Hi, I think this is a pretty interesting problem so I wouldn't mind to keep the discussion on the list.
At first I was a bit skeptical about you theory that the connected NIC cards can burn out a switch blade, but if nothing else ever fails, not one supervisor card, backplane, power supply, nor any of the PCs, always only the interface cards in the switches (is this the case?), then it must be something more specific than a surge on the power grid. I would think however that if this (ie. that powering up connected PCs at the same time can harm the switch) was a generic problem, then many people would face it (the infamous rolling blackouts here in California come to mind... :). One thing that could explain what you see is if your company used the same, exotic NIC model in every PC, and that model had some special behavior. If this is the case, then perhaps you can contact the vendor of those NICs to see if they know anything (of course I wouldn't tell, if I were them, hehe :). It might be also useful to pay attention to the time of the switch failures vs. when the PCs booted up. If there is a big gap, or especially if the failure happened first, then it's less likely that it's the NIC cards' fault. Apart from this, my only suggestion is to send back the switch blades to the respective vendors and ask them to test out which component failed. That might give a hint of the cause. What kind of Cisco switches do you have? Btw, when you say that the cards have "earmarks of an electrical problem", do you mean that you can actually see signs of burns on the card? Thanks, Zsombor At 02:46 PM 7/11/2003 +0000, [EMAIL PROTECTED] wrote: >This is completely off topic and, in my case, of an operational nature, so >I need to apologize in advance. Unless there is some expressed interest >please respond to me privately if you would, thanks. > >We have suffered over the last couple of years what I would consider an >unusually high number of blade failures in various switches (from several >vendors including Extreme and Cisco). They have the earmarks of an >electrical problem, that is, supplied current spiking and valleying. For >instance, where I am located - Northern Virginia - we had some major storms >move through last night. This morning I had two blade failures in two >separate closets. These closets all have UPS units (usually more than one) >and the switches, all of which have dual power supplies, and are split with >one power supply going to one UPS, the other to another UPS. Granted both >are in the same room and usually plugged into the same circuit. Seemingly >this would eliminate, or at least mitigate, the bulk of any electrical >fluctuations in the closets. The UPS's have been tested with various >diagnostic tools and they appear to check out fine. > >My question is this. The failures seem to occur more often on densely >populated blades (isn't that always the case?). Many people in the building >leave their PC's on at night. When the power goes out in this building, the >PCs in most areas will also power down. When power is restored, even if it >just flickers, all these machines will boot at the same time. Is there any >chance that this surge - if there is one - of current coming from the >rebooted PC NIC cards could cause these issues? Or is there even a current >that flows back from the NIC's (it would seem logical, but I am not an EE). >I do realize that I may be grasping at straws. > >Any thoughts, and again I apologize for the off topic content. There are >many bright folks in this group so I thought I would ask. Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=72160&t=72151 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
