Only steps in bold should cause a downtime.
- change savannah secondary node (will take a long time, need to resync all the drbd)
gnt-instance replace-disks -n z2-7.host.gnt
- update shorewall config on rennes to use new ip (86.65.39.29)
- switch instance to new node :
gnt-instance failover z2-7.host.gnt
This step should take less than 10 minutes (time to stop the vm on primary node, and start it on secondary).
After the failover, vm should be available on new node (z2-7) with secondary on z2-4, estimated downtime < 10min.
Done
savannah vm is now available on 86.65.39.29 .
Downtime was about 20 minutes (some configuration / routing issue on gw.rennes.tld, a route "10.10.1.3 via 10.10.10.254" was missing)
z2-2:~# gnt-instance info savannah.vm.gnt
Instance name: savannah.vm.gnt
State: configured to be up, actual state is up
Nodes:
- primary: z2-7.host.gnt
- secondaries: z2-4.host.gnt
Operating system: debootstrap
Allocated network port: 11096
Hypervisor: kvm
- nic_type: default (paravirtual)
- vnc_tls: default (False)
- kernel_args: default (ro)
- vnc_x509_verify: default (False)
- vnc_x509_path: default ()
- vnc_bind_address: default ()
- root_path: default (/dev/vda1)
- serial_console: default (True)
- cdrom_image_path: default ()
- usb_mouse: default ()
- initrd_path: default (/boot/initrd.img-2.6-domU)
- boot_order: default (disk)
- disk_type: default (paravirtual)
- kernel_path: default (/boot/vmlinuz-2.6-domU)
- acpi: default (True)
Hardware:
- VCPUs: 4
- memory: 5120MiB
- NICs:
- nic/0: MAC: 52:54:cf:00:61:78, IP: None, bridge: br0
Disks:
- disk/0: drbd8, size 20.0G
access mode: rw
nodeA: z2-4.host.gnt, minor=7
nodeB: z2-7.host.gnt, minor=1
port: 11097
auth key: 95ba88401d4996a0673602db089acfebfcd83ef9
on primary: /dev/drbd1 (147:1) in sync, status ok
on secondary: /dev/drbd7 (147:7) in sync, status ok
child devices:
- child 0: lvm, size 20.0G
logical_id: all/470046ff-2394-4d49-9203-886b062f8949.disk0_data
on primary: /dev/all/470046ff-2394-4d49-9203-886b062f8949.disk0_data (254:2)
on secondary: /dev/all/470046ff-2394-4d49-9203-886b062f8949.disk0_data (253:6)
- child 1: lvm, size 128M
logical_id: all/470046ff-2394-4d49-9203-886b062f8949.disk0_meta
on primary: /dev/all/470046ff-2394-4d49-9203-886b062f8949.disk0_meta (254:3)
on secondary: /dev/all/470046ff-2394-4d49-9203-886b062f8949.disk0_meta (253:7)
- disk/1: drbd8, size 500.0G
access mode: rw
nodeA: z2-4.host.gnt, minor=8
nodeB: z2-7.host.gnt, minor=2
port: 11098
auth key: aceee540fd23d88b99e2ba6dd3297e23d6c6e186
on primary: /dev/drbd2 (147:2) in sync, status ok
on secondary: /dev/drbd8 (147:8) in sync, status ok
child devices:
- child 0: lvm, size 500.0G
logical_id: all/f3fece3d-2381-4591-b382-10ddca1cee05.disk1_data
on primary: /dev/all/f3fece3d-2381-4591-b382-10ddca1cee05.disk1_data (254:4)
on secondary: /dev/all/f3fece3d-2381-4591-b382-10ddca1cee05.disk1_data (253:8)
- child 1: lvm, size 128M
logical_id: all/f3fece3d-2381-4591-b382-10ddca1cee05.disk1_meta
on primary: /dev/all/f3fece3d-2381-4591-b382-10ddca1cee05.disk1_meta (254:5)
on secondary: /dev/all/f3fece3d-2381-4591-b382-10ddca1cee05.disk1_meta (253:9)
Dns updated :
-
etc/bind/db.fsffrance.org
a b 9 9 10 10 $TTL 86400 11 11 @ IN SOA ns1.fsffrance.org.org. hostmaster.fsffrance.org. ( 12 200911 0901 ; serial12 2009111601 ; serial 13 13 28800 ; refresh 14 14 7200 ; retry 15 15 604800 ; expire … … 207 207 cmd28 IN A 86.65.39.28 ;; Rennes 28 ;; ubuntu-fr herve@moulticast.net 208 208 cmd29 IN A 86.65.39.29 ;; Rennes 29 ;; savannah loic@gnu.org 209 209 cmd30 IN A 86.65.39.30 ;; Rennes 30 210 savannah IN A 8 7.98.156.150210 savannah IN A 86.65.39.29 ;; savannah loic@gnu.org 211 211 trac IN CNAME main.binbang.org. ;; loic@dachary.org 212 212 jolicloud IN CNAME cmd18 213 213 munin IN CNAME main.binbang.org. ;; loic@dachary.org
