LinuxQuestions.org
Register a domain and help support LQ
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Networking
User Name
Password
Linux - Networking This forum is for any issue related to networks or networking.
Routing, network cards, OSI, etc. Anything is fair game.

Notices

Reply
 
Search this Thread
Old 03-10-2003, 02:09 PM   #1
dunkyb
Member
 
Registered: Nov 2002
Distribution: Debian testing.
Posts: 143

Rep: Reputation: 15
large routing tables


Hello

I run a linux router.. with Windows XP clients.

However, Browsing usually hangs/stalls on all the clients, then resumes after a few seconds. On lots of websites, but not all the time... I am trying to find out why this is, no luck as yet (any ideas?!) ... Im not sure whether the problem lies on teh clients (default windows xp boxes) or the linux router.

Does anyone have "large routing tables" built into their kernel? would this enhance performance? .. would it hinder it if i added it in to my kernel?

Also, is the "fast network address translation" under advanced routing options a good idea to add in?

Thanks for any advice/tips

Does anyone else suffer this problem?

Cheers,

Duncan
 
Old 03-10-2003, 03:49 PM   #2
peter_robb
Senior Member
 
Registered: Feb 2002
Location: Szczecin, Poland
Distribution: Gentoo, Debian
Posts: 2,458

Rep: Reputation: 47
Tell a bit more about the Linux router, which Linux, which firewall, etc.

Large routing tables is for static routes, for large networks/isp's...

You may only have a dns bottleneck...
 
Old 03-10-2003, 08:49 PM   #3
lazlow69
Member
 
Registered: Jan 2003
Location: Central New Jersey
Distribution: Knoppix to play, Slack current, OpenBSD stables
Posts: 111

Rep: Reputation: 15
You mention a DNS bottleneck... I am assuming that in order to enable a Linux router you would also need to enable DNS?
 
Old 03-11-2003, 01:08 AM   #4
dunkyb
Member
 
Registered: Nov 2002
Distribution: Debian testing.
Posts: 143

Original Poster
Rep: Reputation: 15
well, it's running Debian Linux stable 3.0r1..

serves 4 windows XP Pro clients

2 network cards, eth1 -> cable modem
eth0 -> LAN (10/100 switch)

I have an iptables firewalling script running, but these problems seem to occur when I blank the ruleset, effectively removing the firewall ...

Running "dnsmasq" caching dns forwarder.
Running custom 2.4.20 kernel.

Anything else you want to know?

Cheers
 
Old 04-26-2003, 07:26 AM   #5
..Tookers..
LQ Newbie
 
Registered: Nov 2002
Location: Hong Kong
Distribution: Redhat 8.0, 7.4, 7.2
Posts: 8

Rep: Reputation: 0
dunkyb, could you tell me what routes you have setup in your routing table?

I have pretty much the same setup... except im running a custom redhat install which only has iptables and the necessary system scripts.

Each computer on my network uses its own PPPoE client in which to "dial" onto the internet.

I have eth1 (which goes to the cable modem) setup as 10.0.0.1
i have eth0 (which is on the local hub) setup as 192.168.0.1

i set the following

route add 0.0.0.0 gw 10.0.0.1

i have tried other ways but nothing i've tried has worked so far.
any help?
 
Old 04-26-2003, 10:04 AM   #6
dunkyb
Member
 
Registered: Nov 2002
Distribution: Debian testing.
Posts: 143

Original Poster
Rep: Reputation: 15
Here is my routing table (/sbin/route)

Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.0.0 * 255.255.255.0 U 0 0 0 eth0
81.96.53.0 * 255.255.255.0 U 0 0 0 eth1
default pc1-horn1-4-gw. 0.0.0.0 UG 0 0 0 eth1

----

here is route -n, maybe more readable

Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.0.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
81.96.53.0 0.0.0.0 255.255.255.0 U 0 0 0 eth1
0.0.0.0 81.96.53.254 0.0.0.0 UG 0 0 0 eth1


Duncan
 
Old 04-26-2003, 11:18 AM   #7
peter_robb
Senior Member
 
Registered: Feb 2002
Location: Szczecin, Poland
Distribution: Gentoo, Debian
Posts: 2,458

Rep: Reputation: 47
First thing I suggest is checking the performance of the dns servers you have listed in /etc/resolv.conf
dig each one with an address from another distant continent and check the reply times.
'man dig' for details
You may have a slow server first in the list. Put the fastest one first.


Time to start adding -j LOG rules in different parts of your script to see what is getting dropped.
Find out if there is a pattern from the entries.
usu LOG entries go to /var/log/messages if you make them '-j LOG --log-level 6 --log-prefix "something " '
 
Old 04-26-2003, 11:51 AM   #8
dunkyb
Member
 
Registered: Nov 2002
Distribution: Debian testing.
Posts: 143

Original Poster
Rep: Reputation: 15
well i have LOG'd the FORWARD stuff, and nothing ever gets logged, so assume that is ok .. have LOG'd INPUT stuff, and get a lot of rubbish .. mainly to port 80, 113, etc.. just scans and random rubbish i guess. I was getting some packets (DHCP) from 10.69.16.1, so have specifically allowed them through, and no longer get them .. Doesn't seem to be anything out of the ordinary!

Cheers

Duncan
 
Old 04-26-2003, 01:18 PM   #9
peter_robb
Senior Member
 
Registered: Feb 2002
Location: Szczecin, Poland
Distribution: Gentoo, Debian
Posts: 2,458

Rep: Reputation: 47
Also try giving dnsmasq a larger cache number in /etc/dnsmasq.conf
I have found it chews out the default 150 quite quickly.
Try 300...
 
Old 04-26-2003, 02:19 PM   #10
dunkyb
Member
 
Registered: Nov 2002
Distribution: Debian testing.
Posts: 143

Original Poster
Rep: Reputation: 15
The version I have sets it to 300 as default... Should i go any larger? ....

Does dnsmasq not delete cache entries when required?
I thought it purges the cache when they timeout...

Cheers,

Duncan
 
Old 04-26-2003, 03:06 PM   #11
peter_robb
Senior Member
 
Registered: Feb 2002
Location: Szczecin, Poland
Distribution: Gentoo, Debian
Posts: 2,458

Rep: Reputation: 47
Yes, it does purge,
but with 5 or more clients browsing busy sites with lots of links, images, sponsors etc,sometimes the cache is cleared just seconds before the next request for old data arrives.

This all takes time to reload and clear something else out...

You mentioned you have periods where it all grinds to a halt momentarily.
These types of cache policies can make it seem slower because they are all swapping and clearing at the same time.
You will see that from their own log files.

Did you check the external dns server speeds?
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
large interface tables patch alex_anagran Linux - Networking 0 11-08-2005 02:10 PM
Routing tables TPB Linux - Networking 3 07-27-2005 09:01 PM
two routing tables fugzi Linux - Networking 7 12-03-2004 06:02 PM
routing tables dunkyb Linux - Networking 3 05-25-2003 02:10 PM
linux kernel: large routing tables dunkyb Linux - Software 0 02-19-2003 02:19 AM


All times are GMT -5. The time now is 04:09 AM.

Main Menu
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration