all lists on lists.proxmox.com
 help / color / mirror / Atom feed
From: Fiona Ebner <f.ebner@proxmox.com>
To: Erik Fastermann <e.fastermann@proxmox.com>, pve-devel@lists.proxmox.com
Subject: Re: [RFC PATCH pve-qemu] savevm-async: fix stuck paused vm after snapshot
Date: Wed, 26 Aug 2026 15:30:46 +0200	[thread overview]
Message-ID: <d052f289-1c33-4db7-814b-aa4664446666@proxmox.com> (raw)
In-Reply-To: <20260612075345.110118-1-e.fastermann@proxmox.com>

Am 12.06.26 um 9:53 AM schrieb Erik Fastermann:
> When creating a snapshot of a paused vm, the vm hung indefinitely in the
> state "paused (finish-migrate)" and had to be restarted. This patch
> fixes this by only transitioning the vm run state if the vm was
> currently running.
> 
> Signed-off-by: Erik Fastermann <e.fastermann@proxmox.com>
> ---
>  ...c-fix-stuck-paused-vm-after-snapshot.patch | 48 +++++++++++++++++++
>  debian/patches/series                         |  1 +
>  2 files changed, 49 insertions(+)
>  create mode 100644 debian/patches/pve/0047-savevm-async-fix-stuck-paused-vm-after-snapshot.patch
> 
> diff --git a/debian/patches/pve/0047-savevm-async-fix-stuck-paused-vm-after-snapshot.patch b/debian/patches/pve/0047-savevm-async-fix-stuck-paused-vm-after-snapshot.patch
> new file mode 100644
> index 0000000..46f0422
> --- /dev/null
> +++ b/debian/patches/pve/0047-savevm-async-fix-stuck-paused-vm-after-snapshot.patch
> @@ -0,0 +1,48 @@
> +From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
> +From: Erik Fastermann <e.fastermann@proxmox.com>
> +Date: Thu, 11 Jun 2026 13:42:05 +0200
> +Subject: [PATCH] savevm-async: fix stuck paused vm after snapshot
> +
> +When creating a snapshot of a paused vm, the vm hung indefinitely in the
> +state "paused (finish-migrate)" and had to be restarted. This patch
> +fixes this by only transitioning the vm run state if the vm was
> +currently running.
> +
> +Signed-off-by: Erik Fastermann <e.fastermann@proxmox.com>
> +---
> + migration/savevm-async.c | 12 ++++++++----
> + 1 file changed, 8 insertions(+), 4 deletions(-)
> +
> +diff --git a/migration/savevm-async.c b/migration/savevm-async.c
> +index acd1a4de6e..153433f539 100644
> +--- a/migration/savevm-async.c
> ++++ b/migration/savevm-async.c
> +@@ -181,9 +181,11 @@ static void process_savevm_finalize(void *opaque)
> +     blk_set_aio_context(snap_state.target, qemu_get_aio_context(), NULL);
> + 
> +     snap_state.vm_needs_start = runstate_is_running();
> +-    ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
> +-    if (ret < 0) {
> +-        save_snapshot_error("vm_stop_force_state error %d", ret);
> ++    if (snap_state.vm_needs_start) {
> ++        ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);

It feels wrong to me to avoid going into the finish-migrate state, since
we do things like qemu_savevm_state_complete_precopy(),
qemu_savevm_state_cleanup() and save_snapshot_cleanup() below.

In migration/migration.c there is

        if (runstate_is_live(s->vm_old_state)) {
            if (!runstate_check(RUN_STATE_SHUTDOWN)) {
                vm_start();
            }
        } else {
            if (runstate_check(RUN_STATE_FINISH_MIGRATE)) {
                runstate_set(s->vm_old_state);
            }
        }

The old state is saved via runstate_get() in migration_stop_vm() before
stopping the VM. I think we should do the same.


> ++        if (ret < 0) {
> ++            save_snapshot_error("vm_stop_force_state error %d", ret);
> ++        }
> +     }
> + 
> +     if (!aborted) {
> +@@ -370,7 +372,9 @@ void qmp_savevm_start(const char *statefile, Error **errp)
> + 
> +     if (!statefile) {
> +         snap_state.vm_needs_start = runstate_is_running();

And for us, the place to store the old state is here.

> +-        vm_stop(RUN_STATE_SAVE_VM);
> ++        if (snap_state.vm_needs_start) {
> ++            vm_stop(RUN_STATE_SAVE_VM);
> ++        }
> +         snap_state.state = SAVE_STATE_COMPLETED;
> +         return;
> +     }
> +-- 
> +2.47.3
> +
> diff --git a/debian/patches/series b/debian/patches/series
> index 84c0664..6463a4a 100644
> --- a/debian/patches/series
> +++ b/debian/patches/series
> @@ -70,3 +70,4 @@ pve/0043-PVE-backup-get-device-info-allow-caller-to-specify-f.patch
>  pve/0044-PVE-backup-implement-backup-access-setup-and-teardow.patch
>  pve/0045-PVE-backup-prepare-for-the-switch-to-using-blockdev-.patch
>  pve/0046-savevm-async-reuse-migration-blocker-check-for-snaps.patch
> +pve/0047-savevm-async-fix-stuck-paused-vm-after-snapshot.patch





      reply	other threads:[~2026-08-26 13:30 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-12  7:53 [RFC PATCH pve-qemu] savevm-async: fix stuck paused vm after snapshot Erik Fastermann
2026-08-26 13:30 ` Fiona Ebner [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=d052f289-1c33-4db7-814b-aa4664446666@proxmox.com \
    --to=f.ebner@proxmox.com \
    --cc=e.fastermann@proxmox.com \
    --cc=pve-devel@lists.proxmox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal