Description
I'm noticing a significant (10x) imbalance between two identical nodes in the query service when running the following command:
cbc-n1qlback -U couchbase://ec2-54-151-7-224.us-west-1.compute.amazonaws.com/events -f queries -t 1 -vvvv 2> n1qlback
"queries" has just one query in it:
{"statement":"SELECT * FROM \`travel-sample\` USE KEYS a"}I see one node getting ~1k queries/sec with the other only getting 100.
The attached logs show the imbalance as well:
[root@ip-10-249-138-61 /]# grep o -e "ec2.*:8093" n1qlback | sort | uniq -c
40746 ec2-50-18-69-64.us-west-1.compute.amazonaws.com:8093
5118 ec2-54-176-81-47.us-west-1.compute.amazonaws.com:8093
[root@ip-10-249-138-61 /]#
Attachments
For Gerrit Dashboard: CCBC-611 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
51375,3 | CCBC-611: Add better random host selection algorithm | master | libcouchbase | Status: MERGED | +2 | +1 |