locked
UAG array "drain stop" query RRS feed

  • Question

  • We have recently upgraded our stand-alone UAG server to an array with 1 additional array member.

    It appears to be working fine though I don't understand it's logic behind the session balancing sometimes!

    We wanted to apply SP1 Update 1 (to resolve the issue of having lots of dead sessions in WebMonitor) so we followed the instructions in an online TechNet article on how to apply this patch in an array scenario.

    The "Drain Stop" feature doesn't appear to work very quickly.  Even after an hour I'm still seeing new sessions being balanced to the array member that is supposedly being drained.  The webmonitor confirms the NLB array status is "draining".

    I noticed this yesterday when I set the array manager to drain.  When I checked it this morning the legitimate sessions were indeed drained (ignoring the multitudes of dead sessions) so it appears that the drain stop feature did work but just took a long time!

    Can anyone comment on how the drain stop feature behaves in their environment? 

    Or can anyone tell me how long it takes for the server to stop receiving new sessions after enabling drain stop ?

    Wednesday, February 8, 2012 3:21 PM

All replies