Slow performance of ACFS for hosting Oracle CRM Siebel application server
Good day dear Sirs!
We are using GRID 19.8 on IBM AIX 7.1, where a diskgroup was created exclusively to accommodate an ACFS volume of approximately 1TB on it
This file system is assembled on two LUNs in normal redundancy "mirror" mode, stripe columns: 8, stripe width: 1048576, number of snapshots: 0, fine grain, each of these two LUNs is taken from a separate disk storage connected via FC.
All this is intended to host the Siebel application server, which holds open hundreds of thousands of files, in addition, the so-called SiebelFS is located there - a directory with tens of millions of files.
The problem is the huge amount of disk I/O compared to using conventional hard drives (no clustering) and very poor CRM performance on ASFS.