Unanswered: Different responses between UNIX machine
I have a Solaris (2.8) machine UltraAX-i2 which has one internl disk – (I’ll refer to it as machine A).
I have also a Solaris (2.9) machine Sun-Fire-280R which has two internal disks – (I’ll refer to it as machine B).
It is common that machine B is much stronger with much stroger than machine A from CPU and disks perspective.
I have a schema creation script that includes ~3000 sements (tables and indexes).
The script creation on machine A takes ~7 min, and on machine B it takes ~15min.
I have checked for wait events for the session running the creation script and saw ‘log file sync’ wait event for both machines.
I have also tried various sizes of log_buffers for machine B but nothing seems to help.
When checking machines I/O (using iostat –Mxnz 2 1000) I see that while creating the objects machine A writes 200 writes with total of ~0.1Mbytes per second and machine B writes 40 writes with total of ~0.2Mbytes per second.
Also while creating the segments the cpu of process oracle.exe on machine A is ~20% and machine B is ~3%
I brought my unix sysadmin who checked the machine and found no problems plus installed all last patches required.
All indicates that the Oracle is blocked by something, I can not find what, and need your help.
Maybe there is a parameter in Oracle or O/S that needs to be configured.