Hello,
iftop is fantastic tool! Thanks for it.
On the other hand, I found little strange when it's scale ends at 977kb with traffic in the ranges 256+ .. ~1000 kb and then jumps right to 9.54Mb. If someone has 1-1.5Mbit (and I'm one of such), then suddenly all the graphs become very small. I tried to add more steps and at the same time move to base 1024.
Attached patch solves these problems. Unfortunately, I had no setup to test correctness of this patch for speeds above 2Mbps.
On Mon, 14 Nov 2005 17:23:49 +0200 Aidas Kasparas a.kasparas@gmc.lt wrote:
iftop is fantastic tool! Thanks for it.
On the other hand, I found little strange when it's scale ends at 977kb with traffic in the ranges 256+ .. ~1000 kb and then jumps right to 9.54Mb. If someone has 1-1.5Mbit (and I'm one of such), then suddenly all the graphs become very small. I tried to add more steps and at the same time move to base 1024.
Attached patch solves these problems. Unfortunately, I had no setup to test correctness of this patch for speeds above 2Mbps.
What about the option e.g. "-m 1536k"?
R.
richard lucassen wrote:
On Mon, 14 Nov 2005 17:23:49 +0200 Aidas Kasparas a.kasparas@gmc.lt wrote:
iftop is fantastic tool! Thanks for it.
On the other hand, I found little strange when it's scale ends at 977kb with traffic in the ranges 256+ .. ~1000 kb and then jumps right to 9.54Mb. If someone has 1-1.5Mbit (and I'm one of such), then suddenly all the graphs become very small. I tried to add more steps and at the same time move to base 1024.
Attached patch solves these problems. Unfortunately, I had no setup to test correctness of this patch for speeds above 2Mbps.
What about the option e.g. "-m 1536k"?
OK, that option works, but is not allways convienient: -- it requires either extra typing every time or use of extra script/alias; -- it is one extra thing to tell to a person to whom you introduce iftop; -- it is one more thing to maintain when one changes speeds (agree, not a everyday activity, but if you support a lot of people to whom you're not an ISP, and they tend to enter commands letter by letter, how you wrote them, catching wrong limit is a nuisance).
On the other hand, I do not propose significant/complicated code change in my patch. Infrastructure for automatic scaling is in present code, what I propose is to add several extra steps at better defined points.
What I would like to see more is an optional automatic scaling back, if traffic is significantly less than current scale for extended period of time.
On Tue, 15 Nov 2005 13:35:55 +0200 Aidas Kasparas a.kasparas@gmc.lt wrote:
What about the option e.g. "-m 1536k"?
OK, that option works, but is not allways convienient: -- it requires either extra typing every time or use of extra script/alias; -- it is one extra thing to tell to a person to whom you introduce iftop; -- it is one more thing to maintain when one changes speeds (agree, not a everyday activity, but if you support a lot of people to whom you're not an ISP, and they tend to enter commands letter by letter, how you wrote them, catching wrong limit is a nuisance).
use ~/.iftoprc, you'll find an example here:
http://tech3.cam.nl/download/linux/iftop/
And rename it to .iftoprc.
On the other hand, I do not propose significant/complicated code change in my patch. Infrastructure for automatic scaling is in present code, what I propose is to add several extra steps at better defined points.
What I would like to see more is an optional automatic scaling back, if traffic is significantly less than current scale for extended period of time.
AFAIK there is no scaling back, perhaps that would be a nice feature :)
R.
On Tue, Nov 15, 2005 at 01:35:55PM +0200, Aidas Kasparas wrote:
On the other hand, I do not propose significant/complicated code change in my patch. Infrastructure for automatic scaling is in present code, what I propose is to add several extra steps at better defined points.
This seems reasonable. I don't have much spare capacity to work on iftop ATM, but I've put the patch in my "iftop to-do" folder for when I do.
What I would like to see more is an optional automatic scaling back, if traffic is significantly less than current scale for extended period of time.
Yes, me too. Shouldn't be that hard to implement as iftop already calculates the peak BW within the last 40s. I've put it on my to-do list, but if someone beats me to it I'll be delighted to receive a patch.
Paul