Getting memory Error - 8, TTL not affecting memory eviction


#1

Hi, My Aerospike (2 nudes cluster), is configured as follows: replication-factor 1 memory-size 2G default-ttl 1d # 30 days, use 0 to never expire/evict. high-water-memory-pct 80 storage-engine memory

I’m writing to it, using a java client with TTL above > 0 - 2G of data, I would expect it to be evicted, when getting to 80% of memory. instead, data not seems to be cleaned and i’m getting error o- Memory Error - 8, after data exceeded 2G.

btw,each record is a about 20k of data…

Thanks for help.


#2

Hi,

Could you please provide the exact command you are running to put data into Aerospike? Depending on the speed at which this is happening Aerospike, regardless of high-water-memory-pct, would not be able to keep up.

Please also note that setting high-water-memory-pct at this level would be considered “living dangerously” and would never be recommended unless you have a specific use case.

Thanks Petter