difference between LWT/Paxos queries's timestamp and normal queries' timestamp in cassandra -
when using cassandra, if execute following 2 cql, second 1 not success.
insert job_info (id, create_time) values (5be224c6-8231-11e7-9619-9801b2a97471, 0) if not exists; insert job_info (id, create_time) values (5be224c6-8231-11e7-9619-9801b2a97471, 1);
the keyspace , table this:
create keyspace scheduler replication = {'class': 'simplestrategy', 'replication_factor': '1'} , durable_writes = true; create table scheduler.job_info ( id timeuuid primary key, create_time int ) caching = {'keys':'all', 'rows_per_partition':'none'};
i posted issue on https://issues.apache.org/jira/browse/cassandra-13767, , commant said that:
you're mixing lwt/paxos queries ({{if not exists}}, deals timestamp in own way, "normal" queries, , known lead unpredictable results in term of ordering (what happens here not 2nd insert fails, it's gets assigned timestamp lower 1st query , ignored). long story short, don't mix lwt/paxos queries other queries (at least on same partition) , you'll avoid issue.
my question is, difference between lwt/paxos queries , "normal" queries' timestamp?
the first cql , second cql can executed interval of many seconds, second 1 still not success. , problem caused timestamp?
Comments
Post a Comment