Checkpoint loguniq 请教！
两台服务器，分别做dbimport导入同一个数据文件，结果机器配置差的花了一个小时，机器配置好的花了二个小时，看日志，慢的
00:01:21Checkpoint Completed:duration was 0 seconds.
00:01:21Checkpoint loguniq 275, logpos 0xc240a8
00:06:25Checkpoint Completed:duration was 0 seconds.
00:06:25Checkpoint loguniq 275, logpos 0xc96018
快的
12:52:50Checkpoint Completed:duration was 0 seconds.
12:52:50Checkpoint loguniq 1792, logpos 0x153a9098
12:57:56Checkpoint Completed:duration was 0 seconds.
12:57:56Checkpoint loguniq 1792, logpos 0x153f31f0
想请教一下Checkpoint loguniq 后面的数字为何相差这么大，他表示什么意思，是不是这个原因造成机器配置好的服务器反而慢。谢谢！ loguniq 是指逻辑日志的唯一序号~~用得久的数据库这个id会比较大~
从你这只能看到两个机子的checkpoint的间隔都是5分钟,还有就是 "慢的"那台服务器的数据库是更近安装的而已.. liaosnet好敬业好有速度啊！:lol: 昨晚又试了一次，发现 duration 的延迟很长，请各位看看
17:00:37Checkpoint Completed:duration was 150 seconds.
17:00:37Checkpoint loguniq 1, logpos 0xd78018
17:10:55Checkpoint Completed:duration was 313 seconds.
17:10:55Checkpoint loguniq 1, logpos 0xdea018
17:22:54Checkpoint Completed:duration was 414 seconds.
17:22:54Checkpoint loguniq 1, logpos 0xe4d510
17:28:37Checkpoint Completed:duration was 38 seconds.
17:28:37Checkpoint loguniq 1, logpos 0xe52018
17:34:49Checkpoint Completed:duration was 68 seconds.
17:34:49Checkpoint loguniq 1, logpos 0xe71018
17:51:51Checkpoint Completed:duration was 717 seconds.
17:51:51Checkpoint loguniq 1, logpos 0xf2c018
17:57:52Checkpoint Completed:duration was 56 seconds.
17:57:52Checkpoint loguniq 1, logpos 0xf5c018 回复 4# nicolasch
持续时间长: 如果你是在import的话,由于在逻辑日志里的记录较多,写回磁盘可以要慢些~
要不你在环境变量里加上 TRACECKPT=1 ,然后重启跟踪下检查点的时间在做什么.. 但是另外一台机器：
12:52:50Checkpoint Completed:duration was 0 seconds.
12:52:50Checkpoint loguniq 1792, logpos 0x153a9098
12:57:56Checkpoint Completed:duration was 0 seconds.
12:57:56Checkpoint loguniq 1792, logpos 0x153f31f0
为什么他的Checkpoint 时间都是0，是不是慢的那台参数设置有问题呢？ 回复 6# nicolasch
所以让你跟踪看下呀~~看它的等待时间是在那?是flush lru还是flush disk
页:
[1]