We are having issues building a secondary from a sharded replica, it keeps getting to the same database and fails with the following error
2021-03-16T21:03:06.756+0000 I REPL [repl writer worker 7] CollectionCloner::start called, on ns:customer.tmpdqhEZ.rename
2021-03-16T21:03:06.758+0000 F STORAGE [repl writer worker 8] Attempted to create a new collection customer.tmpdqhEZ.rename without a UUID
2021-03-16T21:03:06.765+0000 I REPL [repl writer worker 8] CollectionCloner ns:customer.tmpdqhEZ.rename finished cloning with status: InvalidOptions: Attempted to create a new collection customer.tmpdqhEZ.rename without a UUID
2021-03-16T21:03:06.765+0000 W REPL [repl writer worker 8] collection clone for ‘customer.tmpdqhEZ.rename’ failed due to InvalidOptions: While cloning collection ‘customer.tmpdqhEZ.rename’ there was an error ‘Attempted to create a new collection customer.tmpdqhEZ.rename without a UUID’
Is there a way to drop this dead collection from the replicaset? It is not showing up when looking for collections in customer db?
Can you confirm the specific version(s) of MongoDB server used in your cluster?
Can you also share more context on the operations you are doing leading to this error?
It sounds like you may be trying to perform initial sync of a new secondary in a shard replica set, in which case it would be helpful to know fi you are seeding this secondary by copying data files from another member or using the normal initial sync process.
Thank you, this is a sharded cluster with each shard consisting of a primary and 2 replicas. The version currently running is 3.6.20. We found the remaining collection on the secondary and not the primary so it didn’t get the UUID with the Feature Compatibility change. Syncing the new secondary from the primary allowed the initial sync to complete.