Show commands

Hi all,

I have a problem at work that is getting a little bit personal now. We have an application that spans our local network (the web part) and the database where the usernames come from (over a VPN, 100Mb).

This application fails every morning with an unspecified error and we are told without a doubt it's a network issue. Me being the network engineer have exhausted every possibility going looking into this.

I have been thinking, is the 4506 getting over-loaded, or are the un-contended blades using up too much of their backplane? Is there a show command that will tell me how much bandwidth is being used for the blade involved? I've been looking for this fault now for ages and am now lost.

I have Cisco Works LMS 2.6 if that helps, but I'm not a complete expert on the fault-finding side of it.

Does anyone have any ideas?

Kind regards

James M

Reply to
Loading thread data ...

What is failing? Logon? Database replication? Transfer? Queries or transactions? In my years, I have seen a handful of switch problems, but mostly these are related to bad modules, bad ports, bad ASICs, etc. Generally speaking, your switch will not be getting overloaded unless you are running some hardcore stuff like multiple gig ether- channel database replication, and even then, multiple sessions. Not to say that folks can't nail a switch with poor configuration (spanning vlans or trunks, etc).

Where does the VPN go? What kind of bandwidth is it? 100 meg tells me the fast ethernet connection to the VPN, but does not tell me if that is LAN or WAN and where the other side of the VPN is. If this VPN is over a internet T1, and the transaction that fails is a gig file transfer to kick off the day, then you may be hitting a bottleneck, but generally speaking, nothing should 'fail'. I guess if someone is nailing the pipe and this is a sensitive application, it could be timing out, but that is what needs to be analyzed.

Is the transaction that fails run at other times of the day? Does it complete successfully? How do you get beyond the error, wait and retry? Once we see some answers, we can hopefully go to the next steps...but I would very much doubt that you have a switch backplane or hardware problem here.

Reply to

Have you been able to predict the time that it fails based on past failures. You may be looking at a time out of some sort. My guess would be the VPN. Is it an IPSec VPN? What kind of LAN network do you have? If you have any LANE, I've known some weird things to happen sscop.

Have you tried implementing QoS on the VPN? Is there any other traffic in compitition for bandwidth with the VPN?

I have to agree with Trendkill... it's not a switch problem - unless all the applications that flow through the switch fails at the same time.

It seems there are more questions than answers here.


Reply to


Thank you for the replys. This is my setup,

two clustered 6509s feeding with etherchannel a 4506. The core via a firewall has a 100Mb link to another site. The application that had a problem was a web server based on our site, that got the username and password from the other site. Things were timing out when connecting early in the morning.

On chatting with a few of our systems guys, it was apparent there were a few more problems than just this, and it all started happening the time our SAN was introduced. I had a look where our SAN was patched onto our

4506, and all the network cards for it had been put onto a 48 port, contended blade and not one of the uncontended blades. Basically lots of our servers weren't spread out evenly enough over the switch, so I had a sort out, and so far, things are a lot better.

Next time I ask a question I'll supply a lot more detail, but at the time of writing, I just wanted to get something down.

Many thanks


J.Cott>> This application fails every morning with an unspecified error and we

Reply to
James Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.