[ 4.0.3.0-113] Provisioning fails on re triggering provision_containers.sh without making any change in the json file
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R4.0 |
Fix Committed
|
Undecided
|
Abhay Joshi | |||
R4.1 |
Fix Committed
|
Undecided
|
Abhay Joshi | |||
Trunk |
Fix Committed
|
Undecided
|
Abhay Joshi |
Bug Description
Steps:
1) Bring up the setup using the json file. Server manager is on nodei5
/opt/contrail/
2) Trigger /opt/contrail/
outcome: CN becomes unreachable
Reason seems to be the following:
preconfig tried to configure eth0 with the management IP using script interface_setup which failed because vhost0 was already up and configured with the IP which was a conflict and the VM became unreachable
json file:
{
"cluster" : [
{
"id" : "nodel11_cluster",
}
}
}
],
"server": [
{
"id": "nodel11-vm1",
]
},
},
{
"id": "nodel11-vm2",
]
},
},
{
"id": "nodel11-vm3",
]
},
},
{
"id": "nodel11-vm4",
]
},
},
{
"id": "nodel11-vm5",
]
},
}
],
"image": [
{
"id": "regress_image",
"type": "contrail-
"path": "/root/
}
]
}
Logs and Json file copied at :
/cs-shared/
Fix:-
root@nodei5: /opt/contrail/ server_ manager/ client# diff preconfig.py preconfig_old.py interfaces' /opt/contrail/ server_ manager/ client#
301c301
< cmd = 'grep -rn "set vhost0 address" /etc/network/'
---
> cmd = 'grep "set vhost0 address" /etc/network/
root@nodei5: