How to Prevent Split-Brain and Configure Quorum in Proxmox HA Clusters

·6 views

Proxmox VE high-availability clusters face a critical risk called split-brain, where a loss of communication between nodes causes each node to independently assume control, potentially running duplicate VMs and corrupting data. The quorum mechanism addresses this by requiring a majority of nodes to agree before any cluster decisions are made, with a minimum of two votes needed in a three-node setup. Administrators can use the pvecm command-line tool to check cluster status, set expected node counts, and add a dedicated witness server that acts as a neutral tiebreaker without hosting any virtual machines. Common pitfalls include network latency causing false node failures, misconfigured quorum expectations, and relying on a single witness node that itself becomes a point of failure. Best practices recommend using a dedicated cluster network, regularly verifying quorum settings with pvecm status, and deploying two witness servers for added redundancy.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Codex benchmark tests when multi-agent AI engineering actually beats solo work

A new dependency-free benchmark published in Codex How To evaluates whether splitting a software task across multiple AI agents genuinely outperforms a single-agent approach. The test uses an incident-response application with two clearly separated code surfaces — backend persistence and a browser client — assigned to distinct agents under a controller. In an initial smoke run, both single-agent and orchestrated setups passed all supplied tests and external evaluator checks, though the multi-agent run additionally produced live browser interaction evidence. The benchmark concludes that parallel agents are only justified when tasks have frozen interfaces, exclusive write paths, and coordination costs that are measurably outweighed by gains. The framework is designed to be rerun repeatedly so teams can build evidence before committing to an orchestrated workflow.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Launches Curated Directory to Help Users Find n8n Community Nodes

A developer named Bhav Shekhaliya has created 'Awesome n8n Community Nodes', an open-source curated directory hosted on GitHub to simplify the discovery of community-built integrations for the n8n workflow automation platform. The project addresses a growing discoverability problem, as thousands of community nodes already exist on npm but are difficult to find without knowing exactly what to search for. The directory organizes nodes into practical categories such as AI and LLMs, browser automation, communication tools, databases, and CRM integrations. Each entry links to its corresponding npm package and, where available, its public source repository. The maintainer cautions that the directory is not a security endorsement, and users should review source code, dependencies, and licenses before installing any community node.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

US BankChain Alliance and Zerohash Push Blockchain Into Mainstream Finance

The digital asset market is undergoing a major shift as blockchain technology moves from a disruptive force to an integrated part of mainstream finance. Backed by 39 state banking associations, the BankChain Alliance was announced with a 2027 launch target to develop stablecoins, payments, and tokenized deposits within regulated banking frameworks. Simultaneously, crypto infrastructure firm Zerohash is pursuing a national trust bank charter from the OCC, reflecting the crypto industry's push for regulatory legitimacy. These developments highlight a dual trend: traditional financial institutions adopting blockchain to improve efficiency, while crypto firms seek compliance within established legal frameworks. Together, they signal the emergence of a hybrid financial architecture built on distributed ledger technology.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Mac Studio M5 Ultra vs NVIDIA DGX Spark: Which $5,500 Local AI Box Wins?

Apple's newly announced Mac Studio M5 Ultra and NVIDIA's DGX Spark are two competing desktop AI machines priced around $5,500, both designed for running large language models locally without relying on cloud services. The M5 Ultra offers up to 512GB of unified memory with 1.2TB/s bandwidth, giving it roughly 4.4 times the memory throughput of the DGX Spark's 273GB/s — a significant advantage for LLM inference speed. The DGX Spark, priced at $4,699 and available now with 128GB of memory, counters with full CUDA ecosystem support including TensorRT-LLM, vLLM, and NVIDIA's fine-tuning toolchain. For raw token generation on large models, the Mac Studio holds the edge, while the DGX Spark is better suited for fine-tuning workflows and CUDA-dependent software stacks. The two machines ultimately target slightly different users, with Apple prioritizing massive model capacity and NVIDIA focusing on developer ecosystem depth.

0 comments Read more at DEV Community