Database record ID management

Database record IDs are allocated to replicas in ID blocks. The range of record IDs (for both stateful and stateless records) is broken into blocks that are assigned to each replica to be allocated locally within those replicas.

The idblockinfo command lists information about the ID blocks for a replica family. When the command is run at the master site, it provides the latest information about the overall state of the record ID allocations. When the command is run at a replica site, the information is accurate for the last syncreplica -import operation from the master site. The number of available IDs in the ID blocks assigned to each replica is current for the last syncreplica -import operation from those replicas.

You can use the idblockinfo command before you import a large number of records to ensure that the replica has sufficient IDs available for the imported records. You can also monitor ID block consumption to determine appropriate values for the replica ID block size and threshold. Determine the consumption rate of each replica by comparing the available counts in individual blocks over time or the allocation times of a series of blocks. You can use this information to increase the block sizes for the replicas that are using record IDs faster than others.