Sabitlenmiş Tweet
Ankit ☁︎ ❯
733 posts

Ankit ☁︎ ❯
@ankit_ops2799
Cloud & DevOps Engineer 🛠️ | AWS • Azure • Docker • Kubernetes ☸️ Building & Automating Scalable Infrastructure 🏗️
India Katılım Eylül 2011
2K Takip Edilen158 Takipçiler

Docker Image vs Container
• Images are immutable blueprints
• Containers are disposable runtime instances
• Fix the Dockerfile, not the container
This mindset is what makes deployments reproducible.
#Docker #DevOps #Cloud #Kubernetes

English

@suni_code Because compression costs time and CPU, and data is usually written for speed first not optimization.
English

@javarevisited Because container shares the host OS kernel, so it avoids full VM overhead but it still carries everything your app needs to run.
container itself is lightweight but the image can still be huge.
English

@CyberRacheal Quick Format deletes the index, not the data itself.
The files stay on the USB until overwritten.
English

Interviewer:
A user performs a 'Quick Format' on their 32 GB USB drive, thinking they have wiped all their data.
However, a technician is later able to restore 100% of those files.
Where exactly was that data stored during the period the drive appeared empty, and why did the formatting process fail to remove it? 🤔
English

@CyberRacheal C. 10.0.0.50
It belongs to the 10.0.0.0/8 private range, which supports millions of internal hosts.
English

@Sarthak4Alpha Pause autoscaling, stop failing pods, then inspect pod events/logs.
Since no code changed, there can be possibilities of infra issues like bad probes, expired secrets, dependency outages, or resource exhaustion.
English

The correct answer is:
2️⃣ Memory (RAM)
Why?
CPU is compressible → Kubernetes/Linux can throttle CPU usage, slowing containers down while they continue running.
Memory is incompressible → Once RAM is exhausted, the system cannot throttle it safely.
If a Pod exceeds memory limits or the node runs out of RAM:
Kubernetes may trigger an OOMKilled
The Pod can be evicted immediately
CPU → Throttled under pressure
Memory → Hard limit, causes eviction/OOM
Disk I/O → Can slow down workloads
Network bandwidth → Causes latency/congestion
In short:
CPU pressure slows Pods down.
Memory exhaustion kills them
Ankit ☁︎ ❯@ankit_ops2799
When your Ubuntu worker nodes are under heavy CPU load, the Linux kernel can "compress" the usage, slowing down the pods but allowing them to keep running. Which resource, when fully exhausted, is incompressible, forcing Kubernetes to instantly evict the pod? 1️⃣ CPU Cycles 2️⃣ Memory (RAM) 3️⃣ Disk I/O (IOPS) 4️⃣ Network Bandwidth Test your #K8s knowledge! #Kubernetes #SRE #DevOps
English

@uday_devops COPY → used for most file transfers (recommended)
ADD → used only when you need auto-extraction or URL fetching
Use COPY unless you specifically need ADD features.
English































