I think I found the problem. I tested this with a regular computer with a serial port, so the issue has nothing to do with my IBM serial adapter.
Hyperterminal will work with Cisco or 3Com devices if flow control is configured to "none" or "hardware".
When consoling to a Cisco device, (2924XL switch in this case) using Absolute Telnet, flow control must be set to "hardware". If it is set to "none" no output is displayed on the screen.
According to documentation on the Cisco site flow control should be set to "none". [url= http://www.cisco.com/en/US/products/hw/switches/ps607/products_tech_note09186a00800946e5.shtml#consoleaccess ]http://www.cisco.com/en/US/products/hw/switches/ps607/products_tech_note09186a00800946e5.shtml#consoleaccess[/url]
When consoling to a 3Com device, (3500 switch in this case) using Absolute Telnet, flow control can be set to "none" or "hardware".
----- Message Text -----
That's very odd. Now, I need a Cisco device to test!!! Do you see this
as common across all Cisco devices or specific to the 3924XL?
This was a 2924XL switch.
I have other Cisco devices, I'll check them next week and let you know.
Preferably something cheap I can pick up on ebay that exhibits this same behavior.
And this behavior is the same with the USB adapter?
Brian
I have done some testing.
These Cisco devices do not exibit the behavior:
Routers:
3640
2600
2501
Switches:
3550
These Cisco devices do exibit the behavior:
WS-C3512-XL-EN
WS-C2924M-XL-EN
Could this be caused by a configuration setting on the switch perhaps? I'm pretty sure I have seen this on switches that had a default configuration also though.
How about this one?
<Broken Link Removed>
Yes, that is the same model I have.
Brian, I would feel really bad if you purchased a switch to test with and the problem turned out to be "user" error...
I don't have any spare 2900 series swirches, or I'd send you one.
Then I'll sell you the switch!!!
I didn't win the auction, btw.
Brian
BTW, looking at the pictures of this thing..... where's the serial port?
Brian
The "console" connection is on the back, it has a RJ-45 type connection. Cisco supplies a cable with a serial connector on one side for the computer and a RJ-45 on the other for the switch.
Dude,
I did some more testing with this over the weekend and I have some more information. I found a difference in how hyperterminal and AbsoluteTelnet operate using the 'none' flow control setting.
There are flags you must set when doing hardware flow control to indicate to the other side that it is ok to send data. When AbsoluteTelnet uses 'none' flow control, these flags are defaulted to their unused (disabled) state. Hyperterminal, on the other hand, defaults these to an enabled state. Now, in a perfect world, it wouldn't matter because by choosing the 'none' flow control option, we are stating that the other side doesn't care and isn't looking at these flags. Apparently in the case of the Cisco, this isn't true.
In this state, apparently the switch is obeying the hardware flow control flag and is NOT sending data to AbsoluteTelnet, but *is* sending data to HyperTerminal. Based on this, it is becoming more and more obvious to me that the Cisco has some ambiguous behavior (and documentation) in regard to flow control.
I have no problem changing the default behavior of AbsoluteTelnet to match Hyperterminal. It will make this situation work, but tends to mask Cisco's obvious dependency on hardware flow control.
As to how this relates to your other thread about output truncation, that is yet to be seen. I'll have to make this change for you and let you test it to see.
Sounds good, just let me know when its ready for testing. Maybe this will kill two birds with one stone!
Ok, I think I've got it. The new version differs in this way:
With AbsoluteTelnet flow control set to 'none', bot the rts and dtr signals are held high, indicating to the other side that we are ready to receive. If the other side is configured for no flow control, it never looks at these signals. If the other side is set for 'hardware' flow control, then these signals allow it to send without stopping (effectively no flow control).
With AbsoluteTelnet flow control set to 'hardware', cts and rts signals control the flow of data, while the dtr signal is held high. Try it in 'none' and 'hardware' mode and compare the results against Hyperterminal in the same modes and let me know if there's a difference in behavior.
Thanks, Brian
Version 3.68 fixed the issue.
I can use "none" or "hardware" flow control methods on the 2924M-XL switch. Both produce output equivalent to hyperterminal.
Thanks for the excellent support!
Great job Brian!
I'm glad we finally got to the bottom of this one!!!
Hyperterminal has its roots in serial communication, so I trust its behavior is probably more accurate. AbsoluteTelnet started as a telnet client and then added serial/modem/ssh1/ssh2 later.
I'm going to run this version by some other users who also do serial communication just to make sure we didn't break anything...
Thanks for you help and testing! I *am* glad I didn't have to buy any hardware!
Brian
