual offsets in the target kernel's BTF (BPF Type Format — a description of every kernel struct, embedded in <span class='inlinecode'>/sys/kernel/btf/vmlinux</span> on any modern kernel) and patches the program in place. The same <span class='inlinecode'>.bpf.o</span> that ran on a 5.10 Debian kernel runs on a 6.8 Fedora kernel without recompilation.</span><br /> +<br /> +<span>Pictorially, the contrast looks like this:</span><br /> +<br /> +<pre> +Old I/O Riot (Systemtap) New ior (libbpf + CO-RE) +───────────────────────── ──────────────────────────── + .stp source .bpf.c source + │ │ + │ needs THIS kernel's headers │ build ONCE against vmlinux.h + │ + debuginfo package installed │ (generated from any kernel BTF) + ▼ ▼ + per-host translate + compile one portable .bpf.o + │ │ + ▼ ▼ + per-host kernel module same binary on every host + │ │ + insmod / modprobe libbpf loader: + │ │ • read /sys/kernel/btf/vmlinux + ▼ │ • patch field offsets + attached, this kernel only │ • verify + load + ▼ + attached, runs anywhere +</pre> +<br /> +<span>What that buys ior in practice: I ship a single <span class='inlinecode'>ior</span> binary. On any Linux ≥4.18-ish with BTF available (which is almost all of them now — Debian, Ubuntu, Fedora, Arch, RHEL, and ElRepo's <span class='inlinecode'>kernel-ml</span> builds all ship <span class='inlinecode'>CONFIG_DEBUG_INFO_BTF=y</span> by default), it just works. No kernel-debuginfo dependency, no per-kernel build matrix, no DKMS hooks. The first time I tried <span class='inlinecode'>scp ior fedora-box:</span> and it ran without complaint after a 6-month gap I had to double-check it wasn't silently doing nothing.</span><br /> +<br /> +<span>So the operational shape is: pick one box, do the install dance from the Rocky section above (or the Fedora one in the README) once, build, then distribute the 23 MB binary wherever you want to trace. The build host needs Go and clang and the static libraries. The trace hosts need a BTF-enabled kernel and <span class='inlinecode'>sudo</span>. That's it.</span><br /> +<br /> +<span>The runtime shape of a trace pipeline lines up with that:</span><br /> +<br /> +<pre> + kernel side userspace (this binary) + ─────────── ─────────────────────── + ┌──────────────────────┐ + tracepoint: │ Go process │ + sys_enter_openat │ ┌────────────────┐ │ + │ │ │ aggregator │ │ + ▼ │ │ (latency, │ │ + ┌─────────┐ │ │ stacks, │ │ + │ BPF prog│ ─── perf ring buf ──────────>│──│ filters) │ │ + │ (verified │ └─────┬──────────┘ │ + │ bytecode) │ │ │ + └─────────┘ │ ▼ │ + │ Bubble Tea TUI / │ + │ parquet writer / │ + │ CSV stdout │ + └──────────────────────┘ +</pre> +<br /> +<h2 style='display: inline' id='a-note-on-cgo-overhead'>A note on cgo overhead</h2><br /> +<br /> +<span>The cost of being a libbpf wrapper rather than a pure-Go reimplementation is cgo. Every call from Go into libbpf crosses the cgo boundary, which historically meant tens to ~hundred-ish nanoseconds of overhead per call — register save/restore, a stack switch onto g0, goroutine state bookkeeping. Cheap in absolute terms, but it adds up if you call into C inside a tight loop. ior keeps the actual hot path on the kernel side and only crosses into Go once per drained batch of events from the ring buffer, so the per-call cost is amortized over thousands of events. In practice it doesn't show up in profiles.</span><br /> +<br /> +<span>Go 1.26, the current release at the time of writing (early May 2026), is the one that finally took a serious bite out of cgo's per-call cost — the runtime can elide a chunk of the bookkeeping for calls that don't need it. Real-world wins depend heavily on the workload, but the rough direction is that cgo now feels closer to "an unusually expensive function call" than to "a context switch", which is the right mental model for almost everyone touching a C library from Go. The shorter version: cgo overhead used to be a real footgun for ports that called into C in the inner loop. With Go 1.26 it's a footnote unless you're doing many millions of small calls per second, in which case batching across the boundary still fixes it.</span><br /> +<br /> +<h2 style='display: inline' id='if-you-want-to-go-deeper'>If you want to go deeper</h2><br /> +<br /> +<span>If any of this sounds interesting and you want to learn how to write your own BPF programs, two books are the standard recommendations and both well worth the time:</span><br /> +<br /> +<ul> +<li>"Learning eBPF" by Liz Rice (O'Reilly, 2023) is the friendlier on-ramp. It walks through writing your first programs end-to-end, covers CO-RE and BTF in plain English, and is the book I'd hand to someone who has never touched the kernel side before. Liz also gave the canonical "what is eBPF" conference talk floating around YouTube, which makes a good 40-minute companion.</li> +<li>"BPF Performance Tools: Linux System and Application Observability" by Brendan Gregg (Addison-Wesley, 2019) is the encyclopedia. It's where you go after you've understood the basics and now want a complete reference for tracing every subsystem in the kernel — file systems, networking, scheduler, languages, applications — with worked tools for each. The flame-graph-driven analysis style throughout is also exactly how ior's own flamegraph tab thinks about a workload.</li> +</ul><br /> +<span>Between the two, Rice teaches you the moving parts and Gregg teaches you what to do with them.</span><br /> +<br /> +<h2 style='display: inline' id='wrapping-up'>Wrapping up</h2><br /> +<br /> +<span>That's the install dance and the why-it's-portable story. Part 3 is the bottom of the data stack — what's actually in each event row, the syscall-coverage safeguard against new kernels, async-syscall caveats, and how to query the parquet output with ClickHouse Local. Part 1, if you haven't read it, is the demo-driven tour with all the GIFs of the dashboard.</span><br /> +<br /> +<a class='textlink' href='./DRAFT-unveiling-ior-ng-part-1.html'>Part 1: a guided tour</a><br /> +<a class='textlink' href='./DRAFT-unveiling-ior-ng-part-3.html'>Part 3: under the hood (schema, probe generator, ClickHouse)</a><br /> +<a class='textlink' href='https://codeberg.org/snonux/ior'>Source on Codeberg</a><br /> +<a class='textlink' href='https://codeberg.org/snonux/ior/src/branch/main/docs/tutorial/tutorial.md'>The full in-repo tutorial</a><br /> +<p class="footer"> + Generated with <a href="https://codeberg.org/snonux/gemtexter">Gemtexter 3.0.1-develop</a> | + served by <a href="https://www.OpenBSD.org">OpenBSD</a>/<a href="https://man.openbsd.org/relayd.8">relayd(8)</a>+<a href="https://man.openbsd.org/httpd.8">httpd(8)</a> | + <a href="https://foo.zone/site-mirrors.html">Site Mirrors</a> + <br /> + Webring: <a href="https://shring.sh/foo.zone/previous">previous</a> | <a href="https://shring.sh">shring</a> | <a href="https://shring.sh/foo.zone/next">next</a> +</p> +</body> +</html> |