OT: Network latency benchmarks or tests ? (re NFS)
From: kyle Hailey <kylelf_at_gmail.com>
Date: Thu, 12 May 2011 11:27:29 -0700
Message-ID: <BANLkTikQ==DVqPjmRvG8R5+dXnZEu2S34g_at_mail.gmail.com>
Bit off topic, but I haven't been able to track down a good network latency test for timing of block transfers off of NFS mounts. There are some good network throughput test tools, most notably netio ( http://www.ars.de/ars/ars.nsf/docs/netio<http://www.google.com/url?q=http%3A%2F%2Fwww.ars.de%2Fars%2Fars.nsf%2Fdocs%2Fnetio&sa=D&sntz=1&usg=AFrqEzdVNzyguB3g4dOsd2BOx11KCCOz4Q>) which I've been using,
but I'm surprised at the seeming lack of latency tests. Sure there is ping, but ping only goes through the ICMP layer and there is not much in the way of configuration.
There is a program called nfs_ping (https://github.com/mprovost/NFSping <%20https://github.com/mprovost/NFSping>) which at least goes through to the NFS layer but what I'm looking for is a program that does just a ping pong of data between to machines and tracks timings and is configurable to set the size of the data being ping ponged back and forth.
Date: Thu, 12 May 2011 11:27:29 -0700
Message-ID: <BANLkTikQ==DVqPjmRvG8R5+dXnZEu2S34g_at_mail.gmail.com>
Bit off topic, but I haven't been able to track down a good network latency test for timing of block transfers off of NFS mounts. There are some good network throughput test tools, most notably netio ( http://www.ars.de/ars/ars.nsf/docs/netio<http://www.google.com/url?q=http%3A%2F%2Fwww.ars.de%2Fars%2Fars.nsf%2Fdocs%2Fnetio&sa=D&sntz=1&usg=AFrqEzdVNzyguB3g4dOsd2BOx11KCCOz4Q>) which I've been using,
but I'm surprised at the seeming lack of latency tests. Sure there is ping, but ping only goes through the ICMP layer and there is not much in the way of configuration.
There is a program called nfs_ping (https://github.com/mprovost/NFSping <%20https://github.com/mprovost/NFSping>) which at least goes through to the NFS layer but what I'm looking for is a program that does just a ping pong of data between to machines and tracks timings and is configurable to set the size of the data being ping ponged back and forth.
On a site note, I've been using happy with fio ( http://freshmeat.net/projects/fio/<http://www.google.com/url?q=http%3A%2F%2Ffreshmeat.net%2Fprojects%2Ffio%2F&sa=D&sntz=1&usg=AFrqEze5L3Tt-qcw2rwQ0tokz_IxljGTPg>) for disk I/O testing. I tried IOZone, bonnie++ and Orion but had issues with all of them. Interestingly Orion seems to have a bug, in some situations, where the random single block reads are not so random and end up rereading a small subset of the block range specified.
-- http://www.freelists.org/webpage/oracle-lReceived on Thu May 12 2011 - 13:27:29 CDT