Hello,
This already works better, but now I get another weird Problem: The
secondary VM would start normally, but then the kernel always complains
that the initrd is corrupt and panics. After that, colo will resync the
VM state, but the Secondary VM then always resets after every RESUME
and will never form a Valid replica. I always copy the colo-disk0
from primary to secondary before every run and freshly create the
active and hidden disks. The Secondary output is below.
What I find interesting is the '"data": {"guest": true}' part. Does
that mean that the Guest is resetting itself (because of data corruption)?
Regards,
Lukas Straub
Secondary Log:
{"timestamp": {"seconds": 1521366904, "microseconds": 684478}, "event":
"RESUME"}
9598@1521366904.684578:colo_vm_state_change Change 'stop' => 'run'
9598@1521366904.684655:colo_send_message Send 'checkpoint-ready' message
9598@1521366924.627341:colo_receive_message Receive 'checkpoint-request' message
{"timestamp": {"seconds": 1521366924, "microseconds": 633321}, "event": "STOP"}
9598@1521366924.647473:colo_vm_state_change Change 'run' => 'stop'
9598@1521366924.647940:colo_send_message Send 'checkpoint-reply' message
9598@1521366924.648569:colo_receive_message Receive 'vmstate-send' message
9598@1521366924.872320:colo_receive_message Receive 'vmstate-size' message
9598@1521366924.872432:colo_send_message Send 'vmstate-received' message
{"timestamp": {"seconds": 1521366924, "microseconds": 945653}, "event":
"RESUME"}
9598@1521366924.945746:colo_vm_state_change Change 'stop' => 'run'
9598@1521366924.945827:colo_send_message Send 'vmstate-loaded' message
{"timestamp": {"seconds": 1521366924, "microseconds": 951073}, "event":
"RESET", "data": {"guest": true}}
{"timestamp": {"seconds": 1521366924, "microseconds": 978172}, "event":
"RESET", "data": {"guest": true}}
9598@1521366944.627094:colo_receive_message Receive 'checkpoint-request' message
{"timestamp": {"seconds": 1521366944, "microseconds": 627314}, "event": "STOP"}
9598@1521366944.660195:colo_vm_state_change Change 'run' => 'stop'
9598@1521366944.660304:colo_send_message Send 'checkpoint-reply' message
9598@1521366944.662097:colo_receive_message Receive 'vmstate-send' message
9598@1521366944.803816:colo_receive_message Receive 'vmstate-size' message
9598@1521366944.803983:colo_send_message Send 'vmstate-received' message
{"timestamp": {"seconds": 1521366944, "microseconds": 871406}, "event":
"RESUME"}
9598@1521366944.871492:colo_vm_state_change Change 'stop' => 'run'
9598@1521366944.871719:colo_send_message Send 'vmstate-loaded' message
{"timestamp": {"seconds": 1521366944, "microseconds": 877710}, "event":
"RESET", "data": {"guest": true}}
{"timestamp": {"seconds": 1521366944, "microseconds": 898549}, "event":
"RESET", "data": {"guest": true}}
<here I kill the Primary but the secondary is running the BIOS after the reset,
so it can't take over>
qemu-system-x86_64: Can't receive COLO message: Input/output error
{"timestamp": {"seconds": 1521366949, "microseconds": 718227}, "event":
"COLO_EXIT", "data": {"mode": "secondary", "reason": "error"}}
qemu-system-x86_64: Unable to connect character device red0: Failed to connect
socket: Connection refused
^Cqemu-system-x86_64: terminating on signal 2
{"timestamp": {"seconds": 1521366951, "microseconds": 423301}, "event":
"SHUTDOWN", "data": {"guest": false}}
On Sat, 17 Mar 2018 22:22:01 +0000
Zhang Chen <zhangc...@gmail.com> wrote:
> Yes, you can see the Colo wiki, I just updated it.
> https://wiki.qemu.org/Features/COLO
>
> Thanks
> Zhang Chen
>
