Multiple failed instances during deployment
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
New
|
Undecided
|
Unassigned |
Bug Description
During deployment of openstack using juju+maas multiple units fail.
1/lxd/4 down pending focal Failed creating instance record: Add instance info to the database: This instance already exists
1/lxd/5 started 10.244.40.109 juju-352f48-1-lxd-5 focal zone3 Container started
2 started 10.244.40.103 node3 focal zone1 Deployed
2/lxd/0 down pending focal Failed creating instance record: Add instance info to the database: This instance already exists
2/lxd/1 down pending focal Failed creating instance record: Failed initialising instance: Failed creating storage record: UNIQUE constraint failed: storage_
2/lxd/2 down pending focal Failed creating instance record: Add instance info to the database: This instance already exists
2/lxd/3 pending pending focal Creating container
2/lxd/4 down pending focal Failed creating instance record: Failed initialising instance: Failed creating storage record: UNIQUE constraint failed: storage_
2/lxd/5 down pending focal Failed creating instance record: Add instance info to the database: This instance already exists
2/lxd/6 down pending focal Failed creating instance record: Failed initialising instance: Failed creating storage record: UNIQUE constraint failed: storage_
3 started 10.244.40.107 node4 focal zone2 Deployed
3/lxd/0 down pending focal Failed creating instance record: Failed initialising instance: Failed creating storage record: UNIQUE constraint failed: storage_
3/lxd/1 started 10.244.40.216 juju-352f48-3-lxd-1 focal zone2 Container started
3/lxd/2 started 10.244.40.209 juju-352f48-3-lxd-2 focal zone2 Container started
3/lxd/3 down pending focal Failed creating instance record: Failed initialising instance: Failed creating storage record: UNIQUE constraint failed: storage_
3/lxd/4 down pending focal Failed creating instance record: Add instance info to the database: This instance already exists
The deployment was done in a constraint environment. It can be due to resource contention as it is a deployment into virsh machines on the same server
I believe this is a duplicate of 1945813. We can check there for where we were digging deeper. IIRC the issue is that if the initial launch of an LXD container fails, we weren't reusing the pieces that did get started correctly.