Quantcast
Channel: Debian User Forums
Viewing all articles
Browse latest Browse all 3567

General Questions • how can shutdown hang in md_notify_reboot?

$
0
0
Hi.

I have a (quite large) script that basically creates a fresh qemu VM, with fresh virtual disks etc, sets up things like a partition table (incl. a raid1 setup via mdadm) and debootstraps it. Beyond the inherent complexity of that task, it doesn't do anything particularly magical. It's mostly straight-forward. I think the script itself doesn't matter, but note that I constantly create fresh VMs with an (always fresh) mdadm raid1 inside, and use them only for short times.

Every now and then (around 10% of times?!), when I shutdown or reboot such a VM, it just hangs in a very late step, showing two pointless lines of generic mdadm output or similar and a screen distortion. CPU usage is constantly at 100% then. I could leave it for many hours, it doesn't change.

Today it seems I had luck as I got some actual output:
Image

But what does it mean? Is it maybe confused because the raid was not fully synced yet? But that should not prevent shutdown in this way, right?

At the moment I'm debootstrapping Debian trixie VMs, but I had the same issue with bookworm as well.

Statistics: Posted by pino83 — 2025-02-18 23:08 — Replies 0 — Views 20



Viewing all articles
Browse latest Browse all 3567

Trending Articles