[Xen-devel] I have killed hypervisor. wow

[Alexander Benaguev <abenaguev@xxxxxxxxx>]

Greetings

all my pv domains have failed today after reboot (hvm domains are ok). I 
found out that virtual block devices just stopped mounting to guests. 
Then I tried to destroy failed domains and got a lot of messages to 
dmesg like:
blk_tap: Bad number of segments in request (179)
#snip 1511 messages
blk_tap: unknown operation [152]
#snip 1526 messages

Then dom0 froze dead producing the above output only. After the reboot 
everything was the same except dom0 kept produce same output for a 
couple of minutes, and domU finally was destroyed. after 'xm create -c 
ubuntu810' (domain falls into initramfs shell) hypervisor dmesg shows:
blktap: ring-ref 2047, event-channel 4, protocol 1 (x86_64-abi)
tap_frontend_changed: backend/tap/1/51712: prepare for reconnect
blktap: ring-ref 8, event-channel 9, protocol 1 (x86_64-abi)

guest dmesg:
0.403082] device-mapper: uevent: version 1.0.3
[    0.403297] device-mapper: ioctl: 4.14.0-ioctl (2008-04-23) 
initialised: dm-devel@xxxxxxxxxx
[    0.535802]  xvda:
#it's not mistake. dmesg just ends output.

I destroyed domain again and everything repeats.

I think this is caused by 8.10 kernel, which has updated to 2.6.27.7.14 
or 16 recently.
But I have no idea why this keeps crashing the system.
Any ideas?

domU config:
kernel = "/usr/lib/xen/boot/pv-grub-x86_64.gz"
extra = "(hd0,0)/grub/menu.lst"
memory = 2048
vcpus = 2
name = "postgres1"
uuid = "5bb1d470-ae11-389f-b49a-49ff2a84a9f3"
vif = [ 'mac=00:16:3e:36:7f:a9, bridge=eth0' ]
disk = [ 'tap:aio:/var/lib/xen/images/postgres1.img,xvda,w' ]

p.s. please, tell me, if I should provide more information
p.p.s. thanks to Mikhail Kozyrev for spelling

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel