高手博客收藏夹

iptables 拦截 bridge 包的问题排查 via 卡瓦邦噶！ (author: laixintao)

Linux 内核之旅（二）：VFS 0x00 前言 Linux支持多种文件系统格式（如ext2、ext3、reiserfs、FAT、NTFS、iso9660等），不同的磁盘分区或其它存储设备都有不同的文件系统格式，然而这些文件系统都可以mount到某个目录下，使开发者看到一个统一的目录树，各种文件系统上的目录和文件，读写操作用起来也都是一样的。Linux内核在各种不同的文件系统格式之上做了一个抽象层，使得文件、目录、读写访问等概念成为抽象层的概念，因此各种文件系统看起来用起来都一样，这个抽象层称为虚拟文件系统（VFS，Virtual Filesystem） 基础结构 ● inode：定义，与磁盘上真实文件的一对一映射 ● file：定义，一个 file 结构体代表一个物理文件的上下文，不同的进程，甚至同一个进程可以多次打开一个文件，因此具有多个 file struct ● dentry：定义，多个 dentry 可以指向同一个 inode 内核数据结构 0x0 VFS 关联task_struct 0x0 files_struct 0x0 参考 ● VFS ● vfs ● File Descriptor Table ● Linux的虚拟文件系统VFS ● linux vfs轮廓 ● 从文件 I/O 看 Linux 的虚拟文件系统 via 熊喵君的博客

多动脑多思考 今天看到 Elon Musk 在 X 上发了一行动态。说是怎样从一串连续的数字中快速找到 缺少一个数字。我大致瞄了一眼，没找到。然后想到了一个巧妙的算法，快速找出了答案。 然后我把问题发给老婆。她居然也很快找到了结果。一问果然是从头开始找的。从而引出 了本文的主旨，要多动脑多思考。 via 涛叔

支持多开的docker版微信 docker-wechat 是一个开源项目，在 Docker 容器运行微信 linux 客户端，可以通过 web 或者 VNC 访问。 via wbsu2004的博客

我关注的项目这周更新了什么-20241206 简介 每周五更新，所有的更新也会发送到我的频道 https://t.me/bboyapp 规则 ● pre release的不记录 ● 只记录重要的更新 ● 上一版本版本的bugfix不记录，比如某一个项目最新版本是7.0.0 上一个版本6.0.0发布了一个bugfix 6.0.1 那么不会记录 ● 比如lens发布了一个新版本，那么openlens也会发布一个新版本，此时只会记录lens发布的信息，不会记录openlens发布的信息 via Bboysoul's Blog

使用 Fluid 和 S3FS 对接 S3 存储及性能测试 本文使用的是 Fluid 1.0 版本，高版本的配置文件路径发生了变化，需要根据实际情况调整。 1. 制作镜像 1.1 fluid_config_init.py 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 #!/usr/bin/env python import json import os rawStr = "" with open("/etc/fluid/config.json", "r") as f: rawStr = f.readlines() rawStr = rawStr[0] script = """ #!/bin/sh set -ex MNT_TO=$targetPath trap "umount ${MNT_TO}" SIGTERM mkdir -p ${MNT_TO} via 陈少文的网站 (author: 微信公众号)

Automatically group events and reduce noise with AI-powered Intelligent Correlation When you have a complex IT environment with many disparate tools, data sources, and teams, alert noise becomes overwhelming. This can delay incident response and cause missed alerts, ultimately leading to critical incidents and outages. Datadog Event Management’s Event Correlation groups and deduplicates events and alerts, reducing noise and helping response teams act on alerts faster. Traditionally, setting up and maintaining effective correlations required advanced knowledge in all of the services in your organization and teams and understanding what alerts might relate to the same issue. To help with this, Event Management provides an AIOps-driven solution that automates the setup and management of alert grouping and continuously improves existing correlations. via The monitor

Troubleshoot infrastructure changes faster with Recent Changes in the Resource Catalog Organizations often struggle to maintain visibility and control over their distributed cloud infrastructure, where changes in a single resource can have cascading effects throughout the system and potentially cause disruptions. In these environments, infrastructure changes that lead to incidents are often hard to troubleshoot—especially when teams are using disparate tools with siloed data—leading to longer resolution times, more downtime, and negative business outcomes. To troubleshoot infrastructure issues effectively, teams need a centralized view where they can track configuration changes to their cloud resources. via The monitor (author: sriram.raman@datadoghq.com (Sriram Raman))

Autodiscover Confluent Cloud connectors and easily monitor performance in Data Streams Monitoring Confluent Cloud is a Kafka–as-a-service solution that simplifies the deployment, scaling, and operation of Kafka clusters. A popular feature is its Apache Kafka connectors, which make it easy to connect your Kafka clusters to any of 120+ third-party streaming data sources and destinations. Connectors allow you to pull data into your pipelines without having to manage your own dedicated service, call any APIs, or integrate your Kafka streams into another system (such as a data warehouse or CRM application). via The monitor (author: jane.wang@datadoghq.com (Jane Wang))

Kubernetes 1.32 – What’s new? via Sysdig (author: Nigel Douglas)

抽奖组件 lucky-canvas 素材大全 最近需要用到抽奖组件，网上使用最多的是 lucky-canvas 项目。该项目完成度很高， 有大转盘、九宫格和老虎机三种样式，支持 Web 和微信小程序等多种平台。开发者只需要 做简单配置就能快速实现抽奖功能。官方还提供大量示例供大家参考。但美中不足，可能 是为了避免版权争议，官方的完整示例部分并没有提供源码和素材。而官方文档中的素材又 太素，与实战项目相距甚远。所以我抽空梳理出对应的源码和素材供读者学习参考之用。 via 涛叔

xdp-tutorial-Base via 默

使用 Fluid 对接 S3 存储及性能测试 1. Jindo 挂载 S3 配置环境变量 1 2 3 4 export ENDPOINT=obs.ap-southeast-3.myhuaweicloud.com export BUCKET= export AK= export SK= 创建凭证 1 2 3 4 5 6 7 8 9 10 kubectl apply -f - <&LTEOF apiVersion: v1 kind: Secret metadata: name: mys3secret type: Opaque stringData: fs.s3.accessKeyId: ${AK} fs.s3.accessKeySecret: ${SK} EOF 创建 Dataset 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 kubectl apply -f - <&LTEOF apiVersion: data.fluid.io/v1alpha1 kind: Dataset metadata: name: mys3-jindo spec: mounts: - mountPoint: s3://${BUCKET}/test2/ options: fs.s3.endpoint: ${ENDPOINT} encryptOptions: - name: fs.s3.accessKeyId valueFrom: secretKeyRef: name: mys3secret key: fs.s3.accessKeyId - name: fs.s3.accessKeySecret valueFrom: secretKeyRef: name: via 陈少文的网站 (author: 微信公众号)

再来看 TCP D-SACK 和 Example 5 类似，只是用 SACK 覆盖而不是用 ACK 覆盖，第一个 SACK 块确认第一个重复段，第二个 SACK 块确认它所属的较大块，似乎解决了 Example 5 的问题，但也是碰巧受到了每个 SACK 块必须连续的约束，即便如此，还是丢掉了多个重复段被确认的信息，如果只算一个重复段，可能少了，如果把第二个 SACK 块全部计入，可能就多了。但要明确，1970 年代的当时，TCP 只能是那个样子，不预留空间不是设计没有前瞻性，后续的事更像是生态演化，而非根据详细设计的迭代。 via Netfilter,iptables/OpenVPN/TCP guard:-( (author: dog250)

IPTV直播源格式化工具allinone_format allinone_format 是一个对 youshandefeiyang/allinone 的 /tv.m3u 和 /tptv.m3u 进行聚合 & 重新分组的工具。 via wbsu2004的博客

Optimize and troubleshoot cloud storage at scale with Storage Monitoring via The monitor (author: mahashree.rajendran@datadoghq.com (Mahashree Rajendran))

探索Docker默认网络NAT映射的分配与过滤行为 via Tony Bai (author: bigwhite)

Security for Amazon EKS hybrid nodes: Bridging cloud and on-premises Kubernetes security via Sysdig (author: Eric Carter)

Secure your container images with signature verification The use of version control systems, continuous integration (CI), container services, and other tools in software development have enabled developers to ship code more quickly and efficiently. However, as organizations expand their build and packaging ecosystems, they also increase the number of entry points for malicious code injections that can ultimately make their way to production environments. CI/CD pipelines have privileged permissions and access to downstream container registries, making them a valuable target for attackers—and to make matters worse, many organizations continue to rely on risky long-lived credentials for their CI/CD. via The monitor (author: bowen.chen@datadoghq.com (Bowen Chen))