Closed nickumia-reisys closed 2 years ago
I tried to force the SSB to change the DB by renaming it before our recent deployment. But it didn't work. The terraform plan just renamed the old DB to the correct name because it had the same ID.
If after the deployment, the issue persists, we will have to do some diagnosis on the DB or just delete it and recreate it to see if the error comes back.
Edit: Delete the DB with the caveat that the broker will no longer be able to manage any of solr clusters that were in the DB.
After investigation of the cloud-service-broker configuration for ssb-solrcloud
, it seems like DB_HOST
, DB_USERNAME
and DB_PASSWORD
were being set, but not DB_NAME
. Because of this, the DB_NAME
was defaulting to servicebroker
which was the wrong name for the database in question.
Those environment variables should not need to be set (as they are not set in the ssb-solr
, ssb-smtp
and ssb-eks
brokers) so we removed them from the User-provided environment variables
and restaged the ssb-solrcloud
broker. This was able to fix the issue. However, we're not sure how this happened and are not sure this won't happen again.
cloud.gov datagov-ssb @ b12dd98da5f659e3ee275ecd288b738cc61d6663
How it was discovered
cf t -s managment
cf ds ssb-solrcloud-k8s
cf create-user-provided-service ssb-solrcloud-k8s -p k8s-xxx.json
cf bs ssb-solrcloud ssb-solrcloud-k8s
cf rs ssb-solrcloud
cf logs ssb-solrcloud --recent
)Expected behavior
Upon restart, ssb-solrcloud should be working again.
Actual behavior
DB somehow broke.
Sketch
Not sure.