goto http://www.speedguide.net/

be sure to 'enable scripts' at top of screen is broadband, windows2k/xp there it will run a test on you, etc. they give tips to tweak windows and linux there.

Thanks. The situation worsened considerably. The ISP also has a test server- and I get like 1.5M out of it :confused:

However, they're not serving users from it normally. The guy never heard of traffic/bandwidth shaping and can't ask anyone. The "support" is in a separate location from the server farm and they aren't allowed to call them.

P.S. Fogie that site had some ideas about tweaking Linux stuff and it seemed to help a lot on one specific website but everything else stayed the same. This is a wierd one.

Hey Randux;

To check for proxy settings in 'bloze, ans assuming you're using IE, you should check your settins under 'tools' -> 'internet options' and then select the 'connections' tab. In that tab you'll find a 'lan settings' button which will exihbit a window with the proxy settings. I'm translating the strings from my portuguese 'bloze menus so I hope they correspond to what you get on your screen. :)

Edit: You can also reach those through 'control panel' -> 'internet options'...

Cheers

Thanks Bono, I thought you meant some other proxy- actually, I don't know what I thought. I am not using any web browser proxy ( I don't use exploder, even in bloze LOL).

Ok, here is the interesting thing: I spoke to the ISP and they told me I could try their proxy in my browser. When I use the proxy in the browser on my Slackware machine, I am getting full speed again.

I don't understand what this really means or what it proves. Also....what are the disadvantages? I will have to test some ftp and other protocols to see if they were working normally. I mean, I want to check to see if stuff that didn't go through my web browser was not affected by the slowdown I noticed recently, or if it's also being throttled.

Can anybody explain what's going on???

Thanks,
Rand

Randux,

I sure empathize with you on this peculiar problem! There seems to be no clear consensus in this thread to help you troubleshoot, so I will add my own free-flowing thoughts.

I do not have the same setup as you with two separate providers. However, I have noticed in my testing that pinging and testing with the ISP server is useless because that is more or less a direct connection to your PC. In your case, yes, you are passing through the cable company gateway, but you already stated that you experience expected speeds through them. That seems expected because the cable company gateway is not doing any "hard" work, such as DNS lookups. That gateway is merely providing IP forwarding to you. Your problem seems to be downstream of the ISP server.

I doubt the problem is copper (the wires) or other hardware because Windows runs well. Additionally, you have no problems with the cable company gateway.

You say you believe the problem is not DNS, but you did not mention whether you verified your /etc/resolv.conf file. Also, have you looked in your /etc/hosts file to ensure that nothing strange is there? What about your iptables firewall? As long as you have no ports open to the outside world, you could temporarily disable iptables and see if anything changes.

If you are using KDE, there is an environment variable you can add to disable IPV6. In /etc/profile.d/kde.sh add:

KDE_NO_IPV6=1

I'm curious how you are doing your testing. You have not said so, but sounds as though all of your testing thus far has been with a browser. Have you tried some simple wget commands? You can obtain a copy of wget for Windows if you want to compare times.

If you are testing only with a browser, then that is one place I'd investigate. I agree that hardware more than likely is not the problem, but you could have your browser configured differently between OSs. However, a wget test would answer some questions because that is a different program and if you use FTP, a different protocol.

Regarding browsers, have you tried using different browsers while in Slack?

Have you tried running a traceroute?

Regarding running Slackware through Windows, that might help decide whether the browser is the problem. If you configure Slack (rc.d/rc.inet.conf) to use the Windows box as a gateway, and you still experience problems, then you have a clue that the problem is in your Slackware or browser configuration. Connect your Slack box to the Windows box, and connect your Windows box to your router or directly to the cable modem.

That you have faster speeds through the ISP proxy means little because with their proxy they are using caching software. When using the proxy server, download a file that is unlikely to be cached, such as a file size of 100MB or so. Then you would have a better picture about the proxy server speed test. And if I understand correctly, these proxy servers are not for general usage, but are installed for people using accelerator software, which is available only for Windows. And that raises another question: is your Windows box using this proxy server and any accelerator programs?

Although certainly possible, I never have heard of any ISP or service provider purposely discriminating against certain OSs. They might possibly do so unintentionally, but I seriously doubt a provider would do so knowingly. But my thoughts are how would they do this? At the TCP/IP level, there is not a lot of complicated identification going on, just a lot of basic packet handshaking, etc.

Lastly, you said the problems arose recently. That is a key clue. I've been around computers a long time. When a problem arises suddenly, and hardware failure is not obvious, either you or your provider did something. In such cases the first thing I always do is sit back, relax, and ponder what I might have done recently to cause problems. And no, I am not always the problem. I have a local dialup ISP that occasionally craps everywhere. My service is awful for a few hours until somebody resets the servers. Of course, the problem also could be the telco, but the ISP never informs users what caused the problems. But to reflect upon a previous statement, even during those crappy periods, pinging the ISP server proves nothing because often during those times I can dial out and connect, but go no further. As long as I try to go no further than the ISP server, who's home page always shows the servers operational (ha!), I nonetheless can go no further.

So yes, although I do not think much of the idea that your ISP purposely crippled your connection, I have no doubt they can do so unintentionally. Yet, I also would throw some reflection upon anything that you might have done as well.

I hope you soon resolve the problem. I'm just chattering in the hopes of possibly triggering something in your mind that might help.

Hi Woodsman,

Thanks very much for your post. I'll try to answer these questions.

Yes, I checked resolv.conf on both my Slackwares (and I think but can't swear to it now on my OpenBSD machine) and they had the same two DNS servers, in the same order, as the winbloze machine. I did not check /etc/hosts until you mentioned it but all I have in there is two entries for the loopback 127.0.0.1 localhost and 127.0.0.1 hostname.hostname@domain.

I'm not using KDE and I'm running the default 2.4.31 kernel as delivered.

I have done all the testing from a bunch of speed testing websites I've used over the past few years, and all on my primary browser (mozilla) since that's where I noticed the problem. I don't use ftp or much of anything else on this desktop except between machines on my own lan so I didn't think to look at this but it's a good idea. I wanted to do traceroute but I forgot the command until you mentioned it. I kept thinking it was an option on ping that I wasn't finding in the doc :p

I do have a bunch of browsers on my Slackware machines but since I do everything through Mozilla I hadn't thought to change it. I can't remember if I've ever upgraded it or changed anything on it within a year or so but I don't think so.

Someone on bsdforums suggested that and it's a good idea, but I have only one NIC in the winbloze machine and I can't get another one in there. So although that would have been a pretty good test I can't do it.

Actually from the sites I'm using I'm sure nothing is cached. The only thing I tested was the speed check sites and they ran at full speed after I set the proxy in mozilla. I don't have any accelerator anything and I've purposely avoided all uses of proxies in my browsers. On winbloze I am certain that I don't use one. I always thought they were a bad security policy.

I agree with this- and that's why I immediately blamed them :p I can't remember changing anything but now that I'm thinking about this I do remember when I installed OpenBSD 3.9 after having run 3.8 I did notice very poor performance in mozilla. I think it's the same level 1.7.11 that runs on Slackware. Now if I could only remember if I was ever running an old copy of mozilla on these Slackware desktops...

Thanks for all your comments.

Rand

Woodsman, after reading your post somehow things all came together. I remember when I went from mozilla 1.7.8 to mozilla 1.7.11 on OpenBSD performance was horrible. But I don't remember upgrading or changing levels on Slackware 10.2. I think 1.7.11 was the regular mozilla.

I just installed seamonkey (which for anyone who thinks it's new, it's 100% mozilla) and it works (without any proxy) like a house afire.

Thanks very much for helping me figure this out. I don't know why 1.7.11 runs so badly, but it does. Now I have to check what I have on bloze but at least I'm running with no smoke and mirrors. :) I was running 1.7.8 forever and it was great.

Seamonkey 1.0.5 I just got from the mozilla site is super nice because it uses all the old settings and bookmarks from mozilla without me having to do anything.

Good to hear! I hesitated posting because I had nothing "concrete" to offer, and only some generic troubleshooting thoughts.

Hmm. I might give Seamonkey a spin when I find some time (and after 11.0 is official ;)).

Well, dammit, everything is back to being weird again. Yesterday after I tried the proxy and decided to deep-six mozilla 1.7.11 and installed Seamonkey, everything was fine, without the proxy.

Things were even fine on my OpenBSD 3.9 machine running Mozilla 1.7.12 which is running on the same physical box as winbloze (not the one I installed Seamonkey on!) and used to be affected by the same slowdown as my Slackware machines, one on the same box, one on a different box. I knew this wasn't a good sign.

Today we're back to erratic performance just like the good ole days. Sometimes almost full speed, sometimes, almost dead slow. I think the fact that it screams with their web proxy proves that they're doing something. I just don't know what they're doing, why they're doing, or what I can do to get good browser performance without it. Maybe some network guru can figure this out.

I really appreciate all the suggestions and comments and advice from you guys. Even though this sucks :mad: it was still really nice that you tried to help. If anyone has any ideas I'm all ears but I can't play around with this any more at the moment. I have to get back to work.

Rand

P.S. Woodsman I didn't install the Seamonkey package from Slackware packages- if you want you can download it from mozilla.org. The installer is a piece of cake and all your mozilla settings are preserved. A couple of things to watch out for- if you do a custom install like I always do with it, make sure you check personal security manager. It's their stupid name for SSL and you'll be toast of you don't install it because there is no repair-type install. I forgot about that and I had to do a fresh install (it only takes 20 seconds anyway, but it's annoying not to be able to add a feature to an existing install.) You'll have to go get flashplayer again if you didn't save the plugins from your old mozilla/firebox installation. And of course if you do it this way you have to make your own Slackware package if you want one.

Hi,

Just a thought, when you log to the ISP are you 'nailing' the IP? If so the problem could be something as simple as a rack interface problem.

Hi Gary, I didn't understand your question. I'm pretty clueless in networking although learning a little bit here and there.

Thanks,
Rand

hey Randux a while back my network on my pc was going hay-wire and for some reason,
my /etc/resolv.conf showed "search @mydomain" in the top. When I just put in my ISP's dns server addresses and then ran I was fine. The wierd thing is, I know for a fact that I had never put the search part in there, cuz I had backed up my whole /etc folder just after I installed, in case I made a mistake. On occasion, this still happens to me now and again...the "search @mydomain" comes back!

Just a thought...

Hi OF,

Thanks. I checked it several times and it always had the same two servers, in the same order, as all my other machines.

Rand

I have just woken up and am only barely coherent. Am I correct in saying this problem is browser specific, but that a browser's performance deteriorates over time. Is this true? How does wget perform? What is the ping time for google.com on the linux box compared to the bloze box?

If wget performs well, you probably do not have a network problem. Could there be a bottleneck at the disk? How do other disk-intensive programs perform? As silly as it sounds, have you tried clearing out the cache? I don't really know how browsers go about caching things, but perhaps Mozilla started a new cache when you switched to a proxy? Seamonkey most certainly used a new cache.

In the case that it is an actual network problem, someone already mentioned that you run a packet sniffer and this is very good advice. Ethereal has a nice friendly interface. If you're not wise in the way of packet sniffing, maybe capture and upload the output on one of the many sites for posting such things (sendspace.com, for example). (If you're going to post the results, tcpdump works just fine.)

As a sidenote, I once experienced terrible performance under Firefox. On the same partition I had several million small text files left over from benchmarking software. After cleaning house, Firefox worked just fine. Strangely, I only noticed a problem with two programs: Firefox and latex. This led me to investigate and I found that only took several hundred thousand small files to slow Firefox down. This is on resiserfs even.