cassandra - READ 消息在最后 5000 毫秒内被丢弃：2 个内部和 0 个跨节点。平均内部丢弃延迟：2556 毫秒

Question

应用程序使用 go-lang 驱动程序读取查询。像 select * from tab1 where id=123 and datehr='12-10-2019' 这样的查询；

每个分区有 36k 行。每行大小为 120 字节。每个分区大小为 4MB。

但我在 Cassandra 调试日志中获取信息，如下所示。以及应用程序读取超时问题。使用 fetch size 的应用程序 go-lang 驱动程序为 5000。

cassandra 版本是 3.11.4 java-version java 版本 "1.8.0_131" 操作系统版本：centos 7.3 Cpu cores: 16 RAM: 64GB HEAP: 16/16

INFO  [ScheduledTasks:1] 2020-01-20 13:45:32,095 MessagingService.java:1236 - READ messages were dropped in last 5000 ms: 2 internal and 4 cross node. Mean internal dropped latency: 3449 ms and Mean cross-node dropped latency: 3621 ms
INFO  [ScheduledTasks:1] 2020-01-20 13:45:32,129 StatusLogger.java:101 - system_schema.dropped_columns                 0,0
INFO  [ScheduledTasks:1] 2020-01-20 13:46:52,135 MessagingService.java:1236 - READ messages were dropped in last 5000 ms: 1 internal and 0 cross node. Mean internal dropped latency: 2922 ms and Mean cross-node dropped latency: 3449 ms
INFO  [ScheduledTasks:1] 2020-01-20 13:46:52,147 StatusLogger.java:101 - system_schema.dropped_columns                 0,0
INFO  [ScheduledTasks:1] 2020-01-20 13:51:32,157 MessagingService.java:1236 - READ messages were dropped in last 5000 ms: 2 internal and 0 cross node. Mean internal dropped latency: 2556 ms and Mean cross-node dropped latency: 0 ms
INFO  [ScheduledTasks:1] 2020-01-20 13:51:32,166 StatusLogger.java:101 - system_schema.dropped_columns                 0,0
INFO  [ScheduledTasks:1] 2020-01-20 13:52:47,169 MessagingService.java:1236 - READ messages were dropped in last 5000 ms: 0 internal and 2 cross node. Mean internal dropped latency: 0 ms and Mean cross-node dropped latency: 1270 ms
INFO  [ScheduledTasks:1] 2020-01-20 13:52:47,179 StatusLogger.java:101 - system_schema.dropped_columns                 0,0

标签： cassandra

Answer 1

我认为，这里解释了关于 drop 突变和 read 突变的相同内容。请参考下面 在 Cassandra 中删除读取和删除突变