今天发现数据库很慢,看到大量进程在等待log file switch (checkpoint incomplete),日志文件的使用情况如下:
SYS@TEST15>select group#,sequence#,bytes/1024/1024 "BYTES(M)",members,archived,status from v$log; GROUP# SEQUENCE# BYTES(M) MEMBERS ARC STATUS ---------- ---------- ---------- ---------- --- ---------------- 1 205 100 2 YES ACTIVE 2 206 100 2 NO CURRENT 3 204 100 2 YES ACTIVE SYS@TEST15>select group#,status,type,member from v$logfile where type='ONLINE' order by type,group#; GROUP# STATUS TYPE MEMBER ---------- ------- ------- ------------------------------------------------------------ 1 ONLINE /data/oradata/orcl/data/redo01a.log 1 ONLINE /home/oracle/data/orcl/redo01b.log 2 ONLINE /data/oradata/orcl/data/redo02a.log 2 ONLINE /home/oracle/data/orcl/redo02b.log 3 ONLINE /data/oradata/orcl/data/redo03a.log 3 ONLINE /home/oracle/data/orcl/redo03b.log
从上面可以看出,所有的日志文件都是ACTIVE或CURRENT状态,这说明日志文件太小,需增加其大小。
但在线重做日志不可以直接RESIZE,需要通过如下方式:
1)添加新的日志文件组,定义它的大小
SYS@TEST15>alter database add logfile group 4 ('/data/oradata/orcl/data/redo04a.log','/home/oracle/data/orcl/redo04b.log') size 500M reuse; Database altered.
2)持续切换日志,一直到Group 1的status=inactive,且archived=yes(如果为归档的话),如下所示:
SYS@TEST15>alter system switch logfile; SYS@TEST15>alter system checkpoint; SYS@TEST15>select group#,sequence#,bytes/1024/1024 "BYTES(M)",members,archived,status from v$log where group#=1; GROUP# SEQUENCE# BYTES(M) MEMBERS ARC STATUS ---------- ---------- ---------- ---------- --- ---------------- 1 214 100 2 YES INACTIVE3)删除Group 1
SYS@TEST15>alter database drop logfile group 1; Database altered.这样,我们就完成了用更大的Group 4取代原有的Group 1,至于Group 2, Group 3以此类推。
最后,还要记得到操作系统上把日志文件删除。