---
title: "Comment by Kamran Khan on \"PHP2Shell: A Developer’s Guide to Patching CVE-2026-109618 and CVE-2026-109619\" | Hashnode"
description: "One thing that bit me on a patch like this: reloading php-fpm is not the same as restarting it. A reload lets old workers finish their current requests, so a lo"
url: "https://hashnode.com/posts/php2shell-a-developer-s-guide-to-patching-cve-2026-109618-and-cve-2026-109619/6ada86f0f34b8ed726c54f8c/comment/6adb5dc80e2844ea3a4a49b9"
---

[Discussion](https://hashnode.com/posts/php2shell-a-developer-s-guide-to-patching-cve-2026-109618-and-cve-2026-109619/6ada86f0f34b8ed726c54f8c)

Comment [PHP2Shell: A Developer’s Guide to Patching CVE-2026-109618 and CVE-2026-109619](https://hashnode.com/posts/php2shell-a-developer-s-guide-to-patching-cve-2026-109618-and-cve-2026-109619/6ada86f0f34b8ed726c54f8c)

[Kamran Khan](https://hashnode.com/@kamrankhandev)

[Oct 23](https://hashnode.com/posts/php2shell-a-developer-s-guide-to-patching-cve-2026-109618-and-cve-2026-109619/6ada86f0f34b8ed726c54f8c/comment/6adb5dc80e2844ea3a4a49b9)

One thing that bit me on a patch like this: reloading php-fpm is not the same as restarting it. A reload lets old workers finish their current requests, so a long lived upload can keep a vulnerable worker serving for minutes after you've called the box patched. For these I restart outright now, and accept the minute of dropped connections as the price.

[webmix studio](https://hashnode.com/@webmixstudio)

[Nov 2](https://hashnode.com/posts/php2shell-a-developer-s-guide-to-patching-cve-2026-109618-and-cve-2026-109619/6ada86f0f34b8ed726c54f8c/comment/6adb5dc80e2844ea3a4a49b9#reply-6ae7d38543096bd03f3bf886)

Good point! We agree that a full restart is the safer choice for critical security patches. Making sure old workers are completely replaced is just as important as updating the PHP runtime itself. Better to accept a brief interruption than leave vulnerable processes running.
