On Fri, 24 Mar 2006 09:43:40 -0800 Steve Lamb <grey@dmiyu.org> wrote: > Andrew Sackville-West wrote: > > well, Steve, to be honest, I was offended by the responses you were giving > > to everyone's attempts to help you. Maybe its a simple misunderstanding, > > but here's my take on it. I put up the first response telling you "google > > is your friend" and pointing you to a debian admin article on traffic > > shaping. > > Which is not what Netlimiter does and something I described in my initial > post. Traffic shaping allows for queue jumping of some packets over others > and the limitation of other packets based on protocol data which has little to > no bearing at the application level. yes. however, nowhere do you say that traffic shaping isn't what you want. I suppose its implied in your desire for a solution that limits bandwidth by application (which pyshaper seems to do by the way, have you checked that? http://www.freenet.org.nz/python/pyshaper/), but it is only implied and bearing the failures of email in mind, I think its unreasonable to expect us to catch that specifically. > > Netlimiter, on the other hand, limits bandwidth based on the application, > not on information at the network layer. IE, if an application uses random > ports for communication traffic shaper is next to useless because, at the > basic level (something I have configured, thank you) it deals in ports; > *static* ports. > > > This is a subject that I don't know much about, but I'm interested > > to learn so I did a little research. a google of traffic shaping and debian > > came up with tons of responses that all looked useful. > > But did you start off with a search on "Netlimiter" to see what *it* did > first? Why should I? I don't use windows. And your requirement was pretty clear in that you wanted to shape traffic by program. My suggestion to look into that traffic shaping article was just that, a suggestion, take it or leave it. look at it and say to yourself, Oh, I've seen that before, oh well. > Because you know little of traffic shaping and because you wanted to > look into traffic shaping means that *I* knew little and that *I* was looking > for traffic shaping. I formed my question with specifics. huh? how does my knowledge of and desire to learn of traffic shaping have anything to do with your knowledge of traffic shaping? > > > Then you went on to question whether *I* had done adequate > > research into *YOUR* problem. This was followed up with repeated trashing > > of those tools that are out there. > > No, I asked if you had done a search on what I was asking about? Here it > is with flippant remarks and misaddressed answers and I still doubt you've > opened up a browser, hit Google and searched on "Netlimiter". nothing flippant intended. Sorry if it came off that way. I googled "debian traffic shaping" and found good results on the first page. granted I didn't provide my search terms, which I should have done. These misaddressed answers are based on you assuming that we would make assumptions about what you have already done from what you have written in your first post. (phew) thats a lot of assuming. > > > Sorry man, but that just rubbed me the wrong way. Your responses to others > > have been in the same vein. I'm sorry if we haven't been of any help to > > you. We tried. We made some suggestions on where to look and tried to > > provide a solution that we thought fit your requirements. > > No, only one response came close. None of them fit my requirements which > were stated plainly up front. Is there a tool like Netlimiter for Linux? > Answer, no. > > > You mentioned > > apache, so one respondent put up an apache solution. Doesn't work for you > > and that's fine, but its not his responsibility to go find *YOUR* ultimate > > solution, but only to provide what help he deems appropriate. > > But there's the problem. Look what I wrote again... > > 'I want to be able to tell Apache (and other non-self-limiting programs) "you > get this much, NO MORE!"' > > Key phrase *and /other/ non-self-limiting programs*. So tossing out a > solution for Apache, which was just an example of a larger problem, didn't > address much, did it? your key-phrase is parenthetical and you stated > The primary culprit right now is Apache. it is totally reasonable for someone to make an Apache recommendation. That was your primary culprit. As I read it, the "other non-self-limiting programs" would be a nice bonus, but you're really looking for a solution to your apache problem. Again, its just not clear communication. I know to you it looks that way, but to me it doesn't. > > > FTR, I'm just laying it out as I see it. I'm human and fallible and do not > > mean to offend. I'm trying to explain my interpretation of what is likely a > > big communication breakdown. > > Trust me, it's a common breakdown. I mean, here's another good example. > How many people pointed me to traffic shaping? Lots. Did any of them read > the first line of the second paragraph? > > "Yeah, I could do it with iptables and a whole slew of deep magic. " > > Hm, tell me again what traffic shaping is partially configured in? Right, > iptables. *sigh* In your original post you say > Any similar utility, **even if it is just a front for iptables, for Debian?** (emphasis mine) So which is it. traffic shaping or not? in your original post you seem to imply that a front for iptables is okay but that you also don't want to do it with iptables. Now you imply that iptables is not a solution? > > Pardon me for being frustrated when people offer suggestions that I have > looked at and found wanting and indicated I looked at them and found them > wanting in my initial post. Pardon me for being frustrated when I ask about a > large scale problem and someone gives me an answer for one small piece which > doesn't help elsewhere. Pardon me for presuming people should, you know, read > the message first before replying? well, see now you're attacking. I did read the message, and its really rude to suggest otherwise. In two responses I gave you a pointer to a relevant, though sorry, not perfect, reference, made a suggestion to look into google -- something you never mentioned doing, and third suggested a package that looked like it would do exactly what you wanted. What more do you want? Of course you are frustrated by our suggestions that you "have looked at and found wanting". Nowhere do you tell us what you've looked at and found wanting. Where do you say "I've read the debian traffic shaping articles, I've read man tc, I've looked into pyshaper etc and none of these work for me." ALright, I'm done with this. If we were in the same town I'd take you out for a pint (if appropriate) so we could sweep it under the table. I'm sorry that we've had this breakdown. cheers. A > > -- > Steve C. Lamb | I'm your priest, I'm your shrink, I'm your > PGP Key: 8B6E99C5 | main connection to the switchboard of souls. > -------------------------------+--------------------------------------------- > >
Attachment:
pgpruEJAoviPS.pgp
Description: PGP signature