使用默认的三副本方式建表,提示存活节点不够。 MySQL [dbtest]> CREATE TABLE table1 ( siteid INT DEFAULT '10', citycode SMALLINT, username VARCHAR(32) DEFAULT '', pv BIGINT SUM DEFAULT '0' ) AGGREGATE KEY(siteid, citycode, username) DISTRIBUTED BY HASH(siteid) BUCKETS 10 ; ERROR 1050 (42S01): Failed to find enough alive backends. need: 3 MySQL [dbtest]> SHOW PROC '/backends'; +-----------------+-----------+---------------+----------+---------------+--------+----------+---------------------+---------------------+-------+----------------------+-----------------------+-----------+ | Cluster | BackendId | IP | HostName | HeartbeatPort | BePort | HttpPort | LastStartTime | LastHeartbeat | Alive | SystemDecommissioned | ClusterDecommissioned | TabletNum | +-----------------+-----------+---------------+----------+---------------+--------+----------+---------------------+---------------------+-------+----------------------+-----------------------+-----------+ | | 10376 | 192.168.5.100 | FineBI1 | 10050 | 9060 | 10040 | 2017-08-17 11:00:44 | 2017-08-17 11:13:09 | true | false | false | 0 | | example_cluster | 10001 | 192.168.5.105 | FineBI2 | 9050 | 9060 | 8040 | 2017-08-17 11:00:54 | 2017-08-17 11:13:09 | true | false | false | 37 | | example_cluster | 10002 | 192.168.5.106 | FineBI3 | 9050 | 9060 | 8040 | 2017-08-17 11:00:59 | 2017-08-17 11:13:09 | true | false | false | 37 | +-----------------+-----------+---------------+----------+---------------+--------+----------+---------------------+---------------------+-------+----------------------+-----------------------+-----------+ 3 rows in set (0.00 sec)
Activity
chenhao7253886 commentedon Aug 17, 2017
你的cluster内只有两个be啊,所以你只能创建两副本。palo为了保证高可用在每一个host上只会放一个副本,所以你两个be在两台机器上最多创建两副本的表。
chenhao7253886 commentedon Aug 17, 2017
你可以扩容一个节点来创建三副本的,首先在另外一个机器上部署另外一个be,然后通过alter system add backend加进集群,然后通过alter cluster cluster_name properties("instance_num" = "3") 扩容cluster的节点到三个。
zlcnju commentedon Aug 17, 2017
有三个be,上面有列出来,分别是在finebi1,finebi2,finebi3三台机器上,而且都是存活状态
chenhao7253886 commentedon Aug 17, 2017
需要加进你的cluster啊,上面给你说了方法了。show proc "/backends" 注意下cluster这一栏,你的cluster内有几个be,只有两个需要把另外一个再加进去执行alter cluster cluster_name properties("instance_num" = "3")即可
1. Fix the bug of sort node return empty block if child eos is true a…
1. Fix the bug of sort node return empty block if child eos is true a…
1. Fix the bug of sort node return empty block if child eos is true a…
[fix](olapscanner) fix coredump caused by concurrent acccess of olap …
[fix](olapscanner) fix coredump caused by concurrent acccess of olap …
[fix] recover inToMinmaxParitionRewriteThreshold (apache#23)
[Bug](backend-options) fix use after free on BackendOptions::get_loca…
[Bug](backend-options) fix use after free on BackendOptions::get_loca…
[Bug](backend-options) fix use after free on BackendOptions::get_loca…
[Fix](segment iterator) fix shrink non-char column coredump (#36275)
98 remaining items
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](be) Avoid scan executor shutdown use after free (#65220)
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](catalog) catalog-SPI abstraction-mismatch review: paimon time-t…
[fix](be) Avoid scan executor shutdown use after free (#65220)
[fix](fe) Preserve IVM fallback stream offsets (apache#23)
[fix](fe) Preserve IVM fallback stream offsets (apache#23)
[fix](fe) Preserve IVM fallback stream offsets (apache#23)
[fix](fe) Preserve IVM fallback stream offsets (apache#23)
[fix](fe) Preserve IVM fallback stream offsets (apache#23)