You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Could we do something like take a → b → c and make 2 things:
For each arrow, make a record in a cache_relations table that points to this new cache relations
Create a cached relation of a→c in the relations table
When ever we delete a relation (the arrow), we look up whether it existed in the cache_relations table. If so, we delete the cached relation in the relations table.
Would need to record all entites in the graph so that any change could be detected and purged the change.
OR if we are running in an environment where we can see and identify other nodes (kubernetes for example), we could use Raft and elect leaders. But at that rate why even use CRDB if we can just do KV locally then...
Maybe Group Cache? It's not strongly consistent but it is VERY close...
The text was updated successfully, but these errors were encountered:
Could we do something like take a → b → c and make 2 things:
cache_relations
table that points to this new cache relationsrelations
tableWhen ever we delete a relation (the arrow), we look up whether it existed in the
cache_relations
table. If so, we delete the cached relation in therelations
table.Would need to record all entites in the graph so that any change could be detected and purged the change.
OR if we are running in an environment where we can see and identify other nodes (kubernetes for example), we could use Raft and elect leaders. But at that rate why even use CRDB if we can just do KV locally then...
Maybe Group Cache? It's not strongly consistent but it is VERY close...
The text was updated successfully, but these errors were encountered: