<解决中><请教>disk busy 100% 解决及调优

Nikolai

原帖由 vanvesee 于 2008-12-28 11:45 发表
mServer1/#ioo -a
         j2_atimeUpdateSymlink = 0
j2_dynamicBufferPreallocation = 16
             j2_inodeCacheSize = 400
           j2_maxPageReadAhead = 128
             j2_maxRandomWrit ...

可以考虑j2_nBufferPerPagerDevice=2048
# ioo -po j2_nBufferPerPagerDevice=2048

# ioo -h j2_nBufferPerPagerDevice
Help for tunable j2_nBufferPerPagerDevice:
Specifies the number of file system bufstructs for Enhanced JFS. Default: 512. File system must be remounted. This tunable only specifies the number of bufstructs that start on the paging device. Enhanced JFS will allocate more dynamically. Ideally, this value should not be tuned, and instead j2_dynamicBufferPreallocation should be tuned. However, it may be appropriate to change this value if, when using vmstat -v, the number of "external pager filesystem I/Os blocked with no fsbuf" increases quickly (and continues increasing) and j2_dynamicBufferPreallocation tuning has already been attempted. If the kernel must wait for a free bufstruct, it puts the process on a wait list before the start I/O is issued and will wake it up once a bufstruct has become available. May be appropriate to increase if striped logical volumes or disk arrays are being used.

一般来说, 对于磁盘阵列, 提高该值是有帮助的, 具体可以分析为
# vmstat -v |grep "external pager"
如果很busy的时候, external pager增加得很快, 你就有必要增加j2_nBufferPerPagerDevice.

另外, 从vmstat来看, memory应该是足够的.
当idle<10时, 线程一般是大于16, 而且user使用的cpu不高, 说明这个时候cpu都在忙于系统, 而不是在干活. 可以考虑优化应用, 减少线程. 当然io也是一个问题..