Hi,
I want to know why a bulk insert results are far slower than benchmark.
I'm using SQLIO to benchmark my SAN, and I can get between 400MB/s and 1000MB/sec
but a bulk insert results in about 20MB/sec;
I know a benchmark is not a real usage, but 20 times faster than the bulk insert, its a huge difference.
the server is in cluster with 16 cores and 64GB of RAM.
I do the same load test on a VM, where the disks are slower, less RAM, CPU etc... and I got the same bulk insert performance.
the VM is stored on the SAN too, but I know the host server is slower than my dedicated SQL server (the disks are 2 times slower for the same benchmark)
what could be the reason for this?
why both servers performs the same way while 1 is supposed to be at least 3 times faster?
and why so slow compared to the benchmark results?
I try to understand the environment, but I'm not able to find good reasons and also I have a lack of support from the guys in charge of the env.
thanks for your comments!