20 VM startup using new snapshot = general error mounting filesystems
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Invalid
|
Undecided
|
Unassigned |
Bug Description
I have a test that creates an instance from a base image that I have independently verified can boot correctly, then snapshots, then uses the new snapshot to boot 20 instances.
The test polls the OSAPI interface for the status of the booting "master" server every 5 seconds, then snapshots as soon as it sees it is 'ACTIVE', waits until the snapshot has been properly saved, then tears down the master. After that is sets off 20 VMs one after another using the snapshot.
This has generally worked, however I've hit an issue where all 20 VMs have failed with a "General error mounting filesystems"
console output shows :
[ 1.385242] md: ... autorun DONE.
[ 1.386256] EXT3-fs (vda): error: couldn't mount because of unsupported optional features (240)
[ 1.450422] EXT2-fs (vda): error: couldn't mount because of unsupported optional features (240)
[ 1.513197] EXT4-fs (vda): mounted filesystem with ordered data mode. Opts: (null)
[ 1.514564] VFS: Mounted root (ext4 filesystem) readonly on device 252:0.
[ 1.516571] devtmpfs: mounted
[ 1.519673] Freeing unused kernel memory: 880k freed
[ 1.521106] Write protecting the kernel read-only data: 10240k
[ 1.523233] Freeing unused kernel memory: 80k freed
[ 1.531096] Freeing unused kernel memory: 1412k freed
lxcmount stop/pre-start, process 57
init: mountall main process (61) killed by FPE signal
General error mounting filesystems.
A maintenance shell will now be started.
CONTROL-D will terminate this shell and reboot the system.
Give root password for maintenance
(or type Control-D to continue):
I suspect the snapshot is getting corrupted either because there's a timing issue (saving too early) or because something in the path through Glance is broken - we've made some recent changes there.
nova version = 2011.3-1 + patches (hp2.1)
Any error or the nova-compute side when this happens ? Do you reproduce with KVM instead of LXC ?