I would tend to agree with a first step of replacing the Cat 5 cabling. I 
corrected a similar issue onsite by replacing the network cabling. In this 
case the network would be running "as usual" with no problems then all of a 
sudden there would be a 90% plus packet loss. This issue would also correct 
itself over time. 

Nagios is an excellent software package that can monitor this for you.


cheers,

Jim




On Fri, 24 Jun 2005 08:11:51 -0600, Dany Allard wrote
> Shawn
> 
>  Here is a real quick test to check if it is a network problem with Telus.
> Start a ping from the firewall to a well known website (eg 
> www.google.com). Let the pings run long enough to catch the problem, 
> then see if you have any packet loss. If you have anything more than 
> a 3% packet loss, there is likely a network problem with telus. I 
> would also suggest changing the ethernet cable for your firewall as 
> well. I have seen strange things with bad cables. Also if there is 
> no problem with the pings but you still have problems with the name 
> resolution, you can write a cronjob to check the name resolution 
> every few minutes (or you can download the command line nagios 
> packages that do the same thing). Also check for EMI stupid things 
> like that can really affect your internet connection.
> 
>  Hope that helps
> 
>  Dany Allard
> 
> Shawn wrote:
> 
> >Hi all.
> >
> >I've run into a situation that leaves me scratching my head.  I'm looking 
for 
> >any thoughts that might help identify the cause of the problem.
> >
> >A contact of mine is having odd network problems.  The network runs fine 
for a 
> >bit then they have periods of slow response and/or domain resolution 
> >problems.  Then the network returns to normal on it's own.
> >
> >The network is a mixed environment, with IPCop for the firewall, and a 
Linux 
> >server doing web gateway duties.  My contact has spoken with Telus (their 
> >ISP), who pretty much said the problem was an internal issue.  I'm not so 
> >sure of this, but haven't ruled out the possibility.
> >
> >We have tried changing the name servers on the IPCop firewall, and 
changing 
> >the internal DHCP config to use the firewall as the primary name server 
> >(instead of the W2K Active Directory box with DNS).  The problems 
persist.  
> >We have done some simple analysis of the network traffic using EtherApe, 
and 
> >identified a couple of workstations that were using excessive bandwidth 
> >(LimeWire).  We had these workstations shut down the procees causing the 
> >excessive bandwidth, and the problem persists.  A quick scan of traffic 
using 
> >Ethereal indicated normal traffic.  
> >
> >During the troubleshooting process, I had the issue crop up and 
www.google.ca 
> >could not be resolved.  I immediately opened a command shell and tried an 
> >nslookup on www.google.ca.  nslookup reported that it could not connect 
to 
> >the name server by name, which was odd because it only knew the IP 
address so 
> >established the connection anyways.  Sure enough asking for details for 
> >google reported no results found.  A few minutes later, this process 
worked 
> >perfectly fine.
> >
> >So, my gut tells me that we have a name resolution problem, but the steps 
> >we've taken to rectify this don't seem to be making any difference.  
Seeing 
> >as everything works fine for at least some time, I'm assuming the 
switches 
> >involved are good.  (The switches are a combination of 10/100 and Gigabit 
> >Ethernet - 3Com switches with gigabit modules, and a fiber module for 
> >connecting a second building).  As near as I can tell, the network 
> >architecture is within usual standards.
> >
> >My next step is to grab an extended capture session using tcpdump or 
Ethereal, 
> >and track DNS traffic in detail.  We may also try setting up an internal 
Bind 
> >server to see if this helps.  Any other suggestions what we can try?
> >
> >Because it is intermittent, and doesn't seem to be following any pattern, 
this 
> >problem is proving difficult to nail down.  Hmmm...  This does sound like 
a 
> >possible hardware issue, but that shouldn't be the case....
> >
> >Thanks for any tips.  I can offer a little more about the architecture if 
> >needed....
> >
> >Shawn
> >
> >_______________________________________________
> >clug-talk mailing list
> >[email protected]
> >http://clug.ca/mailman/listinfo/clug-talk_clug.ca
> >Mailing List Guidelines (http://clug.ca/ml_guidelines.php)
> >**Please remove these lines when replying
> >  
> >
> 
> This message, including any attachments, is intended only for the 
> person(s) to whom it is addressed. If you received it in error,
>  please let us know and delete the message from your system. This 
> message may be confidential and may fall under the duty of non-
> disclosure. Any use by others than the intended addressee is 
> prohibited. Trema shall not be liable for any damage related to the 
> electronic transmission of this message, such as failure or delay of 
> its delivery, interception or manipulation by third parties, or 
> transmission of viruses or other malicious code.
> 
> _______________________________________________
> clug-talk mailing list
> [email protected]
> http://clug.ca/mailman/listinfo/clug-talk_clug.ca
> Mailing List Guidelines (http://clug.ca/ml_guidelines.php)
> **Please remove these lines when replying


--
-- WiredUnderground Corporation --
Tomorrows Products Today
http://www.wiredunderground.com


_______________________________________________
clug-talk mailing list
[email protected]
http://clug.ca/mailman/listinfo/clug-talk_clug.ca
Mailing List Guidelines (http://clug.ca/ml_guidelines.php)
**Please remove these lines when replying

Reply via email to