Re: Live migration error in Kilo

Rodrigo · ‎2016-02-17

Hello Guys !
I am using the netapp cinder nfs drive in kilo release, when a try execute the live migrate I receive the error:

"f3d3e2ac4e420fb3951984998c3f66 3336ff9d2c464d78b39d8458af37c866 - - -] [instance: 02c278ba-b79a-49bb-bfd0-b5e7c5667679] Live Migration failure: Unsafe migration: Migration may lead to data corruption if disks use cache != none"

Whats the problem ?

All the best !

dcain · ‎2016-02-24

Hi Rodrigo,

Are you certain that /var/lib/nova/instances is hosted on a shared file system like NFS or Gluster? In most distributions by default this is hosted on local storage and won't work for live migrations.

See here if you're using RHEL-OSP7 (Kilo):

https://access.redhat.com/documentation/en/red-hat-enterprise-linux-openstack-platform/version-7/migrating-instances/

Upstream documentation:

http://docs.openstack.org/admin-guide-cloud/compute-configuring-migrations.html

Rodrigo · ‎2016-02-25

Hello dcain !

Exactly, the /var/lib/nova/instance are on local disks on each computer node.

At this time I have 3 computer nodes, I will try create a new NFS share and mount the same NFS share for all nodes.

I will post in result here in few days.

Thank you for anwser.

Rodrigo · ‎2016-02-29

Hello, dcain !

I change the /var/lib/nova/instances in all computer nodes but a I have the same problem.

Error:

<179>Feb 29 17:43:53 node-3 nova-compute 2016-02-29 17:43:53.943 4270 ERROR nova.virt.libvirt.driver [req-dc1a9aba-5997-47ee-9e05-d28884b6fdfb 05f3d3e2ac4e420fb3951984998c3f66 3336ff9d2c464d78b39d8458af37c866 - - -] [instance: 02c278ba-b79a-49bb-bfd0-b5e7c5667679] Live Migration failure: Unsafe migration: Migration may lead to data corruption if disks use cache != none
<179>Feb 29 17:43:54 node-3 nova-compute 2016-02-29 17:43:54.440 4270 ERROR nova.virt.libvirt.driver [req-dc1a9aba-5997-47ee-9e05-d28884b6fdfb 05f3d3e2ac4e420fb3951984998c3f66 3336ff9d2c464d78b39d8458af37c866 - - -] [instance: 02c278ba-b79a-49bb-bfd0-b5e7c5667679] Migration operation has aborted

Config:

root@node-3:~# cat /etc/nova/nova.conf | grep live_migration_flag
#live_migration_flag=VIR_MIGRATE_UNDEFINE_SOURCE, VIR_MIGRATE_PEER2PEER, VIR_MIGRATE_LIVE, VIR_MIGRATE_TUNNELLED
live_migration_flag=VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_PERSIST_DEST

Mount Point:

root@node-3:~# df -h
Filesystem                        Size  Used Avail Use% Mounted on
udev                               16G   12K   16G   1% /dev
tmpfs                             3.2G   16M  3.2G   1% /run
/dev/dm-1                          27G  2.7G   23G  11% /
none                              4.0K     0  4.0K   0% /sys/fs/cgroup
none                              5.0M     0  5.0M   0% /run/lock
none                               16G     0   16G   0% /run/shm
none                              100M     0  100M   0% /run/user
/dev/sda3                         196M   44M  143M  24% /boot
/dev/mapper/vm-nova                47G   33M   47G   1% /var/lib/nova
10.250.3.253:/vol/openstack_inst   10G  1.2M   10G   1% /var/lib/nova/instances
10.250.3.253:/vol/openstack       750G  647G  104G  87% /var/lib/cinder/mnt/36f09148d1243e64fbe04aa72c497a2e
10.250.3.253:/vol/openstack       750G  647G  104G  87% /var/lib/nova/mnt/36f09148d1243e64fbe04aa72c497a2e

dcain · ‎2016-03-07

How are the permissions setup on the "10.250.3.253:/vol/openstack_inst" directory? Should be 0744 owned by nova:nova if memory serves me. If this is a distribution with Selinux configured, are the approiate allowences setup?

How about debug logs for the nova-scheduler and associated nova-compute logs on the two hosts in question? Any insight there?

Rodrigo · ‎2016-03-08

Hello dcain !

We are here again 😉

The permissions are correct:

root@node-3:~# ls -lh
total 4.0K
-rwx------ 1 root root 611 Feb 11 01:41 openrc
root@node-3:~# ls -la /var/lib/nova/
total 8
drwxr-xr-x 13 nova nova  148 Feb 29 17:25 .
drwxr-xr-x 57 root root 4096 Feb  1 22:04 ..
drwxr-xr-x  2 nova nova    6 Oct 19 12:52 buckets
drwxr-xr-x  6 nova nova   58 Feb  1 21:59 CA
drwx------  2 nova nova   33 Feb 29 17:25 .cache
drwxr-xr-x  3 nova nova   45 Feb  2 13:07 .cinderclient
drwxr-xr-x  2 nova nova    6 Oct 19 12:52 images
drwxr-xr-x 27 nova nova 4096 Mar  7 12:55 instances
drwxr-xr-x  2 nova nova    6 Oct 19 12:52 keys
drwxr-xr-x  3 nova nova   45 Feb  2 12:48 mnt
drwxr-xr-x  2 nova nova    6 Oct 19 12:52 networks
drwx------  2 nova nova   71 Feb  1 22:02 .ssh
drwxr-xr-x  2 nova nova    6 Oct 19 12:52 tmp
root@node-3:~# ls -la /var/lib/nova/instances/
total 108
drwxr-xr-x 27 nova nova 4096 Mar  7 12:55 .
drwxr-xr-x 13 nova nova  148 Feb 29 17:25 ..
drwxrwxr-x  2 nova nova 4096 Feb 29 17:30 02c278ba-b79a-49bb-bfd0-b5e7c5667679
drwxr-xr-x  2 nova nova 4096 Mar  4 22:51 0d7cbb6d-afd7-4699-acde-ea3c14c70856
drwxrwxr-x  2 nova nova 4096 Feb 29 17:27 1118709c-8b9e-46da-afe0-e0516c12a54e
drwxr-xr-x  2 nova nova 4096 Mar  5 00:25 136113f4-5986-47fc-b894-0a1d0adbb654
drwxrwxr-x  2 nova nova 4096 Feb 29 17:28 14ae38ad-a1d3-4087-89b6-f0c67f8fa729
drwxr-xr-x  2 nova nova 4096 Mar  7 12:52 16839f46-78ac-4e16-bfb5-eab3e9dd528a
drwxrwxr-x  2 nova nova 4096 Feb 29 17:32 25b923cb-9b5f-4d61-acba-0f3d7c3ff2a0
drwxr-xr-x  2 nova nova 4096 Mar  4 19:16 2cbe37cd-914e-43cb-a71e-48643c0eab28
drwxrwxr-x  2 nova nova 4096 Feb 29 17:34 30cab62a-6f2f-4ce3-a19d-7f6e2a5409f6
drwxrwxr-x  2 nova nova 4096 Feb 29 17:29 3df4ce9d-958c-4e6f-91e2-8c0244714b0c
drwxrwxr-x  2 nova nova 4096 Feb 29 17:31 4820728f-4ed4-4df7-b19e-45b1bc37f168
drwxr-xr-x  2 nova nova 4096 Mar  4 23:33 4d4a864e-3a55-4acf-9f87-b5fd3c011d85
drwxr-xr-x  2 nova nova 4096 Mar  3 12:53 518f29bd-f765-491c-b77a-02b08a3226e1
drwxr-xr-x  2 nova nova 4096 Mar  7 12:52 556ed1b6-49b8-4cb0-b108-76f86e019648
drwxr-xr-x  2 nova nova 4096 Mar  7 12:53 666b92f9-6cce-41ce-8ddd-894dd891a28b
drwxrwxr-x  2 nova nova 4096 Feb 29 17:35 6ca16ad8-3c9c-47a9-83a4-9bd69fae95bf
drwxr-xr-x  2 nova nova 4096 Mar  7 12:52 6fb5b4cb-1bc8-4b6a-8896-330e7cf190d2
drwxrwxr-x  2 nova nova 4096 Feb 29 17:33 8b15ab17-0e36-4cc4-b033-27471d81b572
drwxr-xr-x  2 nova nova 4096 Mar  7 12:55 b11f8803-f5d1-4896-badf-24a432aa61f9
drwxr-xr-x  2 nova nova 4096 Mar  7 12:53 _base
drwxr-xr-x  2 nova nova 4096 Mar  4 19:40 ca6141e5-3b7f-4e1a-8316-1252163df3f8
-rw-r--r--  1 nova nova  129 Mar  8 21:03 compute_nodes
drwxr-xr-x  2 nova nova 4096 Mar  7 12:53 d03f9276-178f-4379-8fa6-67e0dd6ff30d
drwxr-xr-x  2 nova nova 4096 Mar  7 12:53 db4cac52-b284-49a1-8b0d-91d93e852a68
drwxr-xr-x  2 nova nova 4096 Mar  7 12:53 ff830ed3-6d27-4b13-8de8-690733c186c7
drwxr-xr-x  2 nova nova 4096 Mar  7 12:51 locks
root@node-3:~# df -h
Filesystem                        Size  Used Avail Use% Mounted on
udev                               16G   12K   16G   1% /dev
tmpfs                             3.2G   18M  3.2G   1% /run
/dev/dm-1                          27G  2.8G   23G  11% /
none                              4.0K     0  4.0K   0% /sys/fs/cgroup
none                              5.0M     0  5.0M   0% /run/lock
none                               16G     0   16G   0% /run/shm
none                              100M     0  100M   0% /run/user
/dev/sda3                         196M   44M  143M  24% /boot
/dev/mapper/vm-nova                47G   33M   47G   1% /var/lib/nova
10.250.3.253:/vol/openstack_inst   31G   16G   16G  51% /var/lib/nova/instances
10.250.3.253:/vol/openstack       1.7T  706G  974G  43% /var/lib/cinder/mnt/36f09148d1243e64fbe04aa72c497a2e
10.250.3.253:/vol/openstack       1.7T  706G  974G  43% /var/lib/nova/mnt/36f09148d1243e64fbe04aa72c497a2e
root@node-3:~#

If I add the VIR_MIGRATE_UNSAFE flag in nova.conf I can execute live migrate but I am NOT Insurance about this option.

root@node-3:~# cat /etc/nova/nova.conf | grep live_migration_flag
live_migration_flag=VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_PERSIST_DEST
# live_migration_flag=VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_PERSIST_DEST,VIR_MIGRATE_UNSAFE
root@node-3:~#

Best regards !