582e7872e6
- qemu_migration: Refactor qemuMigrationSrcRestoreDomainState (RHEL-1071) - qemu_migration: Do not automatically resume domain after I/O error (RHEL-1071) - qemucapabilitiestest: Add data for the qemu-10.0 dev cycle on x86_64 (RHEL-78398) - qemucapabilitiestest: Update 'caps_10.0.0_x86_64' to 'v9.2.0-1636-gffaf7f0376' (RHEL-78398) - qemu: capabilies: Introduce QEMU_CAPS_BLOCKDEV_SET_ACTIVE (RHEL-78398) - qemu: monitor: Add monitor backend for 'blockdev-set-active' (RHEL-78398) - qemu: migration: Reactivate block nodes after migration if VM is left paused (RHEL-78398) - conf: change virDomainHostdevInsert() to return void (RHEL-77086) - qemu: fix qemu validation to forbid guest-side IP address for type='vdpa' (RHEL-77086) - qemu: validate that model is virtio for vhostuser and vdpa interfaces in the same place (RHEL-77086) - qemu: automatically set model type='virtio' for interface type='vhostuser' (RHEL-77086) - qemu: do all vhostuser attribute validation in qemu driver (RHEL-77086) - conf/qemu: make <source> element *almost* optional for type=vhostuser (RHEL-77086) - qemu: use switch instead of if in qemuProcessPrepareDomainNetwork() (RHEL-77086) - qemu: make qemuPasstCreateSocketPath() public (RHEL-77086) - qemu: complete vhostuser + passt support (RHEL-77086) - qemu: fail validation if a domain def has vhostuser/passt but no shared mem (RHEL-77086) - docs: improve type='user' docs to higlight differences between SLIRP and passt (RHEL-77086) - docs: document using passt backend with <interface type='vhostuser'> (RHEL-77086) - utils: Canonicalize paths before comparing them (RHEL-79165) Resolves: RHEL-1071, RHEL-77086, RHEL-78398, RHEL-79165
39 lines
1.4 KiB
Diff
39 lines
1.4 KiB
Diff
From 33c6c5a2061321e75fa1455a5ad9fa192b1a7825 Mon Sep 17 00:00:00 2001
|
|
Message-ID: <33c6c5a2061321e75fa1455a5ad9fa192b1a7825.1739815540.git.jdenemar@redhat.com>
|
|
From: Jiri Denemark <jdenemar@redhat.com>
|
|
Date: Wed, 15 Jan 2025 16:02:00 +0100
|
|
Subject: [PATCH] qemu_migration: Do not automatically resume domain after I/O
|
|
error
|
|
|
|
When an I/O error happens (causing a domain to be paused) during live
|
|
migration which is later cancelled by a user, trying to resume the
|
|
domain doesn't make sense.
|
|
|
|
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
|
|
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
|
|
(cherry picked from commit 3766891c5954c53370d7eb9c7a0a8e6ba3854ca4)
|
|
|
|
https://issues.redhat.com/browse/RHEL-1071
|
|
---
|
|
src/qemu/qemu_migration.c | 5 +++++
|
|
1 file changed, 5 insertions(+)
|
|
|
|
diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c
|
|
index fe0b9fc672..4112e95a1e 100644
|
|
--- a/src/qemu/qemu_migration.c
|
|
+++ b/src/qemu/qemu_migration.c
|
|
@@ -241,6 +241,11 @@ qemuMigrationSrcRestoreDomainState(virQEMUDriver *driver, virDomainObj *vm)
|
|
reason == VIR_DOMAIN_PAUSED_POSTCOPY_FAILED)
|
|
return;
|
|
|
|
+ if (reason == VIR_DOMAIN_PAUSED_IOERROR) {
|
|
+ VIR_DEBUG("Domain is paused due to I/O error, skipping resume");
|
|
+ return;
|
|
+ }
|
|
+
|
|
VIR_DEBUG("Restoring pre-migration state due to migration error");
|
|
|
|
/* we got here through some sort of failure; start the domain again */
|
|
--
|
|
2.48.1
|