ArangoDB memory overflow - is there no solution?

Question

When I perform AQL queries, I run into this error:

Query: AQL: query would use more memory than allowed (while executing)

My dataset: 9 billion vertices, 19 billion edges

My cluster has three nodes, running Centos 7 with 32 GB RAM

I only hope that the query can be executed and will not be reported wrong. What configuration files can be modified to achieve the purpose? I am new to ArangoDB.

Welcome to SO! Please read how to ask.

Cris Luengo
– Cris Luengo

2017-12-06 02:53:25 +00:00
Commented Dec 6, 2017 at 2:53 — Cris Luengo
– Cris Luengo, Commented Dec 6, 2017 at 2:53

peak · Accepted Answer · 2017-12-07 05:35:20Z

2

Some back-of-the-envelope calculations suggest you will need to use the RocksDB engine:

(100 * 28 * 10^9) / (3 * (32 * 10^9)) => 29.17

See https://www.arangodb.com/why-arangodb/comparing-rocksdb-mmfiles-storage-engines/

edited Dec 7, 2017 at 5:35

answered Dec 6, 2017 at 23:00

peak

119k21 gold badges185 silver badges218 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

feitianStyle Over a year ago

I have already used rocksdb，You can tell me how you calculated this，thank you！

peak Over a year ago

I just used your numbers and guessed an average for the number of bytes per (V+E). If you are already using RocksDB and have enough disk space, then perhaps the cause is that ArangoDB has some memory requirements regarding the indexes it maintains. What indexes have you defined? It sounds like maybe you will have to reach out to ArangoDB more directly , e.g. at github.com/arangodb/arangodb/issues

feitianStyle Over a year ago

I don't have a custom index，It's the default index of arangodb，Yes. My disk space is enough，I've already raised questions on GitHub, but the development team didn't solve me

peak Over a year ago

Maybe the default indices require a lot of RAM? Have you looked at github.com/arangodb/arangodb/issues/3806 ? What is your issue #?

Collectives™ on Stack Overflow

ArangoDB memory overflow - is there no solution?

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related