Something I think my network might be interested: 1. Run etcd at least on version 3.5.10 - Your k8s apiserver watch cache might stay stale for minutes if not hours when there are huge read or write - https://lnkd.in/gsMxruF9 2. Overlayfs in kernel 5.4 very likely has a bug (still trying to identify where the bug is) - https://lnkd.in/g7r2GMR9
Haibing Zhou’s Post
More Relevant Posts
-
Glad this got fixed :p "When FreeBSD switched from using ioctl(2) to using Netlink sockets to configure network interfaces, the logic for determining the default mask in cases where one was not explicitly provided was inadvertantly inverted, resulting in class A addresses getting a prefix size of 24 instead of 8, and vice versa for class C addresses. Class B addresses were not affected." https://lnkd.in/er_tCgbp
To view or add a comment, sign in
-
Did you know ? I came across this one liner today, this is about how to test if the port on a remote host is reachable or not especially when you do not have the light weight utilities such as curl, wget and telnet in containers. </dev/tcp/<hostname>/port-number && echo port is open || echo port is closed It’s very interesting to learn such useful commands and also amazing to see that bash has been able to access TDP and UDP.
To view or add a comment, sign in
-
Raw dogging HTTP server yourself is a fun exercise, especially in low level language like #rust(BTW). and even though i completed all the stages, I know just how many bugs are there related to buffer overflow and so many heap allocations which most likely could be saved 😅 😂 . CodeCrafters.io (YC S22) Project Link: https://lnkd.in/d8kEWJxN My (bad 😅 but accepted 😂 ) Solution(which i plan to fix iterativey): https://lnkd.in/dVNN3PKc PS: This shall be re-written in zig once all the bugs and performance issues are fixed.
To view or add a comment, sign in
-
Do you know there is a simple builtin node.js command, it does same task as nodemon do? You don't have to install nodemon package, the command automatically restarting the node application when file changes in the directory are detected.
To view or add a comment, sign in
-
RPI K3 Cluster with Cloudflare Tunnels So - I have a K3 Cluster - 4-8GB Raspberry Pi 4s running 64bit PI OS on micro SD cards. Recently I had a failure of a worker - no big deal as I don't have much traffic - but then I had failure of my Master which is a problem -) btw use SSDs instead of micro SD cards - only had one SSD fail in many years. At this point I really had wished I wrote down the steps to put this back together but had not... Did it yesterday - here are the steps if interested. Note I used Cloudflare Tunnels to point at my cluster without opening a local Port. https://lnkd.in/g7DyzUXD
K3Cluster.md
gist.github.com
To view or add a comment, sign in
-
Full Stack Developer with a huge love for DSA, PHP, JavaScript, Python, Alpine.js, Vue.js, React, Laravel, Tailwind, Headless Systems, API, MySQL, WordPress, and System Architecture.
This might be the last npm command you’ll ever need. From now on, switch to bun and bunx instead of npm and npx, and experience lightning-fast speeds.
To view or add a comment, sign in
-
Curious about tracing in distributed systems? Join our livestream on April 11 at 2pm PDT to learn about distributed tracing in NextJs. We'll cover general tracing concepts as well as how Highlight's SDKs make it easy to configure. Register: https://lu.ma/b0uz0fiz
To view or add a comment, sign in
-
Building L7 Load balancer from scratch in Go 🚦 I once heard that student majoring computer science should built their own operating system... I started with the idea of making an LB, if not an operating system, at least an LB. Building an LB required some knowledge about various domains including several LB algorithm(RR, Ring hash, Maglev, etc), Go standard net/http library, Concurrency(atomic operations, RW mutexes), and so on. My first goal was to support Yorkie backends, which currently managed by Istio/Envoy LB with resolving so called split-brain issue of long-lived connections(gRPC, Websockets) when backend changes on ring hash based LB. I have now built an LB with maglev consistent hashing algorithm and mechanism to forcefully close connection on split-brain issue. The main idea was to store connection with key-hash value, then recalculate them on backend changes. This is the repo for my LB: https://lnkd.in/g-RvwygV #loadbalancer #Go #Maglev
GitHub - krapie/plumber: Plumber is a L7 load balancer from scratch in Go.
github.com
To view or add a comment, sign in