After one power supply failure in the IGX8420 the "dsppwr" display no temperature and empty power supply. We replaced the damaged power supply so the equipment keeps with two power supply but it is not showing when we perform "dsppwr". The IGX is working properly, all cards and buses are ok.
What it could be?
I do not have administrative experience with IGX. I have only been a user loging in to check network status, but I decided to respond because not many people have experience with such equipment here, unless you are very lucky and someone from cisco responds.
This might be a cosmetic bug, in the sense that the power supply is working, but the code that displays the status to you got somewhat confused and shows garbage. If it was just a router, I would recommend some reboot, but IGX's have so high uptimes that is purely a pity to do such a thing for no good reason. Check any switch leds that would confirm your new power supply is working. Consider removing and re-inserting the new power supply to the IGX. Also try to check the same power supply in an another IGX if you have one. Consider also contacting cisco about this. They can login to the IGX to see what's going on.
Thanks a lot for your return.
I chose to open a SR because I agree with you it is a complex thing to discuss in a forum.
But I appreciated a lot your position.
I just wanted to ask if you have followed all the procedures recommended for power supply replacement, because I see something in the documentation for replacing an AC power supply on the IGX 8420:
"Check the supply-monitoring circuit on the SCM. First, enter the resetcd 0 command at the control terminal (this resets the power supply monitor on the SCM). After waiting about 10 seconds or more, enter the dsppwr command and see if the FAIL indicator for the supply comes on again."
I say this because IGX's can be quite less smarter than routers sometimes, and need you to tell them everything! We used to argue with an IGX expert colleague from time to time about this :-)
Yes, We provided the resetcd 0, the switchcc and after replaced the SCM, but no success. Then we open a SR and we will execute other test.
I can post here the solution when I have it and we can share with everybody.
BTW, thanks a lot for your reply.
Ok, then. It seems to me you did everything by the book. This might be some type of failure associated with the original power supply failure (some type of a unique issue). Good luck!
Just in case, consider a check of your environment's electrical characteristics. We once had an MGX card that got "burnt" due to inappropriate grounding. The colleague I referred to previously told me that in such cases of improper grounding you could see all kinds of weird issues, such as a trunk going up and down while the excess charge discharges. Electricity can be very tricky sometimes. I am thinking this because such electrical environment issues might have caused the original power failure in the first place.
You might be interested in this article (which reminds me that we once also had some LAN switches burnt after a lightning strike):
and this documentation:
p.s. Most people do not suspect such issues, unless they get literally burnt. Even people that their main job is to handle those issues are not so aware. Our operators were astonished when my colleague determined the root cause of the MGX card failure. I was astonished too. Although I am an electrical engineer, I have not worked on the specific field to see with my own eyes the severity of such issues. I have been reading afterwards a book that was saying that it's a miracle all people in Athens have not been electrocuted yet by the huge amount of improper groundings!
Ok, people, speak out the truth: How many of you wear antistatic straps when you are supposed to? Even this small thing is commonly neglected. People just don't believe such things can cause real issues, but they actually can. The problem is that you do not see a log saying "memory failure because you could not wear a simple antistatic strap" and you assume it just doesn't work anymore for an undetermined reason (UFO or something), which is enough to make you stop wondering. When I started working, an experienced engineer (who was also a good one) made fun of me because I wore an antistatic wrist strap! I guess cisco would have much less RMA requests if proper procedures were followed, and I guess we do not have many chances of learning if grounding was good in this case.