日常数据库检查时，发现POSX RAC DB第二台服务器 RHEL5.6 Linux出现如下信息：
DB Alert
Log:
|
Wed Dec 14 15:04:15 CST 2016
Thread 2 advanced to log sequence 43473 (LGWR switch)
Current log# 5 seq# 43473 mem# 0:
+DG1/kkxddb/redo05.log
Wed Dec 14 15:41:18 CST 2016
Could not open audit file:
/oracle/admin/kkxddb/adump/kkxddb2_ora_14443_5.aud
Retry Iteration No: 1 OS Error:
28
Retry Iteration No: 2 OS Error:
28
Retry Iteration No: 3 OS Error:
28
Retry Iteration No: 4 OS Error:
28
Retry Iteration No: 5 OS Error:
28
Wed Dec 14 15:41:23 CST 2016
OS Audit file could not be created; failing after 5 retries
Could not open audit file:
/oracle/admin/kkxddb/adump/kkxddb2_ora_14443_5.aud
Retry Iteration No: 1 OS Error:
28
Retry Iteration No: 2 OS Error:
28
Retry Iteration No: 3 OS Error:
28
Retry Iteration No: 4 OS Error:
28
Retry Iteration No: 5 OS Error:
28
|
OS
Messages:
|
Dec 14 10:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,10:59:51 132861158 Seconds
Dec 14 11:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,11:59:51 132864758 Seconds
Dec 14 12:30:01 xddb2 kernel: EXT3-fs warning (device dm-0):
ext3_dx_add_entry: Directory index full!
Dec 14 12:30:11 xddb2 last message repeated 49 times
Dec 14 12:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,12:59:51 132868358 Seconds
Dec 14 13:15:01 xddb2 auditd[11709]: Audit daemon rotating log
files
Dec 14 13:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,13:59:51 132871957 Seconds
Dec 14 14:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,14:59:51 132875557 Seconds
Dec 14 15:41:18 xddb2 kernel: EXT3-fs warning (device dm-0):
ext3_dx_add_entry: Directory index full!
Dec 14 15:41:27 xddb2 last message repeated 49 times
Dec 14 15:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,15:59:51 132879157 Seconds
Dec 14 16:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,16:59:51 132882757 Seconds
Dec 14 17:59:51 xddb2 MR_MONITOR[23415]: Controller ID:
0 Time established since power on: Time
2016-12-14,17:59:51 132886357 Seconds
|
此问题为inodes(索引节点)已满,引用"inode译成中文就是索引节点，每个存储设备（例如硬盘）或存储设备的分区被格式化为文件系统后，应该有两部份，一部份是inode，另一部份是Block，Block是用来存储数据用的。而inode呢，就是用来存储这些数据的信息，这些信息包括文件大小、属主、归属的用户组、读写权限等。inode为每个文件进行信息索引，所以就有了inode的数值。操作系统根据指令，能通过inode值最快的 找到相对应的文件。"
后分析应该是/oracle
inode用完了。
实际应用需要要根据分区的大小来定，造成此问题通常是产生了大量的小文件。
#df –i
处理方法：删除shell脚本产生的多余的数据库aduit
file。
可参考：http://qiuyt.blog.51cto.com/1229789/1424356
加载中，请稍候......
加载中…
喜欢