jeff.stewart replied on Tue, Oct 14 2008 2:19 PM

rated by 0 users

 Did you ever hear anything about this problem?  I'm seeing the same problem.

Mark Wiggans replied on Tue, Oct 14 2008 2:47 PM

rated by 0 users

Verify all of the Clocks of the Polling Engines, do the Polling Engines and Orion come within 2 Minutes of the SQL server, as well as same Day, Year, and Time Zone?

maz.kassam replied on Tue, Oct 14 2008 3:10 PM

rated by 0 users

Hi Mark,

I have checked the time, time zones and dates on both Orion servers and our seperate SQL server. All of them show the same time, time zones and dates.

I also checked the last database synch times on both Orion servers and compared that to the time on the SQL server - this also seems to be exactly the same.

 Any other ideas for this?

thanks,

Maz

maz.kassam replied on Tue, Oct 14 2008 3:15 PM

rated by 0 users

Additionally, I have reduced the number of interfaces that I am monitoring on both Orion servers. I used to have to re-start the Orion servers once every day or two, now there are some times that I see both Orion servers will run fine for about 4 days.

Thanks,

Maz

Mark Wiggans replied on Tue, Oct 14 2008 4:17 PM

rated by 0 users

 Have you tried using the poller Load Balancer?

maz.kassam replied on Tue, Oct 14 2008 5:20 PM

rated by 0 users

Our setup so far has been that our head office and critical locations are monitored by one server, and all other sites are monitored on the other server. Maybe I need another poller since reducing the number of interfaces per server has helped a bit? My main Orion server that also houses the Orion web site now only has 3222 interfaces, my second server has no web site and has 5000 interfaces.

I was just wondering if the problem was as basic as both engines showing up as Primary in their descriptions - is this correctly set up?

Mark Wiggans replied on Tue, Oct 14 2008 5:44 PM

rated by 0 users

maz.kassam:

 

Polling Engine on ORN02AS299
Last Database Sync	Now
Network Elements	387 Nodes, 8526 Interfaces, 0 Volumes, 8913 Total Elements
Running Since	9/3/2008 4:23:16 PM
Polling Completion	99.50 %
Operating System	Microsoft Windows NT 5.2.3790 Service Pack 1
Service Pack	Service Pack 1
Package	Orion NPM Polling Engine v9 SLX

 

maz.kassam replied on Tue, Oct 14 2008 6:03 PM

rated by 0 users

No we updated to Windows 2003 this year, still using SP1. This is what I see from the DB query:

 Orion NPM v9 SLX Microsoft Windows NT 5.2.3790 Service Pack 1 Service Pack 1   Primary

Orion NPM Polling Engine v9 SLX Microsoft Windows NT 5.2.3790 Service Pack 1 Service Pack  Primary

rgward replied on Tue, Oct 14 2008 7:28 PM

rated by 0 users

Mark Wiggans:

 

Polling Engine on ORN02AS299
Last Database Sync	Now
Network Elements	387 Nodes, 8526 Interfaces, 0 Volumes, 8913 Total Elements
Running Since	9/3/2008 4:23:16 PM
Polling Completion	99.50 %
Operating System	Microsoft Windows NT 5.2.3790 Service Pack 1
Service Pack	Service Pack 1
Package	Orion NPM Polling Engine v9 SLX

 

Mark,

Looks to me like the fix for Case 18005 (Sept 2007) which I reported under v8.5.1 has never been implemented in v9 as promised it would be, at the time, in version 8.6 which obviously never materialized.  Can you confirm?  If not, why not?  I think this is what the issue is with the Operating System being reported as

kvanderploeg replied on Tue, Oct 21 2008 8:38 AM

rated by 0 users

NetPerfMon50Percent.JPG

I've been having the same problem.  I've also notice on the polling engine that loses sync, that the NetPerfMonService is stuck at 50% CPU utilization.  Curious to see if yours is doing the same thing, maz.kassam. Here's the support request I sent to Solarwinds:

I have a server running Orion 9 that has the database and is polling 147 nodes with 2085 interface elements. I also have two additional polling engines. The first polling engine has 593 nodes with 12886 interface elements. The second polling engine has 662 nodes with 13989 interface elements.

My primary engine and first polling engine stay syncronized with the database. The second polling engine falls out of synchronization within approx. 1 hour of rebooting the server or restarting the netperfmon service. Both polling engines are located in the same network, were set up the same way and are the same spec of server. I have also noticed that when the second polling engine falls out of sync, Task Manager shows the NetPerfMonService stuck at 50% CPU utilization. We have tried rebuilding the server from scratch, but the same problem remains. We would appreciate any help you could give.

Kent

maz.kassam replied on Tue, Oct 21 2008 11:18 AM

rated by 0 users

Hi Kent,

 Thanks for your reply. I have not noticed the CPU uitilization gets stuck at 50% in my setup. I will keep an eye out for that though.

For your setup, do all your Orion servers show up as primary in the database manager? I am still trying to find out if my setup is correct. Both of my servers show as Primary. I am assuming that if both servers are Primary, maybe they are both trying to write to the database at the same time and this causes the sync issue?

kvanderploeg replied on Tue, Oct 21 2008 11:30 AM

rated by 0 users

SWDatabase.JPG

I don't know exactly where you are talking about.  Here's a shot of my Database Manager.  I only have the database on the primary server, not on my polling engines.

Kent

jeff.stewart replied on Tue, Oct 21 2008 11:32 AM

rated by 0 users

 I was seeing the same issues as you guys.  I had to remove some elements and now have around 10,000 per server. This seems to be keeping the syncing going.  However, every time I restart my NetPerfMon service I have to wait a very long time, sometimes upwards of 20 minutes for the DB to sync.  It basically is the amount of time for the service to grab the amount of memory it needs.  Is this a bug?

maz.kassam replied on Tue, Oct 21 2008 12:27 PM

rated by 0 users

Hi Jeff,

 I notice my primary Orion server takes about 8 minutes after restart to sync. BUT my secondary server takes about 15-20 minutes. I have looked at the memory usage and on the primary server it only uses about 200mb RAM. on the Secondary, I have to wait for memory usage to go upwards of 420mb for the service to start correctly. The secondary server does have more nodes and pollers, but the primary server houses the web site too.

Its always been like this for our setup. Originally we had version 7.5 on Windows 2000. Even after all the updates of Orion, Service Packs and Operating System update the system behaves the same. The only luck I have had lately is to reduce the number of interfaces to monitor. The system seems a lot more stable recently. If we add more nodes, then I may have to buy another poller - or cost another product!!