When consoled into a Cisco device, I issue the wr t command to display the configuration. Absolute Telnet displays about 26 lines of text then returns to the switch command prompt. There should be many more lines of text displayed. Is this a configuration issue with AT?
Here is a sample:
Switch#wr t
Building configuration...
Current configuration:
!
version 12.0
no service pad
service timestamps debug uptime
service timestamps log uptime
service password-encryption
!
hostname Switch
!
no logging console
logging facility local2
logging source-interface VLAN1
logging 10.90.80.10
enable secret 4 $1$zdbj$vmertiymETy566IjnFcI.
!
!
!
!
!
clock timezone EST -5
clock summer-time EDT recurring
Switch#
[size=1][ January 21, 2005, 09:49 AM: Message edited by: Brian T. Pence ][/size]
This likely stems from the flow control issues raised in your [url= http://www.celestialsoftware.net/absolutetelnet-general/373-laptops-without-serial-ports.html ]other thread[/url].
What is the flow control setting in this situation?
Which Cisco device is this?
Brian
Any response?
To be honest, I cannot remember. I tried so many different switch types and flow control settings for the other thread...
I'll do some more testing next week and give you some definite answers.
I tried this with a Cisco 3550 switch that does not exibit the "flow control" (ie works fine if flow control is set to None or Hardware) issue and Absolute Telnet still displays about 26 lines of text then returns to the switch prompt. I tried this having flow control set to none and hardware, same results.
I also tried this with a Cisco 2924M-XL switch with flow control set to hardware and it too displays about 26 lines of text and returns to the switch prompt.
Were these tests done with the USB adapter or a 'real' serial port? I hate to have to be so specific, but there's obviously something tricky going on here.
Hardware flow control requires a proper cable and support for it at both ends, which Cisco doesn't support from what I've read.
You might try xon/xoff flow control as an alternative.
These tests were done with a real serial port, no USB adapter involved.
I just tried Xon/Xoff, no joy.
An update....
Reading the specific [url= http://www.cisco.com/en/US/products/hw/switches/ps700/products_tech_note09186a008010ff7a.shtml ]Cisco support page[/url] for the 3550 with regard to console support, I see the following, which surprised me:
Follow these steps to connect the PC or terminal to the switch:
1. Be sure that you have configured your PC or terminal emulation software to communicate with the switch via hardware flow control.
2. Configure the baud rate and blah blah blah...
Ok, back to my earlier post, then. Hardware flow control requires a proper cable. The Cisco page goes into some detail in this regard...
These switches require the use of either a rollover or straight-through RJ-45 cable to connect a terminal to the console port. The cable you use depends on the type of Supervisor Engine and other factors.
and references a [url= http://www.cisco.com/en/US/products/hw/routers/ps332/products_tech_note09186a0080094ce6.shtml ]Cabling Guide for Console and AUX Ports[/url]
I have all the cable and connectors mentioned in the document on hand, I'll try a few different combinations and report back.
It's a bit confusing because this document clearly states that the console port requires hardware flow control. I've read in other places on the Cisco site in equally clear terms that the console absolutely does *not* support hardware flow control. Perhaps it is very specific to the hardware and maybe to the IOS version as well.
I've also read some references to attaching to the 'Aux' port instead, as it does support hardware flow control. I'm not sure if your hardware has an 'Aux' port or not.
However, with no flow control enabled, or using the wrong cable, you wind up with exactly the situation you describe with truncated output. Worse yet, it is possible to paste commands to the console that are imcomplete or garbled.
It is possible that with other software set to 'none' flow control, you may not notice the problem, at least not with a small amount of data. However, it is still there. It's not a problem that can be ignored or it will come back to bite you. Proper flow control on a serial line is just a fact of life.
Brian
Any update?
Brian
I had some time today to do some more testing.
I tried three different cable combinations, all three gave the same result. I also tried two different switches, a Cisco 2924M-XL and a 3550, both of these exibited the truncation problem. I also tried different flow controls, which made no difference.
I have discovered what seems to cause the truncation. In hyperterminal, when the screen is full, the router shows:
--More--
This prompts the user to hit Enter to show the next line or hit SpaceBar.
It seems that Absolute Telnet is returning the the Switch# command prompt when it encounters the --More-- prompt.
I am using 2.13, which is the latest version I am licensed for.
That's very strange. AbsoluteTelnet would not of it's own accord skip the '--more--' prompt. Terminals are not very bright and make no assumption about what they're supposed to display. They only display it as it happens. Either the --more-- isn't being sent or it's being sent and lost because of bad flow control.
But even if a flow control problem were to lose data and not display the '--more--', you would still have to hit enter to get the display to continue, etc... And you're saying that Absolute goes *straight* to the next 'Switch#' prompt?
It's presumptuous for the host to even make an assumption of the number of lines on the page. This sounds more like what might happen when you telnet or ssh into the switch, where the host knows the number of lines of the terminal. You *are* testing using serial communication with the same parameters for both Absolute and Hyperterminal, correct?
Have you considered using telnet or ssh to connect to the switch? Does your version of IOS support either? These protocols are buffered at the TCP/IP level and shouldn't have any of these flow control issues.
Would <broken link removed> on ebay suffice for testing? It has an EN-DC at the end of the model name. I don't know if that is important or not. 3 hours left to go and the price is still reasonable.
Brian
Sorry, I cannot access Ebay from work. I think the DC indicates DC power rather than AC. That may be problematic.
I could give you VPN access to my network so you can see the problem. Would that help?
Send me an email and we can schedule some time.
