I executed
> update nics set device_id = 1 where id = 29;
After restarting the router, the interfaces file now looks like this:
root@r-19-VM:~# cat /etc/network/interfaces
auto lo eth0 eth1 eth2
iface lo inet loopback
iface eth0 inet static
address 169.254.3.155
netmask 255.255.0.0
iface
I updated nics.gateway for that network, but the VM still shows the same
behaviour.
If one compares interfaces:
root@r-19-VM:~# cat /etc/network/interfaces
auto lo eth0 eth1 eth2
iface lo inet loopback
iface eth0 inet static
address 169.254.1.242
netmask 255.255.0.0
iface eth1 inet stati
Hi,
select id, name, traffic_type, broadcast_domain_type, cidr, gateway, mode,
state, removed from networks where id = 205;
+-+-+--+---++-+--+---+-+
| id | name| traffic_type | broadcast_domain_type |
No, it doesn’t seem right to me having 2 nics with device id 0. But looks
like they’ve got programmed to correct devices on the backend per your
prev email?
iface eth0 inet static
address 169.254.1.59
netmask 255.255.0.0
iface eth1 inet static
address 10.193.17.1
Netmask
iface eth2 i
Hi,
select id,ip4_address,netmask,gateway,state,removed,network_id,reserver_name
from nics where instance_id=19;
++---+---+-+---+-++--+
| id | ip4_address | netmask | gateway | state | rem
Kambiz, the debug statements below are for the case when eth1 is a control
interface as it was in your old command. I’ve looked at the new command,
eth1 is not control, its either public or guest
eth0: - control
iface eth0 inet static
address 169.254.1.59
netmask 255.255.0.0
eth1:
iface
Kambiz, can you please check the following:
1) Check if the gateway is set on control network:
mysql> select gateway, cidr from networks where traffic_type=‘Control’;
2) For router control nic, check if network/gateway are set.
Select gateway,netmask from nics where instance_id= and
network_id
Hi,
I looked up the startup command of the old router instance which worked
correctly:
/usr/share/cloudstack-common/scripts/vm/hypervisor/kvm/rundomrpre.sh -l r-7-VM
-t all -d /var/lib/libvirt/images/r-7-VM-patchdisk -p
%template=domP%name=r-7-VM%eth2ip=10.193.17.190%eth2mask=255.255.255.0%gate
I don’t think its relevant as the piece we’ve fixed, just eliminated
static nat rule programming for non-existing vm. Missing netmask on eth1
doesn’t seem related to the problem (although we have to figure out why
its missing), as the connection that fails, happening to link local 169.x
eth0 interf
Hi,
thank you, the NullPointerException doesn't occur any more, but there
still seems to be a problem during startup of the router.
When I start the virtual router, it comes up, but in agent.log, there
are lots of 'Could not connect to 169.254.1.x' messages.
Then I logged into the virtual route
Yes, Kambiz, you followed up right, and vm id=15 is the culprit. If vm
id=15 is expunged, we have to clear out the reference to it from
user_ip_address table. Here is the flow:
1) Save the db dump.
2) Run the query to cleanup the reference:
Update user_ip_address set one_to_one_nat=0, instance_id
Hi,
I hope I have understood what you wrote and created the following query
correctly:
select uip.vm_id, uip.network_id, uip.public_ip_address,
n.state as nic_state, n.removed as nic_removed,
vm.state as vm_state, vm.removed as vm_removed
from user_ip_address uip
join nics n on
Kambiz, can you please try one more thing.
1) Locate all the firewall rules for your guest network (205, right?)
Select id, ip_address_id from firewall_rules where network_id=205;
2) Now get all static nat enabled ip addresses for those rules:
Select vm_id, network_id from user_ip_address where
Hi Alena,
thank you for your help.
The query returns no rows, i.e. nics.removed was not null, but I removed
the row though to see what happens: a new virtual router was created
which also couldn't be started due to the same NPE. I reverted the
change by restoring from the dump.
I have to mention
The error happens not because Ip is null, but because the nic in a certain
network can¹t be found. Looks like there is some bug in VPC nic
plug/unplug for Guest networks process.
Kambiz, please do the following to fix it:
1) Stop the MS
2) Take the DB dump of cloud db in case you have to revert
15 matches
Mail list logo
