mongodb分片奔溃
小瑾 2017-03-09 05:13:41 因为资源紧张,一开始使用的是两台服务器,最近加了第3台服务器,很多集合的数据已经上亿,遇到的问题:
1、当数据进行查询的时候,我看了下有42条线程同时在执行查询操作,所有的写入都停止了,有没有办法解决这个问题,在没有查询的情况下,其实每秒的插入量应该到4K+的
insert query update delete getmore command flushes mapped vsize res faults qr|qw ar|aw netIn netOut conn set repl time
*0 *0 *0 *0 0 6|0 0 2.1G 177.0M 0 0|0 0|0 766b 9k 27 RTR 2017-03-09T14:26:41+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:42+08:00
*0 *0 *0 *0 0 2|0 0 2.1G 177.0M 0 0|0 0|0 201b 8k 27 RTR 2017-03-09T14:26:43+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:44+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:45+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:46+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:47+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:48+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:49+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:50+08:00
insert query update delete getmore command flushes mapped vsize res faults qr|qw ar|aw netIn netOut conn set repl time
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:51+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:52+08:00
*0 *0 *0 *0 0 2|0 0 2.1G 177.0M 0 0|0 0|0 201b 8k 27 RTR 2017-03-09T14:26:53+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:54+08:00
*0 *0 *0 *0 0 6|0 0 2.1G 177.0M 0 0|0 0|0 762b 8k 27 RTR 2017-03-09T14:26:55+08:00
*0 *0 *0 *0 0 1|0 0 2.1G 177.0M 0 0|0 0|0 143b 8k 27 RTR 2017-03-09T14:26:56+08:00
1265 *0 *0 *0 0 190|0 0 2.1G 177.0M 0 0|0 0|0 2m 22k 27 RTR 2017-03-09T14:26:57+08:00
2、发现经常出现主分片宕掉,比如昨天:是因为移动分片的时间过长了吗?
2017-03-08T12:23:12.871+0800 I WRITE [conn490] write request to old shard version 1887|1||5896e9b370956d953a47dfe1 waiting for migration commit
2017-03-08T12:23:12.871+0800 I SHARDING [conn490] Waiting for 10 seconds for the migration critical section to end
2017-03-08T12:23:12.871+0800 I - [conn76]