I managed to run stress tool outside the cluster. I a;ready have basic table bank_transactions in bank keyspace on my cassa cluster.
bank.bank_transactions ( customerid text, year int, month int, id timeuuid, amount int, card text, status text, PRIMARY KEY ((customerid, year, month), id) ); This is my profile yaml file:
table: bank_transactions
columnspec:
- name: customerid
size: uniform(5..10)
population: uniform(1..10)
- name: year
size: fixed(4)
- name: month
size: fixed(2)
- name: timeuuid
cluster: uniform(20..40)
- name: amount
size: uniform(0..1000000)
- name: text
size: uniform(10..15)
- name: status
size: uniform(10..20)
insert:
# How many partition to insert per batch
partitions: fixed(2)
# How many rows to update per partition
select: fixed(1)/500
# UNLOGGED or LOGGED batch for insert
batchtype: UNLOGGED
queries:
customer:
cql: select * from bank_transactions where customerid = ? and year = ? and month = ? and id = ?
fields: multirow
#com.datastax.driver.core.exceptions.InvalidQueryException:
#Partition key parts: year, month, id, customerid must be restricted as other parts are
transactions:
cql: select customerid, amount, card, status, id from bank_transactions where id = ? and customerid = ? and year = ? and month = ?
fields: samerow
regularupdate:
cql: update bank_transactions set amount = ? where customerid = ? and year = ? and month = ? and id = ?
fields: samerow
#nestihalo se v casovem limitu, lze navysit pamet, casovy limit...
# updatewithlwt:
# cql: update bank_transactions set amount = ? where customerid = ? and year = ? and month = ? and id = ? IF status = ?
# fields: samerow
I run my stress
./cassandra-stress user profile=./stress_transactions.yaml ops\(insert=3,customer=3,transactions=2,regularupdate=1\) \
n=10000 -node 172.16.20.104,172.16.20.105,172.16.20.106 -port native=9042 thrift=9160 jmx=7199 -graph file=./transactions.html
I was expecting to have something like graph on cassandra.apache site
But the result is really weird:
Could you please help me, what is going wrong?
P.S. Do not mind distributions and so on, it is just playground...