Browser loading slowly after LAN server change, but only for one machine

MGCJerry

Limp Gawd
Joined
Sep 18, 2005
Messages
144
As the title says; My browser loads slowly after LAN server change, but only for my machine. My VMs on my machine are fine as well. More information below.

I had a server named "Hydra" (192.168.0.150) which was my main web/development server. Running Server2k3 (32bit), Apache 2.2, Maria DB 10. Everything runs fine except for a couple changes I wanted to make which would require some hardware changes...

Server #2 entered the game. "Typhon" (192.168.0.10). Running Win7 pro (64bit), Apache 2.4, Maria DB 10. Did several tests of moving/dupicating my software, settings and network shares from "Hydra" to "Typhon". Everything works great. I have a Boltek PCI Stormtracker card that will need testing once I do the swap over.

On my computer, I have several drives mapped to Hydra (G, H, I, & S). This past weekend was my official swap over day where "Typon" would take "Hydra's" place. I moved the Stormtracker to Typhon and minus a program/driver hiccup related to the Stormtracker, everything worked and transitioned just fine. Next is a step by step of what I did

• Unmounted all my network drives on my machine that were mapped to "Hydra".
• Logged onto "Hydra" released the IP address, and shutdown the machine and removed cables.
• Logged into "Typhon" cleaned up some test/transfer files, released the IP address, and shutdown the machine.
• Went into my router (Firebox XTM520), and changed the LAN IP assignment to give "Typhon" "Hydra's" old address assignment (192.168.0.150).
• Booted up "Typhon".
• Verified IP assignment for Typhon, good. Remapped my network drives to match what I had on "Hydra". File copy speeds are great.

• Checked httpd by going to http://192.168.0.150 with my browser, and it took a solid 3-5 seconds before it even loaded in my browser. All subsequent page loads are just as long. Even static content (text, static html, images, etc).
• Googled the problem and did all kinds of Apache 2.4 config changes, clearing browser cache (which is *always* disabled on my machine), fushed dns caches on machines, disabled ipv6, reinstalled apache 2.4 fresh config, and much much more. None of that made any difference.
• Checked from other machines on my LAN, and even virtual machines on my computer and everything is fine and loads *immediately*. Other browsers on my machine are also slow (at least 3 seconds, but up to 20 seconds per page). Other browsers on other machines on my LAN and in virtual machines loads *immediately*

This points to a Windows networking problem on my machine (but fine within 9 Virtualbox VMs). But where? Anyone have and additional advice.
 
Sorry not extremely skilled in this kinda troubleshooting but maybe you have an entry in your host file for the old server? Or DNS Suffix list on your NIC itself under IPv4? What if you ping the server from your desktop versus the VM's is there a latency difference? Sorry kinda just shooting in the dark but maybe it will help you discover something you haven't thought of.
 
Out of curiosity, have you tried another browser on the faulting machine? Or are you just using the one browser you used on the faulting machine in the past for diagnosis?

Also, just out of curiosity, what happens if you change the IP address of Typhoon?
 
OT: If you're developing *always* disabling browser cache is a sure way not to catch cache related problems in your application - so the end users who *always* have the cache enabled will be the only ones suffering from stale cache problems ;)
 
OT: If you're developing *always* disabling browser cache is a sure way not to catch cache related problems in your application - so the end users who *always* have the cache enabled will be the only ones suffering from stale cache problems ;)

This is a valid point....
 
Sorry not extremely skilled in this kinda troubleshooting but maybe you have an entry in your host file for the old server? Or DNS Suffix list on your NIC itself under IPv4? What if you ping the server from your desktop versus the VM's is there a latency difference? Sorry kinda just shooting in the dark but maybe it will help you discover something you haven't thought of.

1. No other entires in the hosts file except for usual localhost 127.0.0.1
While i could be wrong, i think this isn't a dns issue since Im accessing strictly via IP. But, I've also been wrong before making such assumptions. Nonetheless, it was still something to doublecheck.
2. No dns suffix.
3. All ping latencies are under 1ms on my machine as well as VMs hosted on my machine.

That's why I posted here. I'm sure somebody would have thought of something, or pointed something out that I completely missed. The joys of troubleshooting. ;)

Out of curiosity, have you tried another browser on the faulting machine? Or are you just using the one browser you used on the faulting machine in the past for diagnosis?

Also, just out of curiosity, what happens if you change the IP address of Typhoon?

1. Yes, I have tried another version of palemoon (portable and normal), Firefox 32 (portable and normal)... And Internet Exploder, err.. Explorer. :D
2. Still slow page loads.

OT: If you're developing *always* disabling browser cache is a sure way not to catch cache related problems in your application - so the end users who *always* have the cache enabled will be the only ones suffering from stale cache problems ;)
Very true. I do additional testing before pushing the the update to help catch those problems too.


With some additional testing and looking into developer tools in my browser, I see...
DNS Resolution: 0 ms
Connecting: 1 ms
Waiting: 2117 ms
Receiving: 203 ms

Does with with a static HTML page, image, text document, phpinfo page, or anything else.

Ive also just tried messing with the adapter settings by altering the network team, and that also seemed to make no difference. Something else ive seen is that if there are enough images for instance (image gallery), Some images will load, then haveto wait another 2 or so seconds for more to load.... then wait 2 more seconds, then more will load. A gallery page with 60 images (all under 800x600 200K in size) still takes over 5 seconds to load with several of the images waiting an additional 2113 ms before they load.

Strange..

Edit
Apparently when I run apache on my local machine, and try accessing it from another machine, it does the same thing with the excessive wait time. Maybe its something in Windows 7 when running apache.
 
Last edited:
What happens if you update your httpd.conf and add
Code:
AcceptFilter http none
AcceptFilter https none
 
Do you have hostnamelookups off in your apache conf? It hangs if for some reason DNS resolving doesn't work quickly enough.
 
What happens if you update your httpd.conf and add
Code:
AcceptFilter http none
AcceptFilter https none

These were some of the first settings in place. However, this did make a marked improvement with the other LAN devices earlier in diagnosis.


Do you have hostnamelookups off in your apache conf? It hangs if for some reason DNS resolving doesn't work quickly enough.

Yes, I do. Ive encountered this issue back in Apache 1.3 days hosting on windows 2000. I always double check to ensure its off.


After further testing on some 'off time' (jeez I hate working 6 days a week when I have soooo much computer stuff I want to do, but its allowing me to buy my new stuff), I can live-cd boot my machine in Ubuntu and the site loads like lightning. This seems to indicate a problem with my machine's windows 7 installation at this time. All other windows 7, Windows 10, *nix & Android devices on the network are fine loading off the server. If/when I find the problem, I'll report back.

Thanks for the replies everyone :)
 
A few questions/avenues to explore:
Do you have "specialized" network drivers installed on your win7 machine? I saw similar issues with the ASUS drivers about 7 years ago (not too long after Win7 shipped).
I noticed in your second comment that you have PHP, are you using LoadModule or AddType to handle PHP?
Have you modified default settings for services in either the host or client machine?
Finally, do you have any third party firewalls or security software installed on either the host or client?
 
A few questions/avenues to explore:
Do you have "specialized" network drivers installed on your win7 machine? I saw similar issues with the ASUS drivers about 7 years ago (not too long after Win7 shipped).
I noticed in your second comment that you have PHP, are you using LoadModule or AddType to handle PHP?
Have you modified default settings for services in either the host or client machine?
Finally, do you have any third party firewalls or security software installed on either the host or client?

Good questions...
1. No, I just have the standard updated Intel drivers for my board. Same for the server.
2. Both... I have included the relevant portion of my httpd.conf below. I had loaded php the same way when I was using Apache2.2 on Server2k3.... Perhaps its different with 2.4?
3. Yes, I killed a lot of the unnecessary services on the server and my machine (years ago). (remote registry, audio services (on the server), print spooler (on the server), font cache, indexing, wireless crap, media discovery, .net framework, etc).
4. No firewalls or security software

Apache 2.4.27 httpd.conf
Code:
LoadModule php5_module "D:\localhost\php-5.6.31\php5apache2_4.dll"
AddType application/x-httpd-php .php
AddType application/x-httpd-php-source .phps
PHPIniDir D:/localhost/php-5.6.31/

ServerSignature Off
HostnameLookups Off

AcceptFilter http none
AcceptFilter https none
EnableSendfile off
EnableMMAP off
 
Solved!

I logged into my switch to make note of my port configurations because I have a new switch coming soon, and I seen a TON of 'alerts' regarding my machine & the server. Hmm... Cleared all the alerts, and loaded the local site. Alert on the switch console... Loaded it again... Another Alert on the switch.

"Excessive undersized/giant packets on port 13 (My machine)"
"Excessive undersized/giant packets on port 7 (Server - lan1)"
"Excessive undersized/giant packets on port 8 (Server - lan2)"

Medusa : 2810-24G switch said:
Description:
A device on port 13 is transmitting packets shorter than 64 bytes or longer than 1518 bytes (longer than 1522 bytes if tagged), with valid CRCs.
Possible causes:
The possible causes include a misconfigured NIC or malfunctioning NIC, NIC driver, or transceiver.
Actions:
  1. Check the NIC for a misconfiguration.
  2. Update the NIC driver software.
  3. Replace the malfunctioning NIC or transceiver.
  4. Check for a short-circuit in the cable path connected to this port.

So I dove into my NIC's settings, and reset everything to default and rebooted as requested. That solved the issue. I don't remember changing any of those settings, but its definitely something to keep an eye on. I feel stupid that I own a managed switch that monitors for this kind of crap, but it never occurred to me to look there for the problem.

Thanks for the help and assistance everyone :)
 
You probably had the large packet size enabled in the NIC and that's a known source for problems.
 
Back
Top