Actually, it is not a list.. but just something that''ll be very useful to me in my business. If crossbow/virtual interfaces can provide a flow that can capture bandwidth usage over a period of time (including 95 percentile), that''ll be great! Is this possible? -- This message posted from opensolaris.org
Anil: Try acctadm(1m) (type ''net'') to enable link (basic) or extended (flow) stats logging. Then using dladm show-usage or flowadm show-usage with the log file to look at usage for a specified time duration. -venu On Tue, 19 May 2009, Anil wrote:> Actually, it is not a list.. but just something that''ll be very useful to me in my business. If crossbow/virtual interfaces can provide a flow that can capture bandwidth usage over a period of time (including 95 percentile), that''ll be great! > > Is this possible? > -- > This message posted from opensolaris.org > _______________________________________________ > crossbow-discuss mailing list > crossbow-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/crossbow-discuss >
Cool, thanks. That''ll do it. BTW, shouldn''t the column say IBYTES instead of RBYTES? root at vps1:~# dladm show-usage -f /var/account.info LINK DURATION IPACKETS RBYTES OPACKETS OBYTES BANDWIDTH vnic1 2180 4713 280952 57 2842 0.001 Mbps vnic0 2180 5578 886177 939 611206 0.005 Mbps e1000g0 2180 47251 5939057 42730 38342186 0.162 Mbps -- This message posted from opensolaris.org
I think the nomenclature (ipackets, rbytes) is borrowed from kstat which existed before. Moving forward, dladm and flowadm will retain configuration specific commands while two new commands dlstat and flowstat will provide interface to query dynamic network traffic statistics. I will make the naming consistent while implementing dlstat/flowstat. ~ Shri Anil wrote:> Cool, thanks. That''ll do it. > BTW, shouldn''t the column say IBYTES instead of RBYTES? > > root at vps1:~# dladm show-usage -f /var/account.info > LINK DURATION IPACKETS RBYTES OPACKETS OBYTES BANDWIDTH > vnic1 2180 4713 280952 57 2842 0.001 Mbps > vnic0 2180 5578 886177 939 611206 0.005 Mbps > e1000g0 2180 47251 5939057 42730 38342186 0.162 Mbps
One thing I would love to see is it calculate 95 percentile. That is how most service providers calculate how much traffic is being used. Basically you take 5 minute averages, throw the top 5% out and then calculate the average Mbps. -- This message posted from opensolaris.org