I am having a problem w/ two css's working in redundancy modes.
a | b |
c | d |
SW1 --- SW2
| X |
NC = NetCache
WSF = Web Server Farm
SW = switch
The two switches are connected using isl trunking and VTP. The normal traffic
is configured so the traffic flows a>c>nc for all port 80 traffic. The NC will
use a secondary VIP on the CSS to load balance it's requests to the web farm.
When we break link A, we have keepalives set so that all of CSS1's connections
will failover to CSS2 and traffic would flow b>d>NC. This is working properly
and is not an issue.
When link A reconnects, the the virtual interfaces and routers all fail back
over the the MASTER CSS1 and traffic flows a>c>NC again. This is working
properly and is not an issue.
When we break link C, the CSS1 looses it's keepalive to SW1 and fails over to
CSS2. This is working as well. When we reconnect link C, we loose connectivity
to the secondary VIP (the VIP that the netcache used to get content from the
origin servers) The netcaches are set up in a virtual interface configuration,
but testing has proved these not to be the issue. When the Secondary VIP fails,
so do the NetCaches, because they can't get reach the origin servers. The CSS
shows mastership for all content rules. We can ping the netcaches, web servers,
and other network devices.
When the client tries to browse the secondary VIP directly, you get a 'web site
found, waiting for reply", but never a reply, just a page timeout error. When
you try to browse the NetCache directly, you get the same error as you would the
When we unplug and plug in links C and D, things go back to normal and the
Secondary VIP works again.
The Load Balancing Mailing List
MRTG with SLB: http://vegan.net/MRTG
Hosted by: http://www.tokkisystems.com
This archive was generated by hypermail 2b30 : Wed Jan 30 2002 - 19:01:46 EST