when I use kvm and spice to run vm(win7 32bit with spice guest tool),use spice client to connet to vm and then I use libvrit to migrate the vm to another host,qemu crash in libspiceserver,the callback is that: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7faf9ffff700 (LWP 9974)] 0x00007fb0456143a0 in qemu_del_timer () (gdb) bt #0 0x00007fb0456143a0 in qemu_del_timer () #1 0x00007fb040e919f1 in spice_server_vm_stop (s=<optimized out>) at reds.c:3810 #2 0x00007fb045631934 in vm_change_state_handler () #3 0x00007fb04564bceb in vm_state_notify () #4 0x00007fb0456516f2 in vm_stop () #5 0x00007fb0455e91b9 in migration_thread () #6 0x00007fb0434aadf3 in start_thread () from /lib64/libpthread.so.0 #7 0x00007fb0401a63dd in clone () from /lib64/libc.so.6 dose anybody konw about this? thanks very much!
Please, add information about versions of libvirtd, qemu and spice-server of both hosts.
the version information： virsh # version Compiled against library: libvirt 1.2.7 Using library: libvirt 1.2.7 Using API: QEMU 1.2.7 Running hypervisor: QEMU 1.5.3 spice-server-0.12.4-5.el7.x86_64 qemu-kvm-common-1.5.3-60.el7.x86_64 libvirt-daemon-kvm-1.2.7-1.el7.centos.x86_64 qemu-kvm-1.5.3-60.el7.x86_64 simple debug info: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7f0d63cfe700 (LWP 29780)] 0x00007f0e0b3f06c0 in timer_cancel () (gdb) bt #0 0x00007f0e0b3f06c0 in timer_cancel () #1 0x00007f0e06c56841 in spice_server_vm_stop (s=<optimized out>) at reds.c:4535 #2 0x00007f0e0b3f0934 in vm_change_state_handler () #3 0x00007f0e0b40aceb in vm_state_notify () #4 0x00007f0e0b4106f2 in vm_stop () #5 0x00007f0e0b3a81b9 in migration_thread () #6 0x00007f0e09269df3 in start_thread () from /lib64/libpthread.so.0 #7 0x00007f0e05f6c3dd in clone () from /lib64/libc.so.6 #1 0x00007f0e06c56841 in spice_server_vm_stop (s=<optimized out>) at reds.c:4535 warning: Source file is more recent than executable. 4535 spice_char_device_stop(st_item->st); (gdb) l 4530 reds->vm_running = FALSE; 4531 RING_FOREACH(item, &reds->char_devs_states) { 4532 SpiceCharDeviceStateItem *st_item; 4533 4534 st_item = SPICE_CONTAINEROF(item, SpiceCharDeviceStateItem, link); 4535 spice_char_device_stop(st_item->st); 4536 } 4537 red_dispatcher_on_vm_stop(); 4538 } 4539 (gdb) p *st_item $2 = {link = {prev = 0x7f0e0daee410, next = 0x7f0e0d8e5020}, st = 0x7f0e0da04580} (gdb) p *st_item->st $3 = {running = 0, active = 0, wait_for_migrate_data = 0, refs = 0, write_queue = { prev = 0x0, next = 0x0}, write_bufs_pool = {prev = 0x0, next = 0x0}, cur_write_buf = 0x0, cur_write_buf_pos = 0x0, write_to_dev_timer = 0x0, num_self_tokens = 0, clients = {prev = 0x0, next = 0x0}, num_clients = 0, client_tokens_interval = 0, sin = 0x0, during_read_from_device = 0, cbs = { read_one_msg_from_device = 0x0, ref_msg_to_client = 0x0, unref_msg_to_client = 0x0, send_msg_to_client = 0x0, send_tokens_to_client = 0x0, on_free_self_token = 0x0, remove_client = 0x0}, opaque = 0x0} (gdb) it doesn't crash all time,about 3-5 times try meet 1 crash,hung the vm may meet the same crash also,but need more try.would you please take this situation ，thanks!
Do you do anything special during the migration - what are the steps to reproduce it? I haven't encountered this issue (I am using the same versions - except QEMU 2.0.0 as the hypervisor).
imbadyc, please add virsh dumpxml / qemu command line of the vm.
I'm closing this due inactivity: NEEDINFO since 2015-03-06
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.