v1.4.2 Bug Reports and Comments

[lligetfa]

However trying to upgrade to wispswitch-1.4.2 and it stops half way through and stops doing the upgrade.

I would try downloading the file again from another computer and compare the two in case it was corrupted.

[AlexC]

I've discovered that if I use a password containing the character '$' I am unable to login to the serial console. The same password works fine for ssh and the web interface.

Edit: I've determined that you can login via the console if you escape the $ characters, ie put a \ before them so for example you type '\$ecure' to use the password '$ecure'.

[Eric Stern]

I've discovered that if I use a password containing the character '$' I am unable to login to the serial console. The same password works fine for ssh and the web interface.

Edit: I've determined that you can login via the console if you escape the $ characters, ie put a \ before them so for example you type '\$ecure' to use the password '$secure'.

This will be fixed in the next release.

[semakka]

WS-24-400B - upgraded from 1.3.9 and lost communication with the switch, looks like some VLANs are messed up.
I've also lost comm with one unit powered by the switch but the rest seem to be OK.
any ideas before sending staff on sine?

[sirhc]

Possibly a corrupt config file?

v1.4.3rc4 and newer we are adding in code to deal with corrupt config files.

Sadly you may need to factory default and set it back up.

[amp]

I'm not sure this is reported yet but its so small display issue that i guess it probably wouldn't be. I have 100M-F device plugged into a gigabit sfp glc-t. Shows green on the interface, and it knows its at 100-F. WS-8-150-DC on 1.4.2. Figure it makes more since for the display of the speed to match the connected speed and not the port capability.

[highlands]

last night I upgraded a WS-12_DC D model from 1.3.9 to 1.4.2 and did not notice till this morning that there were tons of CRC errors exactly since I did the UG. was there a stack change for the Ethernet ports?

Any thoughts?

I just rolled back to 1.3.9 and all the CRCs stopped!

John

[sirhc]

last night I upgraded a WS-12_DC D model from 1.3.9 to 1.4.2 and did not notice till this morning that there were tons of CRC errors exactly since I did the UG. was there a stack change for the Ethernet ports?

Any thoughts?

I just rolled back to 1.3.9 and all the CRCs stopped!

John

No, "we do not have anything to do with the switch core code itself. The switch core itself is a self contained chip and we have nothing to do with the code that runs it.

All we do is write the UI/CLI/STATS and control the Daemons that control things like STP/LACP and then we only have a limited affect on them such as steering a car, all the steering mechanisms are already built into the car.

I would say if anything it was when the device re-negotiated the link the link was poor sort of like when you call someone and the line is crappy so you hang up and call back.

I would have tried rebooting the device or disabling and enabling the port.

But once again "we have nothing to do with the switch core code and packet handling per say, this is all done in the switch core chip and it has it's own firmware which we do not write."

[highlands]

Thanks for the update. I will admit it is a long cable run with double jacked STP CAT5E over 200 feet. But that should not be an issue. When testing the cable I get the normal alert about Pairs 3 and 4.

I'll try that late tonight and see what happens.

[sirhc]

The cable diagnostics tool is a CRUDE TOOL only meant to find shorts and cross shorts and report length.

It has no ability to test the cable quality or ability to handle communications as Ethernet cable testers costs "start" at several thousand dollars.

Just because the cable diagnostics reports OK does NOT mean the cable is not stretched or has bad attenuation between pairs. As I said a good cable tester is several thousand dollars.