Hi,
I am working on a monitoring software which get metrics from many system command (df, netstat, iostat, ...). It works well on Linux but some metrics are missing on FreeBSD.
My problem concerns disk latency: on Linux I am using [cmd=""]iostat -x -d 2 1[/cmd] which give me everything I need. On FreeBSD I am using the same command, however it is slightly different. Some columns are equivalent (%util on Linux is %b, rkB/s is kr/s, ...) but others are not (await, svtm, avgqu_sz, svc_t and wait).
My issue is to find the equivalent of await. The best candidate seems to be svc_t. But is it exactly the same measure? What does svc_t include? (queue time? seek time? transfer time?).
The question is a bit specific and I found nobody answered it yet.
Thank you in advance.
I am working on a monitoring software which get metrics from many system command (df, netstat, iostat, ...). It works well on Linux but some metrics are missing on FreeBSD.
My problem concerns disk latency: on Linux I am using [cmd=""]iostat -x -d 2 1[/cmd] which give me everything I need. On FreeBSD I am using the same command, however it is slightly different. Some columns are equivalent (%util on Linux is %b, rkB/s is kr/s, ...) but others are not (await, svtm, avgqu_sz, svc_t and wait).
My issue is to find the equivalent of await. The best candidate seems to be svc_t. But is it exactly the same measure? What does svc_t include? (queue time? seek time? transfer time?).
The question is a bit specific and I found nobody answered it yet.
Thank you in advance.