|
View:
New views
5 Messages
—
Rating Filter:
Alert me
|
|
|
NFS - network read / write throughput per protocolHi
there, is
there a possibility to get the read / write throughput in kilobyte per protocol
(separated NFS / CIFS) on a filer? I
found only the operations but nothing else. Thanks
and regards, Steffen
|
|
|
RE: NFS - network read / write throughput per protocolStats show with the –e
will let you use expression matching. stats
show -i 15 -e volume:(nfs|vmware:shares):(nfs) Would report all of
the NFS traffic on volumes with vmware or shares in the name. There is a ton more it
can do, check out the man page for more. From:
owner-toasters@... [mailto:owner-toasters@...] On Behalf Of Steffen Kammerer Hi there, is there a possibility to get the read / write throughput in
kilobyte per protocol (separated NFS / CIFS) on a filer? I found only the operations but nothing else. Thanks and regards, Steffen
Please be advised that this email may contain confidential information. |
|
|
Re: NFS - network read / write throughput per protocolI don't see a counter for what you are looking for in 7.3.1.1, but you
could add up all the volume stats for nfs_read_data and nfs_write_data and cifs_read_data/cifs_write_data. Add that all together, to get the total nfs throughput / sec and cifs throughput per second. It sounds like a good RFE for something under system:system:* -Blake On Thu, Oct 29, 2009 at 9:37 AM, Steffen Kammerer <steffen.kammerer@...> wrote: > Hi there, > > > > is there a possibility to get the read / write throughput in kilobyte per > protocol (separated NFS / CIFS) on a filer? > > > > I found only the operations but nothing else. > > > > Thanks and regards, > > > > Steffen |
|
|
|
|
|
Re: NFS - network read / write throughput per protocol
Steffen,
Here is something I use quite ofter from a unix box. However this is split per volume. If you want the whole system output you need to sdo some maths and add all the volumes together. (miso:samuell/home/samuell> ssh samuellf820 "priv set diag; stats show -i 1 -r volume " | grep ccase | grep "_data" volume:ccase:read_data:0b/s volume:ccase:write_data:0b/s volume:ccase:nfs_read_data:0b/s <--- NFS read volume:ccase:nfs_write_data:0b/s <--- NFS write volume:ccase:cifs_read_data:0b/s <-- cifs volume:ccase:cifs_write_data:0b/s <-- cifs volume:ccase:san_read_data:0b/s volume:ccase:san_write_data:0b/s volume:ccase:flexcache_read_data:0b/s volume:ccase:flexcache_write_data:0b/s volume:ccase:flexcache_send_data:0b/s volume:ccase:flexcache_receive_data:0b/s -i = seconds ccase = volume name in question ( remove grep ccase to get all volumes) hope it helps Blake Golliher wrote: There's no specific counter for bytes / sec per protocol in priv set diag. For iops and latency's and size histograms, but nothing specific. -Blake On Thu, Oct 29, 2009 at 1:54 PM, Maxwell Reid slinkywizard@... wrote: |
| Free embeddable forum powered by Nabble | Forum Help |