On Feb 27, 2007, at 4:55 PM, Leon Koll wrote: > Hello, gurus > I need your help. During the benchmark test of NFS-shared ZFS file > systems at some moment the number of NFS threads jumps to the > maximal value, 1027 (NFSD_SERVERS was set to 1024). The latency > also grows and the number of IOPS is going down. > I've collected the output of > echo "::pgrep nfsd | ::walk thread | ::findstack -v" | mdb -k > that can be seen here: > http://tinyurl.com/yrvn4z > > Could you please look at it and tell me what's wrong with my NFS > server.
The NFS server is waiting on the underlying filesystem to complete the work it has been asked to do on behalf of the clients. The NFS server is structured such that each thread is acting on a single incoming NFS request. Therefore, you have a set of clients that can simultaneously submit 1024 requests. And with respect to latency, it is a feedback loop; the longer it takes the requests to be serviced the fewer number of of IOPs that are going to be achieved. So, it appears that some of the questions you have asked about ZFS behavior are relevant here. The NFS server is waiting on ZFS to complete the work, etc. Spencer
