informix Checkpoint 问题

tstt

一台IBM225 scounix 506 informix731的服务器，共4个online硬盘，70个Chunks, 每天都处理大量数据，下面是在处理数据时的日志，看看Checkpoint  时间是否正常? 有问题改怎样调整。
Thu Jun 30 17:24:03 2005

17:24:03  Event alarms enabled.  ALARMPROG = '/usr/informix/etc/log_full.sh'
17:24:11  DR: DRAUTO is 0 (Off)
17:24:11  Informix Dynamic Server Version 7.31.UC5    Software Serial Number AAD
#J228943
17:24:11  Informix Dynamic Server Initialized -- Shared Memory Initialized.
17:24:11  Physical Recovery Started.
17:24:11  Physical Recovery Complete: 0 Pages Restored.
17:24:11  Logical Recovery Started.
17:24:14  Logical Recovery Complete.
   0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks

17:24:15  Onconfig parameter NUMAIOVPS modified from 2 to 20.
17:24:15  Dataskip is now OFF for all dbspaces
17:24:15  On-Line Mode
17:24:15  Checkpoint Completed:  duration was 0 seconds.
17:30:51  Checkpoint Completed:  duration was 16 seconds.
17:35:44  Checkpoint Completed:  duration was 53 seconds.
17:40:12  Checkpoint Completed:  duration was 54 seconds.
......
......
Fri Jul  1 00:01:02 2005

00:01:02  Checkpoint Completed:  duration was 69 seconds.
00:03:04  Checkpoint Completed:  duration was 67 seconds.
00:05:07  Checkpoint Completed:  duration was 67 seconds.
00:07:12  Checkpoint Completed:  duration was 69 seconds.
00:09:16  Checkpoint Completed:  duration was 69 seconds.
00:11:21  Checkpoint Completed:  duration was 68 seconds.
........
07:48:34  Checkpoint Completed:  duration was 72 seconds.
07:54:03  Checkpoint Completed:  duration was 71 seconds.
07:59:29  Checkpoint Completed:  duration was 71 seconds.
08:04:57  Checkpoint Completed:  duration was 71 seconds.
08:10:16  Checkpoint Completed:  duration was 71 seconds.

yu_xiao_64

修改ONCONFIG文件，可以把你当前的文件配置，贴出来大家共同探讨一下

yipy

看不出有什么问题啊，数据大的话，checkpoint这个时间也是正常的

d_y_liu

同意，5分钟一次基本正常

tstt

Checkpoint 的时间是否过长? 几十秒

zzjijun

执行Checkpoint时间过长了。可以调整LRU试试。
不过关键看楼主是什么应用，如果就是处理大批量数据的，不是立即等结果的OLTP那么就这样了。调优主要是看你目前满不满意系统性能，不是一定要看那个参数如何。