Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Assign Polling Engine to 1 NIC?

We have multiple NIC cards in the server, some to local machines to do backups, but only ONE that connects to the LAN. How do I assign the polling engine to just that one NIC card, so it will stop trying to poll through the other connections?

Find more posts tagged with

Accepted answers

All comments

bleearg13

We had a similar problem, but fixed it by going to the Network Connections control panel in Windows 2000 (works in W2K3, too) and clicking the "Advanced" menu option and selecting "Advanced Settings". Move the Primary NIC to the top of the list of NICs. This should ensure that any packets leaving your machine to the general network will be sourced from that NIC.

-evt

ccsdnoc

I too did this to my Windows Server 2003 and it worked perfectly. I have 1 NIC used for data collection and monitoring (the higher prioritized NIC in the advanced settings) and the other NIC is for access to the server (HTTP/TERMSERV).

Donald D. Quindardo
CCSD Network Operations Center
webmaster@craw.org

Cygnuz

Hello to all,
I can't understand, isn't it a routing problem? or are you talking about multiple interface on the same network?
I have problems connecting with "Orion sistem manager" from the primary polling engine to the secondary, can you give me any hint?
Thanks

ajsburns

Hi everyone,

We are having a related problem. We have dual Ethernet ports on the W2K3 server running SolarWinds & the SQL server, one for access to the WAN,etc, and one purely for backup.

We have recently discovered packets on the WAN with source addresses of the backup interface. This shouldn't have happened as far as we were concerned and a bit more investigation seems to have tied it down to the rediscovery polling requests that are sent every hour. In particular, src & dst ports are UDP/137 - how/why are these packets being sourced with an address that does not have access to the WAN?

We have now changed the order of the interfaces in the Advanced Network Settings as suggested in a previous post to see if this helps. There is still the problem that the IP address of the backup interface appears in the database table 'Engines, field 'IP'.

What we would like to know is what is this field used for (i.e. is it actually display only)? And is it 'safe' to change it to the correct address, and if so do we just edit the field directly, or do we need to re-run the configuration wizard or ....?

Thanks in advance.

ajsburns

quote:
What we would like to know is what is this field used for (i.e. is it actually display only)? And is it 'safe' to change it to the correct address, and if so do we just edit the field directly, or do we need to re-run the configuration wizard or ....?

We have found that if we do reset the field in the database, it gets overwritten with the old address when the NetPerfMon service restarts.

So the first question now is where does the service pick up this address from? I can't see anything in the registry, and we have changed the order of the cards to be 'correct'?

We still haven't re-run the configuration wizard - would there be any harm in doing so on a live server or should we try an upgrade first?

Thanks for any suggestions...

bleearg13

Interesting...I've noticed the same problem, but I thought that it was a Windows RPC issue. We have the order correct *and* have the right IP in the Engines table in the DB. If you find an answer, please let me know. The invalid packets are really annoying, as they throw up some errors in the firewall due them being seen as "spoofed" packets.

Network_Guru

The config wizard is safe to run at any time as long as you point it to the existing DB & remember the old passwords for the DB & Web engine.
You also only need to run the part you want to fix, IE. - to just recreate the web site & permissions - only run the Web install portion.

-=Cheers=-
NG

rmverdi

"I too did this to my Windows Server 2003 and it worked perfectly. I have 1 NIC used for data collection and monitoring (the higher prioritized NIC in the advanced settings) and the other NIC is for access to the server (HTTP/TERMSERV)."

We have the same setup , yet the polling engine is still switching polling interfaces at least once a day. When this happens we loose all nodes (everything appears down) because of the way we route traffic from our core to our remote head end routers. It's killing our statistics. We did not have this problem with WIN2K. Called SolarWinds they say it's the way WIN2K3 handles the NICs that is the problem. I believe this, but is there a work around? Can I "unbind" SNMP and ICMP from one interface? I have also applied a metric of 1 on the preferred interface and a metric of 20 on the other to try to help. Any further suggestions would be appreciated. Unfortunately, 1 NIC for everything is not a solution for us.

Thanks, Ron V.

FormerMember

We ran into this problem a while back.
Our Backup network is a Class C subnet
The other (Primary) NIC is intended for all other network traffic.

To resolve this we ended up not setting a default gateway on the Backup Network NIC.

We also do not set DNS Server entries and uncheck the Dynamic Register with DNS on that NIC to resolve problems with Windows DDNS and overlapping DNS entries.

The end result is that all network traffic, aside from the Backup traffic goes across the Primary NIC.