Sunday, September 24, 2006

Network Down - 2006/09/20, night time

Cause:

  • large file being transferred from BACH to a certain number of clients in B403, causing failure on B403 VLAN.
  • system reboot didn't fix the problem; the reboot caused failure on all other VLANs
    all switches were OK; "Directed Broadcast" (checked enabled) was not the reason for the failure this time (DHCP won't be able to give IPs if checked disabled)

Diagnosis:

Before BACH restarted:

  • all switches restarted, and could all be ping thru 172.16.0.0/24 subnet
  • 172.16.14.0/24 (B403 subnet) was the only one that failed PING test; all other VLANs
    ping'ed

After BACH restarted

  • 172.16.0.0/24 was OK
  • all VLANs failed; DHCPSRVR couldn't receive broadcast from clients
  • NIC for VLAN connections was checked by assigning an IP on 172.16.0.0/24 and switch port swapped to the default LAN; could ping 172.16.0.0/24 and thus NIC failure was not the issue

Fix:

  • before, NETINFO.CFG (\\BACH\SYS\ETC\) was configured manually. The whole thing is re-configured using INETCFG.
  • all configs on INETCFG were first removed; and all NIC settings were reconfigured BASP and VLAN settings were done on INETCFG
  • INETCFG doesn't bind IPs on BASP and has to be done manually on the console using the
    following command:
    BASP BIND BASP_V2_EII B57_2_EII
  • An IP has to be assigned to the VLAN NIC, or else BASP cannot find the NIC name and
    BASP BIND will not work. The VLAN NIC is now assigned with 192.168.1.1/24

Things to do:

  • remove the IP on the VLAN NIC (ie. 192.168.1.1)
  • design a way so that BACH reboots properly and without manual intervention

0 Comments:

Post a Comment

<< Home