Speeding up node adding to Janusgraph
Hi Everybody, I am using Janusgraph with Berkley DB JE. In my use case I have to add initially nodes one by one. The number could be quite huge in certain cases, i.e., 200k-500k. It is currently taking quite some time, as compared to inmemory test setup.
Have tried the following to expedite and got some improvement:
.set("storage.berkeleyje.cache-percentage", CACHE_PERCENTAGE)
.set("cache.db-cache",true)
.set("cache.db-cache-size", DB_CACHE_SIZE)
private const val CACHE_PERCENTAGE = 80
private const val DB_CACHE_SIZE = 0.8
Cannot change the backend due to some limitations posed by customer.
Any suggestion for improvements are welcome.
Thanks and regards
Tanvir
Have tried the following to expedite and got some improvement:
.set("storage.berkeleyje.cache-percentage", CACHE_PERCENTAGE)
.set("cache.db-cache",true)
.set("cache.db-cache-size", DB_CACHE_SIZE)
private const val CACHE_PERCENTAGE = 80
private const val DB_CACHE_SIZE = 0.8
Cannot change the backend due to some limitations posed by customer.
Any suggestion for improvements are welcome.
Thanks and regards
Tanvir
