Concourse deployments have been failing with "Expected to find variables: git_private_key" upon further investigation it look like the concourse deployment database db/########-####-####-####-############, is full. An attempt to clean up the database results in "No space left on device" error.
db/########-####-####-####-############:~# df -h
Filesystem Size Used Avail Use% Mounted on
/dev 3.9G 4.0K 3.9G 1%
/dev tmpfs 799M 300K 799M 1%
/run /dev/vda1 2.8G 1.2G 1.5G 44%
/dev/vdb1 962G 913G 0 100%
/var/vcap/store
### relation | total_size
-----------------------------------+------------
public.request_audit_record | 631 GB
public.event_audit_record | 278 GB
public.auth_failure_audit_record | 1560 MB
public.credential | 400 kB
public.access_entry | 296 kB
public.credential_name | 296 kB
public.schema_version | 80 kB
public.encryption_key_canary | 64 kB
public.password_secret | 16 kB
public.user_credential | 16 kB
public.certificate_credential | 16 kB
public.ssh_credential | 16 kB
public.rsa_credential | 16 kB
public.hibernate_sequence | 8192 bytes
(14 rows)
db/########-####-####-####-############:~# /var/vcap/packages/postgres-9.6.6/bin/psql -U credhub
psql (9.6.6)
Type "help" for help.
credhub=> VACUUM FULL;
ERROR: could not access status of transaction 0
DETAIL: Could not write to file "pg_subtrans/1BF0" at offset 114688: No space left on device.
credhub=>
To free up space, you must enlarge the persistent disk. This requires creating a new, larger disk type and assigning it to the persistent disk in the Concourse manifest.
Create a new disk type
Run bosh cloud-config and capture the output into a file.
Edit the file to add a new persistent_disk_type definition.
Apply the changes by running:
Update the Concourse deployment manifest
Run bosh -d concourse manifest > manifest.yml to capture the current manifest.
Edit the manifest so that the db instance uses the new persistent_disk_type for its persistent disk.
Redeploy the changes by running:
After redeployment, the db instance uses the larger persistent disk.