Hot- Standby feature

Question

Well, I can't seem to get any response from SW on my ticket that has been open for over a week now on this problem so hopefully this forum can give me some assistance.

We recently installed the Hot-Standby feature.  We have found the feature only fails over one time and then we have to stop and restart the NetPerfMonService to restore it's ability to fail-over.

Here is how we tested fail-over (Poller1), fail-back, fail-over (Poller2).
 
1. Stopped/restarted NetPerfMonService on Hot-Standby system (HS poller) to clear previous stopped condition and start new.  
2. Stopped NetPerfMonService on Poller1 system.
3. Hot-Standby system detected failure and triggered Fail-over.  Fail-over event event logged and Fail-over action notification sent.
4/27/2007 11:15 PM  Primary Engine (Poller1) Failed. Hot-Standby Engine (HS poller) taking over.  
 
4. Let HS run until 4/29. Fail-over action notifications resent on hourly basis.
5. Restarted NetPerfMonService on Poller1 system.
6. Hot-Standby system detected poller restored and triggered Fail-back.  Fail-back event message logged and Fail-back action notification sent.  Monitor Polling Engines status as Hot-Standby Dormant (Poller1), then just to Hot-Standby.  Poller1 to Running.
 4/29/2007 10:46 AM  Hot-Standby Engine (HS poller) shutting down. Primary Engine (Poller1) active.

7. Hot-Standby status (in Monitor Polling Engines) returned to Hot-Standby, Poller1 Running.
8. Stopped NetPerfMonService on Poller2 system.
4/29/2007 10:06 PM  Network Performance Monitor Engine (Poller2) Shutdown

9. Hot-Standby system detected failover but no Fail-over event message logged.  HS only triggered fail-over action notification. Monitor Polling Engines status as Hot-Standby (Poller2), Poller2 Stopped. 
10.  Maintained this state for approx. 22 minutes.  No database statistics collected for Poller2 nodes.  HS seemed to be stuck at 25% CPU and memory usage remained unchanged at 202,240K.
11. Restarted NetPerfMonService on Poller2.
4/29/2007 10:28 PM  Network Performance Monitor Engine (Poller2) Started

12. Hot-Standby system detected fail-back but no Fail-back event message logged.  HS only triggered fail-back action notification. Monitor Polling Engines status as Hot-Standby Dormant (Poller2), Poller2 Running.  Then Hot-Standby (HS poller) went to Stopped.  HS service still stuck at 25% CPU and memory usage unchanged at 202,240K
13. Could not stop NetPerfMonService using Shutdown NetPerfMon on Hot-Standby (status Stopping).  Had to kill service and restart to return HS to operational Standby status.

Anyone notice this problem with their Hot-Standby system?

I also have questions on website failover.  How does HS detect website failure and take-over?   I seem to recall this was listed as one of the features of Hot-Standby on the HS fact sheet.  Interesting! I can no longer find it on SolarWinds website.  I wonder?

jtimes · Answer

My HSB really hasn't worked correctly since 7.8.4 (Jan 2006), of course my Orion version is now at 8.0.2 (Mar 2007). Even when HSB and Orion where at the same release, HSB was questionable at best.  I have come to the sad conclusion that HSB is really not worth the maint cost (right now.)

The HSB server has to be setup as a HSB to the Orion master server. Here is a kicker on this (Same as remote admin pack)...  Almost everything you change on the Orion server has to be manually syc with the HSB (maps, views, reports)

Right now HSB and Remote Admin Pack are a frustrating, not to mention the lack of maint and support for both!

sorry to rant on your post...