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
For a 1TB database shard, master-record is bout 10MB
For a 10TB database shard approx 100MB.
It's important master record is not to large as we store it in the object store on each metadata update.
This is not bad, but we can consider making it smaller still with some strategies:
using varint for integer fields
for rangestart and range end, we can store the range relative to the entry before it (for levels > 0) as we know entries are ordered by range and there are no overlapping ranges.
This could potentially allow us to halve the size.
The text was updated successfully, but these errors were encountered:
For a 1TB database shard, master-record is bout 10MB
For a 10TB database shard approx 100MB.
It's important master record is not to large as we store it in the object store on each metadata update.
This is not bad, but we can consider making it smaller still with some strategies:
This could potentially allow us to halve the size.
The text was updated successfully, but these errors were encountered: