diff options
Diffstat (limited to 'results/classifier/deepseek-2-tmp/output/manual-review/2482')
| -rw-r--r-- | results/classifier/deepseek-2-tmp/output/manual-review/2482 | 137 |
1 files changed, 0 insertions, 137 deletions
diff --git a/results/classifier/deepseek-2-tmp/output/manual-review/2482 b/results/classifier/deepseek-2-tmp/output/manual-review/2482 deleted file mode 100644 index fb427300d..000000000 --- a/results/classifier/deepseek-2-tmp/output/manual-review/2482 +++ /dev/null @@ -1,137 +0,0 @@ - -qemu-system-x86_64: Live Migration fails with BLOCK_JOB_ERROR -Description of problem: -After disk migration is completed and RAM migration is being performed, migration status switches to 'pre-switchover'. -In the 'pre-switchover' migration state, block jobs status is still set to 'ready' instead of 'running' -on queried for block job status when 'offset' and 'length' diverged. Thus, It results in BLOCK_JOB_ERROR. -Steps to reproduce: -On source host -1. Add disk(s) that needed to be migrated by issuing 'blockdev-add' QMP command. -2. start blockdev-mirror operations to perform disk(s) transfer by issuing QMP command -3. start RAM migration. (send HMP commands - listed below -4. Migration status changed to 'pre-switchover'. While in 'pre-switchover', check for disk activity - -While RAM migration is happening, Migration status is changed to 'pre-switchover' -and observe that block jobs 'offset' and 'length' diverged. But, block job status is still set to 'ready' instead of 'running'. - -On destination host -1. Launch the VM in listening mode (-incoming) for migrations -2. start NBD server -3. add disks to NBD server. -4. set migration parameters by sending HMP commands -Additional information: -# On SOURCE Host, start all blockdev-add operations -# Issue QMP commands (blockdev-add) for all block devices ("drive-scsi-disk-0" and "drive-scsi-disk-1") of VM - -``` - { - "execute" => "blockdev-add", - "arguments" => { - "driver" => "raw", - "node-name" => "node_drive-scsi-disk-0", - "auto-read-only" => false, - "read-only" => false, - "file" => { - "driver" => "nbd", - "export" => "drive-scsi-disk-0", - "server" => { - "type" => "inet", - "host" => "2600:3c0f:17:14::21", - "port" => "37552", - }, - "tls-creds" => "tlscreds0" - } - } - } -``` - - { - "execute" => "blockdev-add", - "arguments" => { - "driver" => "raw", - "node-name" => "node_drive-scsi-disk-1", - "auto-read-only" => false, - "read-only" => false, - "file" => { - "driver" => "nbd", - "export" => "drive-scsi-disk-1", - "server" => { - "type" => "inet", - "host" => "2600:3c0f:17:14::21", - "port" => "37552", - }, - "tls-creds" => "tlscreds0" - } - } - } - -# On SOURCE Host, start all blockdev-mirror operations to start disk transfer -# i.e Issue QMP commands (blockdev-mirror) for each of those block devices ("drive-scsi-disk-0" and "drive-scsi-disk-1") - -``` - { - "execute" => "blockdev-mirror", - "arguments" => { - "device" => "drive-scsi-disk0", - "target" => "node_drive-scsi-disk-0", - "speed" => 100000000, - "sync" => "full", - } - } -``` - -``` - { - "execute" => "blockdev-mirror", - "arguments" => { - "device" => "drive-scsi-disk1", - "target" => "node_drive-scsi-disk-1", - "speed" => 100000000, - "sync" => "full", - } - } -``` - -# NBD server configuration on destination host by issuing QMP command -# Start NBD server -``` - { - "execute" => "nbd-server-start", - "arguments" => { - "addr" => { - "type" => "inet", - "data" => { - "host" => "2600:3c0f:17:14::21", - "port" => "37552" - } - }, - "tls-creds" => "tlscreds0" - } - } -``` - -# On DESTINATION Host -# Register incoming disks(2) with NBD server by issuing QMP commands to VM on the destination host -# Disk# 1 -``` - { - "execute" => "nbd-server-add", - "arguments" => { - "device" => "drive-scsi-disk0", - "writable" => true - } - } -``` -# Disk# 2 -``` - { - "execute" => "nbd-server-add", - "arguments" => { - "device" => "drive-scsi-disk1", - "writable" => true - } - } -``` - -# Wait for disks to finish the bulk of the data migration. -# |