Hello everyone !

Happy new year to you :)

Coming back from holidays, I've been doing some tests to run Pregel on 
ArangoDB using a moderate dataset ( graph Around 50 millions nodes + edges 
).

I just wanted to run a pagerank to see how the pregel execution behaves on 
our server, so I launched a pregel job, it ran for one hour and when I 
stopped it, I saw this warning into the logs : 

2019-01-02T15:54:33Z [33639] WARNING {pregel} Aborted loading graph

It ran for a complete hour and yet the graph was still loading into memory 
? What bothers me is that the server I'm using is quite powerful, its all 
bare metal and the data are stored on local SSD with fast access, network 
problem / bottleneck is non relevant.
So why does iotop shows that ArangoDB is only loading data at ~650 
KiloBytes / s when it can go up to 30 MegaBytes / s !?

Loading data within the Pregel job seems to be monothread, fine. But still, 
even in monothread it should be much faster,

What am i missing ?

Thank you for your answer,

Best Regards

Cyprien.

-- 
You received this message because you are subscribed to the Google Groups 
"ArangoDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to