Migrating Stores¤
Sizing and Deployment¤
- size and deploy of the new store (refer to the capacity planning / sizing considerations, refer to the docker container / orchestration)
- store specific config (e.g. search-all-graphs in SD)
Transferring Data and Configuration¤
- backing-up / exporting and restore / import of graphs, DI-projects, configuration (if any)
- graphs
- blacklisting the DI projects graphs
- config
- DM
- stardog text match support (this is a DM parameter!)
- search queries
- navigation
- DM
- DP
- configure the resp. store
- DI
- nothing to do … just duplicate / copy the configuration as-is
cmemc admin workspace export / import
- graphs
Test and Validation¤
- best practice:
- run all (SELECT) queries in the query catalog and compare results (e.g. with
cmemc
)- theoretically this could also be applied to INSERT queries (by re-writing into SELECTS in case you want / need to omit altering your graphs)
- count all triples in all graphs on both instances before/after export/import (
cmemc graph count --all
)
- run all (SELECT) queries in the query catalog and compare results (e.g. with
Optimizing Your Setup¤
- optimizing customization (e.g. queries in SHAPES; DI; DM-config)
- “textmatch” / “lucene” queries need to be migrated (a query can be helpful to find these queries…)
- performance comparisons could be automated via
cmemc query replay
- identify query that won’t run or run slow
- general query best practices
- → query optimization guide
- use
VALUE
instead ofFILTER (?x IN (...))
(esp. on GDB)
- use
- → query optimization guide