Can we have HA VIP used as source IP for Polling all nodes

abdulraheemsidz

As if now we know that in case of HA solution we have 3 IP addresses (HA VIP, Primary server IP, Secondary Server IP).

- HA VIP is mainly used for incoming requests like (SNMP Traps, Net flow data etc.)

- In case of polling the devices either of the 3 IPs is used based as source IPs.

Which IP address is used as the source when using a VIP?

It would be very ease and helpful if we fix HA VIP also as Source Polling IP address.

Find more posts tagged with

ha vip

high availability polling ip

ha polling ip

high availability vip

Status: None

Comments

mad-stevo

Between all three IP addresses (HA VIP, Primary server IP, Secondary Server IP) Windows will automatically choose the closest IP address to the Network ID as the source address. So in order to make sure all outbound communication is sourced from the VIP, just make sure you set your VIP to be the closest IP address to the Network ID, out of the three SolarWinds addresses. See Which IP address is used as the source? - SolarWinds Worldwide, LLC. Help and Support

dsimpkins

Although that is fine and if you are able to architect your poller IP's in this fashion, asking your network engineer/firewall engineer to allow 3 IP's for each poller can be a difficult ask especially as your number of pollers increase.

Having the software locking in the source polling IP would be much easier for device configs as you scale out.

johan

I tried following this methodology, still does not work in a single subnet deployment model??

GW = x.x.x.254 11111110

VIP = x.x.x.60 00111100 <-- Used to be my primary poller IPv4 address

Primary Poller = x.x.x.1 00000001

Secondary Poller = x.x.x.11 00001011

Still gives issues. Best of all is, Solarwinds' response - oh, here are some tech guides to follow, best of luck to you mentality. Support has REALLY been disappointing of late vs what it used to be.

oiram

While this might be obvious that the closest IP address will be chosen (that concept fooled me as well) it is not always the case. There is good reading on how Windows behave in case of multiple IP(s) on single NIC (when it comes to multi-NIC server you can play with metrics) and how to work around it:

https://www.virtualizationhowto.com/2016/02/force-primary-outgoing-ip-address-windows-server-2012-r2/

It is for WS2012R2 but I have it tested also on WS2016 and solution works just fine.

For your case johan I would suggest creating following PowerShell script and execute it after switchover/failover:

Main:

Set-NetIPAddress -IPAddress 196.34.139.60 -SkipAsSource $False

Set-NetIPAddress -IPAddress 196.34.139.1 -SkipAsSource $True

Backup:

Set-NetIPAddress -IPAddress 196.34.139.60 -SkipAsSource $False

Set-NetIPAddress -IPAddress 196.34.139.11 -SkipAsSource $True

johan

Hello Oiram,

Thank you for the above, as mentioned in a previous conversation, if I buy a service from a Vendor, I would expect this functionality to exist within the capabilities of the application to make this intelligent decision. If you buy a car, you don't need to have a steering wheel fitted after market to get it on the road. Similar with HA - this functionality should reside in the app itself.

I think it is worth your Developers having a look at this request??

dsimpkins

I was fortunate that my pollers are in a /23 subnet so was able to put the VIP address in the first half and the server IP's in the second, i did however feel the pain of trying to work out the binary and metrics when we were designing our platform to make certain it would work the way expected.

The way the actual failing over of the IP address is very rude and crude and we have had to supress duplicate IP alerts in SCOM because of it and it does leave a mess behind when you delete the pool especially if you want to recreate it again.

The impression I've got from the product team and trainers I've spoken to is that HA is still in the early stages and development time is ramping up with the level of interest so the more we share our concerns/feedback the better it will get.

Fingers crossed for a much improved version 2

oiram

I read you dsimpkins you are lucky you did not have to battle with FoE, but I'm sure your colleagues told you some stories Some of the behaviours, like leaving IP on the NIC after destroying the pool, were intentional. Imagine accidentally removing the pool... Also, the VIP shall be removed as one of the first tasks of failover process, resulting of the VIP not visible on the network at all for short period of time, simply to prevent duplicate.

What I'm thinking would be a good option for the product is to have the ability to configure the pool to use VIP as a source of outgoing traffic, if VIP is configured on the pool (pools in HA v2 may have only VHN).

aLTeReGo

For those of you depserate for a solution which will force Orion to utilize the VIP as the source for all polling without changing the IP addresses of your HA pool members, I have posted a workarond in the thread below.

Force Orion High Availability to Utilize VIP as Source IP Address

oiram

Much appreciated aLTeReGo

aabdul

One customer i worked with this week got this problem fixed by changing the default gateway to a different IP in such a way so that VIP has highest number of matching significant bits as suggested in Which IP address is used as the source when using a VIP?

In their case, the VIP and Primary poller IP's had same number of matching bits i.e 25 which may have resulted windows to pickup the primary poller IP

The issue here is more to do with how Windows picks up the primary IP then it has to do with Solarwinds Application as already suggested by @mad-stevo