This is why I wont be reporting any more bugs.
-
there is "assumptions" been made that things are due to errors I must have made myself
If I had a dollar for every person who stormed in here screaming about bugs that turned out to be either their own lack of understanding or user error, I could retire.
and I have even been called a liar
He didn't call you a liar, he said that something you said wasn't true. To lie you must try to deceive, and nobody thinks you're trying to trick us.
I was involved in that thread. You showed up without a clue about what you're doing, and then turn around later saying that you disagree with the pfSense devs because they don't seem to understand ALTQ as well as you do after your 3 entire hours of research. That level of arrogance stopped me from caring any further about your problem. You have yet to provide anything whatsoever to help illustrate or debug your issue, including your floating rules and general shaper config. Even your bug report has no info at all, and now you're arguing with the devs there.
-
As far as forum versus bug tracker… I personally prefer starting a discussion in a forum first, to determine if there is a bug or if it's just something I'm doing wrong. If it's determined that there is a bug, then creating a report in the bug tracker is justified. It keeps a lot of the simple determining if there's an issue out of the tracker. You can always link to a forum thread in the bug report.
-
pfSense is amazing software, but the attitude to bug reports I have made has been pretty shocking, there is "assumptions" been made that things are due to errors I must have made myself, and I have even been called a liar…
I agree, you were treated appallingly. And even if it was not a "real bug", there's this thing called diplomacy. There are a couple of posters in the forums who are prone to fly off the handle and treat people poorly.
If I treated customers or my testers like that I'd be in line at the unemployment office.
-
I will just add here from my real life experience that I totally understand where they come from when they are short with people. This is not software your newbie to networking should be using, if your that new to networking you need to be using a consumer router. This is software for people who have a fairly good understanding of how things work on firewalls and routers. I can't tell you guys how many different forums i visit and have to answer questions like whats a vlan, or whats a static route, or someone who just got rid of their asus router complaining that their vpn doesnt work and with no information wonder why.. It takes a toll.
While some people have been short with me, they have been very helpful, and you wont see many places where the guys who are developing the product are also very network aware and are willing to spend their time helping out the public.
Personally instead of complaining about how short they get, you should be asking yourself why their getting so short, and if you posted any relevant data to help them figure out your problem..
Just my 2cents…
-
KOM it was kill bill, and the post in question where he used the word "bollocks" to describe my report did specifically have the word liar. Maybe I should have done a screencap because now his comments got toned down, its only my word.
-
Please don't make excuses for rudeness. It has no place anywhere for any reason.
-
dcol exactly.
-
Think of kill bill as the Don Rickles of the forum.
Might seem a bit harsh at first but he's always spot-on. When it comes to pfSense I'd trust him blind folded. -
Should have grabbed some more popcorn I guess…
@chrcoluk: Things like the traffic shaping wizard have been used by tens of thousands of users. Don't you think someone would have noticed before you that "rules created by traffic shaper wizard dont do anything"? It's not exactly a fringe use case when you look at the traffic shaping subforum.
Why's this thread even in this section? Cannot see any mention of 2.4 in either of your "bug" reports.
-
TREY GOWDY: PIN DROP SPEECH
https://www.youtube.com/watch?v=N0dCTvX0wzsYou can watch the whole thing or go right to this time mark for some really good points re: communication and persuasion.
https://youtu.be/N0dCTvX0wzs?t=384"You also need a very effective manner or method of communication. In other words you need to learn how to persuade."
"You want to persuade, change people's minds to come around to your of thinking on whatever issue it is?"
"I will tell you what doesn't work. Insulting people."
"Insulting people does not work if your objective is to persuade.""When I see a bumper sticker that says 'Don't blame me, I didn't vote for the idiot.' Do you think that is persuasive?"
"You know what happens when you're insulted? You become even more dogmatic in holding your incorrect belief than you were before you were insulted. So if your goal is to persuade you shouldn't be insulting people."
From a constructive confrontation course the point that struck me most was to focus on the issue/problem and not the people/person.
i.e. Leave the person's knowledge out of it. Don't attack them, belittle them etc.examples:
"If you knew what you were doing then we wouldn't be having this problem in the first place."
"If you would do what you were … then ..."Instead address the issue. You may think the issue is their lack of knowledge etc. And that may very well may be the ultimate source of the issue it is not the issue at hand and is better dealt with elsewhere and perhaps by others and by other means.
If you what to help someone solve a problem. Great. But condescension, belittlement and personal attack is of little help to anyone. Unless it is therapeutic to feel better about oneself. Even that would be questionable compared to benefit of other therapies.
If you really want to feel good about yourself. Help someone you think is an idiot while genuinely abstaining from any criticism, condescension, belittlement, personal attack, product/services attacks, insults (personal or otherwise), etc.
It's kind of like the proverb; "It is more blessed to give than to receive."
Be gracious in giving (helping others) rather than condescending, insulting, belittling, etc.
It is much more fulfilling.
-
Ok regarding the issue in question (bug report), I did some more testing after I received some support.
I have now found a pattern.
I mentioned earlier that the traffic was been matched as the counters went up, well yes but it turns out its only matching the connection setup (initial syn).
I have written a ton of notes but it would be a very big post so will try to summarise.
I wiped the traffic shaper config clean and ran the wizard again, I selected dedicated link, additional to the previous options I set I set http traffic to low priority as http is really easy to test and also that steam downloads are port 80 not the ports used in the wizard.
I observed by watching the queue screen there was a very small amount of qACK traffic but qOthersLow and QothersHigh both remained at a big fat zero.
I went back to the rules page and noticed the counter for the rules only incremented a small amount at the start of a connection but never increased for more traffic, e.g. a http download would increase the counter at the start of the download, but no more during the download. This I believe it is just recognising the initial syn packets and not the data packets afterwards.
I then went to the rule and highlighted LAN so both WAN and LAN are selected. This causes a small amount of traffic to appear in the qOthersLow for http but it is very tiny, not even 1 kbyte/sec, it shows using bytes/sec.
I switched back to default WAN only selected and changed the rule to pass.
Suddenly qOthersLow full of activity during a http download and in addition the packets counter for the rulle accumulated 'during' the download rapidly, suggesting its working (but of course with the security issue as a floating rule).
I then moved the pass rule to a LAN rule away from floating and it carries on working correctly but without the security issue. This behaviour is the same for my generic ack rule also.
I have tried to prove my previous results wrong, as I am the sort of person who will hold their hand up if proven wrong even if embarrassing but I cannot replicate what should be the correct behaviour on my setup. It is matching syn packets it seems but thats it, nothing else when the rule is a 'match' rule.
I need to test UDP dns, which I will test by running a dns benchmark app as I want to confirm if UDP is having the same issues as I know for sure on TCP.
Also this was posted in this section because I am running the 2.4 beta code not 2.2 stable.
-
I wiped the traffic shaper config clean and ran the wizard again, I selected dedicated link
You should have selected Multiple Lan/Wan.
I'd really like to help you out but I think I'm done here. All these posts, all this heat, and yet you have not posted a single thing that anyone requested. No floating rules, no shaper config, nothing. Considering how you seem to want to get this working, I find it baffling that you consistently post nothing for anyone to help you other than your descriptions of what you think you've done. It works for everyone els ein the expected way.
-
@KOM:
I wiped the traffic shaper config clean and ran the wizard again, I selected dedicated link
You should have selected Multiple Lan/Wan.
I'd really like to help you out but I think I'm done here. All these posts, all this heat, and yet you have not posted a single thing that anyone requested. No floating rules, no shaper config, nothing. Considering how you seem to want to get this working, I find it baffling that you consistently post nothing for anyone to help you other than your descriptions of what you think you've done. It works for everyone els ein the expected way.
That is what I selected the first time when I made the bug report, the issue you have is you are still approaching this as a operator error issue, you simply refuse to accept anything else, until you change that line of thinking you will not manage to make any progress on this issue.
Also noone asked me to do do anything apart from one guy who asked for the screenshots which I am posting in a matter of minutes from now, I know you are really hoping to see something that looks like I fiddled with something that must be to blame.
-
Ok here is the results using the dnsbench GRC application which I used to flood my router with outbound dns connections, the results were not the same as TCP tests.
1 - With the default rules created by the wizard it doesnt work but in addition unlike the other match rules there is 0 matches tallied on the rule.
2 - changing to pass whilst still a floating rule is the same result as #1.
3 - Having it as a pass rule on the outbound LAN interface (not floating) it correctly matches the packets and I see dns traffic in qOthersHigh queue. -
attaching floating rules pictures, I have explained already what is there, but for those who want to visualise here it is.
-
the issue you have is you are still approaching this as a operator error issue
Again, it works for everyone else, so yes we're assuming it's PEBKAC.
you simply refuse to accept anything else, until you change that line of thinking you will not manage to make any progress on this issue.
Since you consistently refuse to provide any details whatsoever, we have nothing else to go on, do we? You're saying that even though others (including myself) have it working but you don't we should assume it's a bug? Here's a thought: everyone in here is telling you you're wrong but you stubbornly refuse to listen and instead want to argue with the devs and old-timers. Maybe you're the one in need of a change in thinking?
Also noone asked me to do do anything apart from one guy who asked for the screenshots
I'm pretty sure I've asked at least once before.
I know you are really hoping to see something that looks like I fiddled with something that must be to blame.
You're making this personal for no good reason. I was hoping we could fix your problem so you would stop moaning about it but like I said I'm done. I'm sure someone else here can look at your diagrams.
-
You havent been constructive in any of your posts in this thread. I have obviously personally ruled out operator error by repeating the procedure probably a dozen times now, its not my fault you wont accept that. You said you staying out of the thread which is probably the best post you made in here, of course if you want to offer constructive input go ahead, but first step back take a breather.
I have never said this doesnt work on any pfsense routers. You seem to think something either must be broken on every usage case, or working in every usage case, nothing in between, except the real world doesnt work like that, bugs can surface themselves in mysterious ways. Not to mention that you saying "everyone" has told me that I am wrong in that there is a malfunctioning code problem (or documentation issue). As that is also not the case.
I have spent 100s of hours in my job when having to track down bugs that may only affect sub 1% of people. I didnt reject reports because "it works for most people".
-
A few observations-
- I don't think trying to continue troubleshooting in a thread that was started to comment on the bug reporting process is going to be productive for you. Perhaps continue on a previous thread, or start a clean one.
- Most of the people here are just other users trying to help out, so don't get so offended is someone is short with you. If you are paying me $100 an hour to troubleshoot your problem, I promise I will be very polite and professional. For free help, take what you want and ignore the rest.
- Skill level varies widely, so there is naturally skepticism. I have tried to reason with people who claimed to have years of network experience, but acted like twelve year old kids.
- Document your case, and ignore those who you feel are not constructive. Getting in feuds is not going to help you solve your problem.
I'll stop there. Good luck on your issue. Personally, I find the shaper complex enough that I don't try to give others advice on it.
-
Fair points dotdash.
I will take your advice and start a new thread on the issue alone and we can see if a resolution is found.
-
thread for the issue is here if anyone wants to participate on the problem at hand (not for discussion of the bug report issue.)
https://forum.pfsense.org/index.php?topic=123757.new#new
