If you store your datafiles in filesystems each filesystem should have own log device or use INLINE log
Tune JFS log buffer
- ioo -p -o j2_dynamicBufferPreallocation=256 -o
- ioo -p -o pv_min_pbuf=4096
- vmo -p -o minperm%=3 -o maxperm%=90 -o maxclient%=90 -o lru_file_repage=0 -o strict_maxperm=0
If you delight in new storage re-balancing capabilities and using few number of big disks in volume groups:
- You should increase the depth of run queue of disks from 40 to max. 256
- You should increase num_cmd_elems parameter of FC adapters from 200 to max. 2048
- Be careful changing max_xfer_size from 0x100000 to 0x200000 it could generate FC adapter errors
Monitor your disk IO utilization, if it passes 50-60% increase the number of disks inside the volume group (whatever technical sales suggest)
Niciun comentariu:
Trimiteți un comentariu