---
title: "Comment by StackShortlist on \"Why a Deleted Linux File Can Still Consume Disk Space\" | Hashnode"
description: "The 52GB debug.log example is a very realistic way to show why rm returning instantly means nothing while a process still holds the file descriptor. The df vers"
url: "https://hashnode.com/posts/why-a-deleted-linux-file-can-still-consume-disk-space/6ad24b1a042b56ac7bf5afd5/comment/6ad2b877a305ffe6691885c7"
---

[Discussion](https://hashnode.com/posts/why-a-deleted-linux-file-can-still-consume-disk-space/6ad24b1a042b56ac7bf5afd5)

Comment [Why a Deleted Linux File Can Still Consume Disk Space](https://hashnode.com/posts/why-a-deleted-linux-file-can-still-consume-disk-space/6ad24b1a042b56ac7bf5afd5)

[StackShortlist](https://hashnode.com/@veristria)

[Oct 16](https://hashnode.com/posts/why-a-deleted-linux-file-can-still-consume-disk-space/6ad24b1a042b56ac7bf5afd5/comment/6ad2b877a305ffe6691885c7)

The 52GB debug.log example is a very realistic way to show why rm returning instantly means nothing while a process still holds the file descriptor. The df versus du mismatch is usually the first clue, so having that next to the triage playbook is handy. The O\_APPEND caveat is a good addition too, since truncating through the open descriptor behaves differently depending on how the writer opened the file.
