Eugene Klimov Slach

Счётчики, метрики и трейспоинты в ClickHouse (черновик).

Мотивация

В ClickHouse разным образом реализовано некоторое количество функциональности, имеющей общие черты. Это:

метрики на основе инкрементальных счётчиков (ProfileEvents);
метрики на основе количества одновременных процессов (Metrics);
метрики на основе асинхронно получаемых значений (AsynchronousMetrics);
ограничения на сложность выполнения запроса (встроены в Settings);
прогресс выполнения запроса (Progress, отправляется клиенту);

Proxmox VE Installation on Hetzner Server via Rescue System

Follow these steps to install Proxmox VE on a Hetzner server via the Rescue System. The Rescue System is a Linux-based environment that can be booted into to perform system recovery tasks. We'll be using it to install Proxmox VE.

In order to complete the process, it is indeed necessary to first boot into the Rescue System and then connect to it via SSH. This will allow you to run the commands for installing Proxmox VE. Here are the steps:

Starting the Rescue System

Log into the Hetzner Robot.
Under "Main Functions; Server" select the desired server and then open the tab "Rescue".

Enable `systemd` in WSL 2

**NOTE: If you have Windows 11 there is now an official way to do this in WSL 2, use it if possible - see MS post here (WINDOWS 11 ONLY)**

This guide will enable systemd to run as normal under WSL 2. This will enable services like microk8s, docker and many more to just work during a WSL session. Note: this was tested on Windows 10 Build 2004, running Ubuntu 20.04 LTS in WSL 2.

To enable systemd under WSL we require a tool called systemd-genie
Copy the contents of install-sg.sh to a new file /tmp/install-sg.sh:

In order to compile a fully static binary when using Cgo you'll need to link in a C library like musl.

I find it convenient to have a Docker image ready for building these artifacts.

FROM golang
RUN wget https://musl.libc.org/releases/musl-1.2.5.tar.gz && \
   tar -xzf musl-1.2.5.tar.gz && \
   cd musl-1.2.5 && \
 ./configure --enable-static --disable-shared &amp;&amp; \

	SELECT
	count(),
	arrayStringConcat(arrayMap(x -> concat(demangle(addressToSymbol(x)), '\n ', addressToLine(x)), trace), '\n') AS sym
	FROM system.trace_log
	WHERE (query_id = 'ebca3574-ad0a-400a-9cbc-dca382f5998c') AND (event_date = today())
	GROUP BY trace
	ORDER BY count() DESC
	LIMIT 10

	Row 1:

	FROM centos:7

	RUN yum install -y python3 python3-devel python3-pip unixODBC unixODBC-devel python3-setuptools gcc gcc-c++ https://download.dremio.com/odbc-driver/dremio-odbc-LATEST.x86_64.rpm && \
	pip3 install apache-superset sqlalchemy_dremio && \
	export LC_ALL=en_US.UTF-8 && \
	superset db upgrade && \
	superset init && \
	export FLASK_APP=superset && \
	flask fab create-admin --username admin --firstname admin --lastname admin --email admin@admin --password admin

	By default, EBPF programs will not run on WSL2 due to required kernel modules missing. The following example error is an
	indication of this problem:
	modprobe: ERROR: ../libkmod/libkmod.c:586 kmod_search_moddep() could not open moddep file '/lib/modules/4.19.84-microso
	ft-standard/modules.dep.bin'
	modprobe: FATAL: Module kheaders not found in directory /lib/modules/4.19.84-microsoft-standard
	chdir(/lib/modules/4.19.84-microsoft-standard/build): No such file or directory

	To fix this you need to rebuild the WSL2 kernel with the missing kernel modules. The below instructions are for Ubuntu 18.04 WSL2.

	1. git clone https://github.com/microsoft/WSL2-Linux-Kernel.git

	package seriesList

	import (
	"fmt"
	"math"
	"strconv"

	"github.com/go-graphite/carbonapi/expr/helper"
	"github.com/go-graphite/carbonapi/expr/interfaces"
	"github.com/go-graphite/carbonapi/expr/types"

	-- in different versions of CH it works differently, and does not work at all with circle replication


	drop table t1 on cluster segmented;
	drop table t2 on cluster segmented;
	create table t1 on cluster segmented (A Int64) Engine=MergeTree order by tuple();
	create table t2 on cluster segmented (A Int64) Engine=MergeTree order by tuple();
	create table t1d on cluster segmented as t1 Engine=Distributed(segmented, currentDatabase(), t1, A);
	create table t2d on cluster segmented as t2 Engine=Distributed(segmented, currentDatabase(), t2, A);

	#!/usr/bin/python
	# @lint-avoid-python-3-compatibility-imports
	#
	# profile Profile CPU usage by sampling stack traces at a timed interval.
	# For Linux, uses BCC, BPF, perf_events. Embedded C.
	#
	# This is an efficient profiler, as stack traces are frequency counted in
	# kernel context, rather than passing every stack to user space for frequency
	# counting there. Only the unique stacks and counts are passed to user space
	# at the end of the profile, greatly reducing the kernel<->user transfer.