Why Disk Space Stays Full After Deleting a Large Log File on Linux

·4 views

A server ran out of disk space even after its largest log file was deleted, because df still showed the filesystem as full. On Linux, removing a file only drops its directory name, and the storage is freed only once no names and no running processes still reference it. An application that had the log open keeps its descriptor valid, so the data continues to occupy space and may even grow. Tools behave differently here: du cannot see unlinked files, while df counts allocated blocks, and lsof +L1 can reveal which process is holding the deleted file. The recommended fix is to have that application reopen or reload its log, or restart the service, with truncating via /proc only as a risky emergency option.

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)

Microsoft Offers Voluntary Buyouts to About a Quarter of Its Workforce

Microsoft has reportedly extended voluntary buyout offers to roughly a quarter of its employees. The source is a KUOW report on the company's workforce move. The article linked from Hacker News, which had 20 points and no comments at the time of capture, provides the details of the offer. Employees who accept the buyouts would leave the company in exchange for financial compensation. The full terms and reasoning behind the offer are not available in the summary provided.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

Developer publishes AKS Kubernetes manifests to GitHub and removes Azure lab resources

A developer completed the final stage of a Kubernetes on Azure Kubernetes Service lab by committing nine manifest files covering pods, deployments, services, ConfigMaps, StatefulSets, DaemonSets, jobs and CronJobs. The files were pushed to a public GitHub repository to preserve the implementation as portfolio evidence. The developer then verified namespace removal, deleted the Azure resource group, confirmed the AKS-managed resource group was gone, and removed the obsolete Kubernetes context from the local kubeconfig. The steps were taken to ensure the temporary cloud environment stopped incurring charges while keeping the configuration available for review.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Cleaning Up WordPress XML Sitemaps to Diagnose Google Indexation Problems

A common assumption is that stalled indexing on small WordPress sites stems from crawl budget or soft 404s, but the XML sitemap can be a quieter source of trouble. Noisy sitemaps containing redirects, noindex pages, or unreliable lastmod dates can leave 'Discovered - currently not indexed' counts growing even when coverage looks healthy. The author outlines a triage approach: sitemaps should list only final, canonical, indexable URLs that return HTTP 200, since they serve as a discovery hint rather than a way to force indexing. Guidance also covers which page types to exclude and why lastmod should reflect meaningful content changes rather than being bumped artificially. The piece stresses that sitemap cleanup helps most with certain Search Console statuses and rarely resolves others on its own.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Three Tricky .NET 11 Scenarios Test EF Core 11 and ASP.NET Core 11 Knowledge

A developer has published a 20-question practice quiz to help teams check their code for subtle breaking changes in EF Core 11 and ASP.NET Core 11 before the November 2026 general availability release. The article previews three real-world scenarios, covering null vector columns in a RAG re-ranking service, a Managed Identity connection failure in Azure staging, and validation not applying across assemblies in Minimal API. Readers are invited to answer multiple-choice questions in the comments before checking the answer key. The full quiz also includes 17 additional scenarios, such as complex types on TPT/TPC hierarchies.

0 comments Read more at DEV Community