06-06-2012 07:30 PM
I got problem where I cannot ping server from another server, both connected to the BigIron RX 16. It occur in random time, and in random server. When it occur the ARP entries are not exist on server's ARP table.
I have checked the server connection to RX but there is no problem, each server can ping the BigIron even if the problem occur.
I attached the detail of topology.
I have try to "show arp log", the entries show some "pending" in status column, but i don't understand what does it truly means. Is this entry have something to do with this ARP problem? or is there something from Big Iron configuration that reject ARP broadcast? or Is it have nothing to do with BigIron RX? Is there any other command that could help me troubleshoot this problem?
Please help, I will be thankful for any respond.
06-06-2012 10:58 PM
I would check interface statistics and counters if there are any dropped packets, errors and also look port utilization etc.
if you see no problem on these while error occurs I suggest you to look if you can still ping same ip address of the server while server is not connected (there may be a duplicate ip). show log output may help you if there is.
if you have this problem only on specific IPs then you may have a problem on related ethernet NICs.
hope this helps,
06-06-2012 11:27 PM
Thank you Serhat, I've open case this to brocade support, they said that the problem is in RX module, due to bad TM and DRAM error. They suggest me to replace it, however i don't really know what TM and DRAM mean and what is this correlation with the problem. Could you help me?
And again, thnks for your response Serhat.
06-07-2012 12:06 AM
RX has a close fabric architecture as below. TM means traffic manager which resides in the Clos architecture uses data striping to ensure optimal utilization of fabric interconnects at all times. This mechanism always distributes the load equally across all available links between the input and output interface modules by using fixed-size cells to transport packets across the switch fabric and it uses Local DRAM as a fixed memory. according to TAC response you have an hardware issue and need to change your linecard to fix this problem.