Have never seen anything exactly like what your describing. However, Unless I am reading your posts poorly, it sounds like you are saying that as long as you leave the front office workstations offline you have issues. How many front office workstations are there? Is it feasible to check there ip settings one at a time manually? Sounds a LOT like someone has statically assigned an ip address on a machine rather than letting the dhcp server take care of it. If they have used an ip address that is already assigned elsewhere then that can cause issues .......if they have assigned an ip address that is already assigned to an ethernet port on a router then that can cause LOTS of problems. Having this issue come up after the network builds up a certain amount of load may simply mean that someone finally booted this screwed up workstation that has a errant ip address statically assigned. Also, do any of your computers have multiple nics? I've heard of situations similar to what you are describing, when a computer has multiple nics......with each nic assigned appropriately for the settings needed to allow it to participate in different vlans.....with routing turned on ......and those the computer starts advertising itself as a router would....the other routers begin injecting this information into their routing tables dynamically....and poof........and to make the system even more difficult narrow down, microsoft has it set up so that only the gateway used on the last nic activated will be used, so if someone is troubleshooting and act/deact the multiple nics in a given machine.....different gateways will become active at any given time thus the symptoms are consistent. Hope this gives you some ideas, Cleve
At 10:43 PM 12/5/2003 -0600, you wrote: >Last night, the network was humming along just fine. > >This morning, It appears that once the network got loaded down somewhat >and started doing the freakin' thing again. > >Except, this one situation was slightly different than yesterdays. > >Using EtherApe, the network (from an ip viewpoint) would grow and shrink >as would the ability >to hit certain machines via tcpip. Segments would join and leave the >network at random intervals. > >EtherApe also allows monitoring at the ethernet level. From that >perpective, the network >extents were complete and static.. IPX packets could bounce around all >regions with no problem or loss.. > >As far as the machines go, some routers (but not all) would do the low cpu >utilization thing, then jump to nearly 100% utilization..and then go back down. > >We isolated some hardware - just enough for the plant to run; the front office >is disconnected except for the servers. The network is stable now (at >least the half of it that is still juiced) > >Weird that it would do that ...All of the dark routers excpet for one had >a solid and correct configuration. The one bad apple consistantly >misreported status with what was configured. With the bad one dark, the >other ones freaked as well (even after cold starts) until their power was >extenguished as well.. > >Waiting on spares ... to tide over until the network is replaced with a >new one soon. > >I'd like to know if anyone else has seen anything remotely like this.... > > > >_______________________________________________ >General mailing list >[email protected] >http://brlug.net/mailman/listinfo/general_brlug.net
