Hi to all! I'm new to this forum and am currently trying to set up a HA cluster on three Dell PowerEdge R730 servers, but having some problems with online migration of CTs. Proxmox 3.4 has just been installed on the servers and subscription keys added. The servers have been combined to a cluster but no HA options have been set up yet.As a first test I tried to set up a simple Debian Jessie (minimal) container on a shared storage (GlusterFS via NFS). The container itself boots okay, but then I'm not able to do a live migration:
The CT template for Debian 8.1 64bit minimal was built according to https://pve.proxmox.com/wiki/Debian_Appliance_Builder I've tried the same with a Debian 7.7 64bit minimal container, with exactly the same results. I've found some other reports about migration problems (https://bugzilla.openvz.org/show_bug.cgi?id=2242) but my CT is just a small Debian installation that does neither have *internal* NFS mounts nor does it run mysql or anthing heavy like that. Migrating a VM (using the same shared storage for the HDD) works fine! I googled around but I'm stuck now. Any suggestions?
Code:
Jun 26 09:10:52 starting migration of CT 100 to node 'metal2' (192.168.100.2)
Jun 26 09:10:52 container is running - using online migration
Jun 26 09:10:52 container data is on shared storage 'systems'
Jun 26 09:10:52 start live migration - suspending container
Jun 26 09:10:52 dump container state
Jun 26 09:10:53 dump 2nd level quota
Jun 26 09:10:54 initialize container on remote node 'metal2'
Jun 26 09:10:54 initializing remote quota
Jun 26 09:11:03 turn on remote quota
Jun 26 09:11:03 load 2nd level quota
Jun 26 09:11:03 starting container on remote node 'metal2'
Jun 26 09:11:03 restore container state
Jun 26 09:11:04 # /usr/bin/ssh -o 'BatchMode=yes' root@192.168.100.2 vzctl restore 100 --undump --dumpfile /mnt/pve/systems/dump/dump.100 --skip_arpdetect
Jun 26 09:11:03 Restoring container ...
Jun 26 09:11:03 Starting container ...
Jun 26 09:11:03 Container is mounted
Jun 26 09:11:03 undump...
Jun 26 09:11:03 Adding IP address(es): 192.168.99.1
Jun 26 09:11:03 Setting CPU units: 1000
Jun 26 09:11:03 Setting CPUs: 1
Jun 26 09:11:03 vzquota : (warning) Quota is running for id 100 already
Jun 26 09:11:03 Error: undump failed: Invalid argument
Jun 26 09:11:03 Restoring failed:
Jun 26 09:11:03 Error: bind_unix_socket: can't rebind unix socket -22
Jun 26 09:11:03 Error: bind unix: -22
Jun 26 09:11:03 Error: rst_sockets: open_listening_socket: -22
Jun 26 09:11:03 Error: rst_sockets: -22
Jun 26 09:11:04 Container is unmounted
Jun 26 09:11:04 ERROR: online migrate failure - Failed to restore container: Container start failed
Jun 26 09:11:04 start final cleanup
Jun 26 09:11:04 ERROR: migration finished with problems (duration 00:00:12)
TASK ERROR: migration problemsThe CT template for Debian 8.1 64bit minimal was built according to https://pve.proxmox.com/wiki/Debian_Appliance_Builder I've tried the same with a Debian 7.7 64bit minimal container, with exactly the same results. I've found some other reports about migration problems (https://bugzilla.openvz.org/show_bug.cgi?id=2242) but my CT is just a small Debian installation that does neither have *internal* NFS mounts nor does it run mysql or anthing heavy like that. Migrating a VM (using the same shared storage for the HDD) works fine! I googled around but I'm stuck now. Any suggestions?
Last edited by a moderator: