Re: NetBSD 6.1 NFS server performances

[Sad Clouds <cryintothebluesky%googlemail.com@localhost>]

Can you capture network traces then look at NFS protocol with
wireshark. Most optimal way for NFS client to work would be to issue
multiple WRITE requests, followed by a single COMMIT request for
all of those WRITE requests. If you see each WRITE followed by a COMMIT,
then this can introduce latencies, as it's not fully pipelining WRITE
requests.

Also, where is NFS server writing to? Are you sure local disk is not
the bottleneck? Tried writing via NFS to ramdisk?


On Thu, 3 Dec 2015 13:48:18 +0100 (CET)
"Emile `iMil' Heitor" <imil%home.imil.net@localhost> wrote:

> 
> Continuing on the NAS performances topic, now it's NFS server's turn.
> First things first, I've checked both network and disk throughput,
> neither cause a bottleneck:
> 
> tatooine is the client
> coruscant is the server
> 
> $ iperf -c coruscant -p 2828 -t 10
> ------------------------------------------------------------
> Client connecting to coruscant, TCP port 2828
> TCP window size: 43.8 KByte (default)
> ------------------------------------------------------------
> [  3] local 192.168.1.1 port 51371 connected with 192.168.1.2 port
> 2828 [ ID] Interval       Transfer     Bandwidth
> [  3]  0.0-10.0 sec  1.06 GBytes   908 Mbits/sec
> 
> $ dd if=/dev/zero bs=1024K count=1000 | nc -v coruscant 2828
> Connection to coruscant 2828 port [tcp/*] succeeded!
> 1000+0 records in
> 1000+0 records out
> 1048576000 bytes (1.0 GB) copied, 9.27363 s, 113 MB/s
> 
> Gigabit link, all clear.
> 
> Now using NFS:
> 
> $ dd if=/dev/zero bs=1024K count=1000
> >Desktop/nfs@coruscant/imil/tmp/test 1000+0 records in
> 1000+0 records out
> 1048576000 bytes (1.0 GB) copied, 51.8476 s, 20.2 MB/s
> 
> I know, right? And yes results are identical with differents bs
> values.
> 
> I've tried a bazillion NFS options on the clients (TCP, UDP, {r,w}
> size from 8192 to 64k...), tried many OSes as a client, the NFS
> results are consistent, always between 20 and 30MB/s.
> NFS server is started via rc.d with the following rc.conf variables:
> 
> rpcbind=YES
> mountd=YES
> nfs_server=YES
> nfsd_flags="-6tun 8"
> lockd=YES
> statd=YES
> 
> And yes I tried increasing or reducing thread number.
> 
> /etc/exports is pretty simple:
> 
> /export -alldirs -noresvport -maproot=root:wheel -network
> 192.168.1.0/24
> 
> I've read an extensive number of "NetBSD NFS server performances"
> posts here, applied every suggestion without any luck, any idea would
> be highly appreciated.
> 
> Thanks,
> 
> ----------------------------------------------------------------
> Emile `iMil' Heitor * <imil@{home.imil.net,NetBSD.org,gcu.info}>
>                                                                _
>              | http://imil.net        | ASCII ribbon campaign ( )
>              | http://www.NetBSD.org  |  - against HTML email  X
>              | http://gcu.info        |              & vCards / \
>