求助iostat！！！

fog30000

因为一块镜像坏掉了，现在更换硬盘后，用iostat －E 显示的还是原来的那块硬盘的信息，请问怎么才能在不重启机器的情况下刷新信息呢？？

sun_up

Find a solution (item in Sun website, but I don't have a support account to read the document.

http://sunsolve8.sun.com/search/ ... &repository=cds

风之幻想

devfsadm 刷新一下看看.

faintblue

Description:        Top

When a hard- soft- or transport error occurs the counters are set and
accumulated.

There are some situations where you would like to clear the counters without
rebooting:

- when replacing a disk on the fly
- when using alert software that "alerts" someone in case of a
  "hard error" <-- example


Document Body:        Top

the simple answer is:
You cannot clear the kernel counters without rebooting the system (fcal you
can, see last statement).

Explanation:
iostat scans the system for current usage on the fly. This info is cached
on the kernel. You could be able to "unload" a module from the kernel to
clear the counters, but this is not recommended by Sun.

This really counts for "SCSI" disks.

For "FCAL" disks however, the "WWN" is unique. When a fcal disk is
replaced, the "new WWN" is seen by the system, which makes it see the
replacement disk, so in this case the "kernel counters" should be reset.