Specifications
Page 32 ExtremeWare v6.2.1 Release Notes
Release Notes for ExtremeWare v6.2.1
“down” after the first ping check interval of 30 seconds. The ping health checking can be disabled using
the command:
disable slb node {all | <ipaddress>} ping-check
Server Load Balancing with 3DNS
3DNS is used as a global load balancing and site redundancy tool. Additional information concerning
individual server health and performance can be gathered by 3DNS from the SLB services within the
Extreme switch for more granular and accurate decision making by the 3DNS device. These additional
functions apply when using Transparent or Translational modes. To enable responses to F5's 3DNS
i_query requests from Extreme's SLB services, use the command:
enable slb 3dns iquery-client
To see what 3DNS devices are currently communicating with the SLB enabled switch, use the
command:
show slb 3dns members
To disable responses to 3DNS queries, use the command:
disable slb 3dns iquery-client
The SLB enabled switch responds to directed queries from 3DNS. To direct 3DNS queries to the switch,
you add a "Big/IP" device to the 3DNS configuration. Encrypted communications with 3DNS is
currently not supported. These functions were tested with 3DNS v2.x and should function correctly
with v3.x.
Web Cache Redirection / Policy Based Routing
Flow Redirect Source Changes to Destination
In Flow Redirection, a source port rule will automatically be changed to a destination port rule when a
switch reboot occurs. (1-9YEP9)
Health Checking
Under very high sustained loads a Web Cache Redirect may fail and a cache servers set to the “down”
state and then brought back up. This only occurs during high loads for a duration of more than 2
minutes. The server will come back "up" immediately; however, during that time connections that were
established may be dropped due to a flushing of the associated IP forwarding database entries. A
"down" state is depicted in the log with the following message:
09/01/2000 10:51.56 <INFO:IPRT> redirect next hop test <ip_addr> changed to down
To use more than one next-hop server or router, the IP port must be set to any.
The FDB table will time out before the IPARP table on the ports connected to the cache servers. To work
around this configure the switch to have a higher FDB time-out than the IPARP time-out.
An ICMP PING of the next hop address is turned on by default and cannot be disabled.