Details
-
Story
-
Resolution: Incomplete
-
Critical
-
1.6.4
-
Security Level: Public
-
None
Description
I setup 4 m1.large nodes in EC2 with a latest build. From a client running a local Moxi (configured to point at all 4 nodes) I inserted 1 mm items of size 10k:
./brutis -x localhost -f 8 -s 10240 -r 1:0 -k 1000000 -v
Test case: MB-70 with a Brutis load of pure gets: ./brutis -x localhost -f 8 -r 0:1 -k 1000000 -v
Rebalance one of the nodes out of the cluster and perf drops off with a few disk fetches and a very large disk write queue.