[
https://issues.apache.org/jira/browse/CLOUDSTACK-6741?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Bret Mette updated CLOUDSTACK-6741:
-----------------------------------
Description:
I created an instance using the web GUI. I then created a template from that
instance, stopped the instance and created a new instance from the template.
All was well. I noticed I wanted to change something in the template so I
stopped the instance, destroyed it and started up the instance I made the
template from. The instance I made the template from was not able to get an IP
from DHCP.
I destroyed the instance, destroyed the template and created a new instance
with the same name using the instance wizard.
The new instance was also unable to get an IP from DHCP on the VR. I logged
into the VR and checked the basics. The MAC for the new instance was showing as
"incomplete" when I ran arp -n.
root@r-113-VM:~# cat /etc/hosts
127.0.0.1 localhost
# The following lines are desirable for IPv6 capable hosts
::1 localhost ip6-localhost ip6-loopback
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
10.0.10.254 data-server
10.0.10.254 r-113-VM
10.0.10.10 xencenter
10.0.10.11 bret-dev
10.0.10.13 cpanel-temp
10.0.10.12 centos-6-5-minimal
root@r-113-VM:~# cat /etc/dhcphosts.txt
06:f6:b6:00:0a:b0,set:10_0_10_10,10.0.10.10,xencenter,infinite
06:3d:6e:00:0a:b1,set:10_0_10_11,10.0.10.11,bret-dev,infinite
06:24:a0:00:0a:b3,set:10_0_10_13,10.0.10.13,cpanel-temp,infinite
06:c6:f6:00:0a:b2,set:10_0_10_12,10.0.10.12,centos-6-5-minimal,infinite
root@r-113-VM:~# arp -n
Address HWtype HWaddress Flags Mask Iface
10.0.10.11 ether 06:3d:6e:00:0a:b1 C
eth0
10.0.10.12 (incomplete)
eth0
x.x.x.x ether x:x:x:x:x:x C
eth2
169.254.0.1 ether 4e:ad:18:d9:72:47 C
eth1
I then destroyed (but did not expunge) the instance and changed the MAC in the
cloud.nics table in the database. Restored the instance, it uses the same IP
but a different MAC and now works.
Why would CloudStack reuse a MAC, and why would reusing the MAC cause a routing
issue? Maybe the new instance was being created on a different host in the
cluster and the host's routing table was not updated / confused?
Here is the same information after I restored the instance using the new MAC
root@r-113-VM:/etc/init.d# cat /etc/hosts
127.0.0.1 localhost
# The following lines are desirable for IPv6 capable hosts
::1 localhost ip6-localhost ip6-loopback
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
10.0.10.254 data-server
10.0.10.254 r-113-VM
10.0.10.10 xencenter
10.0.10.11 bret-dev
10.0.10.13 cpanel-temp
10.0.10.12 centos-6-5-minimal
root@r-113-VM:/etc/init.d# cat dhcphosts.txt
cat: dhcphosts.txt: No such file or directory
root@r-113-VM:/etc/init.d# cat /etc/dhcphosts.txt
06:f6:b6:00:0a:b0,set:10_0_10_10,10.0.10.10,xencenter,infinite
06:3d:6e:00:0a:b1,set:10_0_10_11,10.0.10.11,bret-dev,infinite
06:24:a0:00:0a:b3,set:10_0_10_13,10.0.10.13,cpanel-temp,infinite
06:c6:f6:00:0a:b4,set:10_0_10_12,10.0.10.12,centos-6-5-minimal,infinite
root@r-113-VM:/etc/init.d# arp -n
Address HWtype HWaddress Flags Mask Iface
10.0.10.11 ether 06:3d:6e:00:0a:b1 C eth0
10.0.10.12 ether 06:c6:f6:00:0a:b4 C eth0
x.x.x.x ether x:x:x:x:x:x C eth2
169.254.0.1 ether 4e:ad:18:d9:72:47 C eth1
was:
I created an instance using the web GUI. I then created a template from that
instance, stopped the instance and created a new instance from the template.
All was well. I noticed I wanted to change something in the template so I
stopped the instance, destroyed it and started up the instance I made the
template from. The instance I made the template from was not able to get an IP
from DHCP.
I destroyed the instance, destroyed the template and created a new instance
with the same name using the instance wizard.
The new instance was also unable to get an IP from DHCP on the VR. I logged
into the VR and checked the basics. The MAC for the new instance was showing as
"incomplete" when I ran arp -n.
root@r-113-VM:~# cat /etc/hosts
127.0.0.1 localhost
# The following lines are desirable for IPv6 capable hosts
::1 localhost ip6-localhost ip6-loopback
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
10.0.10.254 data-server
10.0.10.254 r-113-VM
10.0.10.10 xencenter
10.0.10.11 bret-dev
10.0.10.13 cpanel-temp
10.0.10.12 centos-6-5-minimal
root@r-113-VM:~# cat /etc/dhcphosts.txt
06:f6:b6:00:0a:b0,set:10_0_10_10,10.0.10.10,xencenter,infinite
06:3d:6e:00:0a:b1,set:10_0_10_11,10.0.10.11,bret-dev,infinite
06:24:a0:00:0a:b3,set:10_0_10_13,10.0.10.13,cpanel-temp,infinite
06:c6:f6:00:0a:b2,set:10_0_10_12,10.0.10.12,centos-6-5-minimal,infinite
root@r-113-VM:~# arp -n
Address HWtype HWaddress Flags Mask Iface
10.0.10.11 ether 06:3d:6e:00:0a:b1 C
eth0
10.0.10.12 (incomplete)
eth0
x.x.x.x ether x:x:x:x:x:x C
eth2
169.254.0.1 ether 4e:ad:18:d9:72:47 C
eth1
I then destroyed (but did not expunge) the instance and changed the MAC in the
cloud.nics table in the database. Restored the instance, it uses the same IP
but a different MAC and now works.
Why would CloudStack reuse a MAC, and why would reusing the MAC cause a routing
issue? Maybe the new instance was being created on a different host in the
cluster and the host's routing table was not updated / confused?
Here is the same information after I restored the instance using the new MAC
root@r-113-VM:/etc/init.d# cat /etc/hosts
127.0.0.1 localhost
# The following lines are desirable for IPv6 capable hosts
::1 localhost ip6-localhost ip6-loopback
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
10.0.10.254 data-server
10.0.10.254 r-113-VM
10.0.10.10 xencenter
10.0.10.11 bret-dev
10.0.10.13 cpanel-temp
10.0.10.12 centos-6-5-minimal
root@r-113-VM:/etc/init.d# cat dhcphosts.txt
cat: dhcphosts.txt: No such file or directory
root@r-113-VM:/etc/init.d# cat /etc/dhcphosts.txt
06:f6:b6:00:0a:b0,set:10_0_10_10,10.0.10.10,xencenter,infinite
06:3d:6e:00:0a:b1,set:10_0_10_11,10.0.10.11,bret-dev,infinite
06:24:a0:00:0a:b3,set:10_0_10_13,10.0.10.13,cpanel-temp,infinite
06:c6:f6:00:0a:b4,set:10_0_10_12,10.0.10.12,centos-6-5-minimal,infinite
root@r-113-VM:/etc/init.d# arp -n
Address HWtype HWaddress Flags Mask Iface
10.0.10.11 ether 06:3d:6e:00:0a:b1 C eth0
10.0.10.12 ether 06:c6:f6:00:0a:b4 C eth0
68.234.25.129 ether 00:15:fa:88:c5:79 C eth2
169.254.0.1 ether 4e:ad:18:d9:72:47 C eth1
> VM in shared networking unable to get IP from virtual router
> ------------------------------------------------------------
>
> Key: CLOUDSTACK-6741
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6741
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: Management Server, Network Devices, Virtual Router
> Affects Versions: 4.3.0
> Environment: CloudStack 4.3.0
> XenServer 6.2.0
> Reporter: Bret Mette
> Priority: Critical
>
> I created an instance using the web GUI. I then created a template from that
> instance, stopped the instance and created a new instance from the template.
> All was well. I noticed I wanted to change something in the template so I
> stopped the instance, destroyed it and started up the instance I made the
> template from. The instance I made the template from was not able to get an
> IP from DHCP.
> I destroyed the instance, destroyed the template and created a new instance
> with the same name using the instance wizard.
> The new instance was also unable to get an IP from DHCP on the VR. I logged
> into the VR and checked the basics. The MAC for the new instance was showing
> as "incomplete" when I ran arp -n.
> root@r-113-VM:~# cat /etc/hosts
> 127.0.0.1 localhost
> # The following lines are desirable for IPv6 capable hosts
> ::1 localhost ip6-localhost ip6-loopback
> ff02::1 ip6-allnodes
> ff02::2 ip6-allrouters
> 10.0.10.254 data-server
> 10.0.10.254 r-113-VM
> 10.0.10.10 xencenter
> 10.0.10.11 bret-dev
> 10.0.10.13 cpanel-temp
> 10.0.10.12 centos-6-5-minimal
> root@r-113-VM:~# cat /etc/dhcphosts.txt
> 06:f6:b6:00:0a:b0,set:10_0_10_10,10.0.10.10,xencenter,infinite
> 06:3d:6e:00:0a:b1,set:10_0_10_11,10.0.10.11,bret-dev,infinite
> 06:24:a0:00:0a:b3,set:10_0_10_13,10.0.10.13,cpanel-temp,infinite
> 06:c6:f6:00:0a:b2,set:10_0_10_12,10.0.10.12,centos-6-5-minimal,infinite
> root@r-113-VM:~# arp -n
> Address HWtype HWaddress Flags Mask
> Iface
> 10.0.10.11 ether 06:3d:6e:00:0a:b1 C
> eth0
> 10.0.10.12 (incomplete)
> eth0
> x.x.x.x ether x:x:x:x:x:x C
> eth2
> 169.254.0.1 ether 4e:ad:18:d9:72:47 C
> eth1
> I then destroyed (but did not expunge) the instance and changed the MAC in
> the cloud.nics table in the database. Restored the instance, it uses the same
> IP but a different MAC and now works.
> Why would CloudStack reuse a MAC, and why would reusing the MAC cause a
> routing issue? Maybe the new instance was being created on a different host
> in the cluster and the host's routing table was not updated / confused?
> Here is the same information after I restored the instance using the new MAC
> root@r-113-VM:/etc/init.d# cat /etc/hosts
> 127.0.0.1 localhost
> # The following lines are desirable for IPv6 capable hosts
> ::1 localhost ip6-localhost ip6-loopback
> ff02::1 ip6-allnodes
> ff02::2 ip6-allrouters
> 10.0.10.254 data-server
> 10.0.10.254 r-113-VM
> 10.0.10.10 xencenter
> 10.0.10.11 bret-dev
> 10.0.10.13 cpanel-temp
> 10.0.10.12 centos-6-5-minimal
> root@r-113-VM:/etc/init.d# cat dhcphosts.txt
> cat: dhcphosts.txt: No such file or directory
> root@r-113-VM:/etc/init.d# cat /etc/dhcphosts.txt
> 06:f6:b6:00:0a:b0,set:10_0_10_10,10.0.10.10,xencenter,infinite
> 06:3d:6e:00:0a:b1,set:10_0_10_11,10.0.10.11,bret-dev,infinite
> 06:24:a0:00:0a:b3,set:10_0_10_13,10.0.10.13,cpanel-temp,infinite
> 06:c6:f6:00:0a:b4,set:10_0_10_12,10.0.10.12,centos-6-5-minimal,infinite
> root@r-113-VM:/etc/init.d# arp -n
> Address HWtype HWaddress Flags Mask
> Iface
> 10.0.10.11 ether 06:3d:6e:00:0a:b1 C
> eth0
> 10.0.10.12 ether 06:c6:f6:00:0a:b4 C
> eth0
> x.x.x.x ether x:x:x:x:x:x C eth2
> 169.254.0.1 ether 4e:ad:18:d9:72:47 C
> eth1
--
This message was sent by Atlassian JIRA
(v6.2#6252)
