cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 8

Newbie Here - Where to begin...

Jump to solution

The organization I work for is the eighth largest school district in the country.  We have approximately 25,000 employees servicing approximately 280,000 students in grades 1 - 12.  Our organization is spread over close to 400 buildings, all networked, each having hundreds of nodes (switches, computers, hand-held devices...etc.).

We just purchased the following modules:

Orion Additional Polling Engine -2017.03.14

Orion Netflow Traffic Analyzer - SLX-4.2.3

Orion Network Configuration Manager v7 - DLX-7.7.0

Orion Network Performance Monitor SLX-12.2.0

Server & Application Monitor - ALX-6.4.0

Database Performance Analyzer - Category1 - Oracle EE  or ASE or DB2 (5 to 9 instances)-11.0.1.1018

Engineer's Toolset for Desktop-11.0.6

Engineer's Toolset for Web-11.0.6

Web Performance Monitor  - WPM50-2.2.1

I was not involved with the selection of these modules.  I am tasked with creating the servers in our VMware environment.  My background is in programming, VMware, DBs and MS System Management.  I read the requirements doc for NPM.  At this point, I know I need a main Orion server and a DB Server (SQL Server).

Questions:

1.  Do each of the modules (NPM, SAM, WPM) need their own application server and corresponding DB server or can all the modules reside on one main server and one DB server (w/ multiple DBs)?  What would be the recommendation for an organization of our size?

2.  As far as the Polling servers, is it a separate server or is it a server that has the Agent installed on it?  Could we use our Domain Controllers as Polling servers?

3.  What documentation would you recommend for me to start out with?  I am looking on the Solar Winds Customer Portal at the NPM documentation.  I also quickly looked at the System Application Monitor documentation in the Portal.

4.  Is there an ERD/Diagram of the entire Orion Platform?

I hope this will get me started.  I appreciate any help that anyone can give me.

Scott

0 Kudos
1 Solution

You're in the right area, sabtampa​.  Either you can find the information here, or someone can share it with you, do it for you, or point you to good resources & links.

1.  Do each of the modules (NPM, SAM, WPM) need their own application server and corresponding DB server or can all the modules reside on one main server and one DB server (w/ multiple DBs)?  What would be the recommendation for an organization of our size?

A:  Lots of Solarwinds can share application space on the main instance of NPM.  For example, NCM, NTA, VNQM, IPAM, UDT can all be installed on the main NMP server.  Some, like NTA, have their own separate database, while others, like NPM and NCM, share a different database.  You can click on each link from the Thwack Community Product forum and read their requirements from up above this page:

pastedImage_0.png

Regarding Engineer's Toolset For Web, I think you can install that for no charge on your main NPM server, and also on your APE's (Additional Polling Engines).

The Engineer's Toolset For Desktop should be installed on your PC or Laptop.  I purchase an additional license of this for each of my Network Analysts so they have a local copy on their laptop to use for troubleshooting wherever they go.

The NPM and NTA and NCM licensing you purchased will determine how much monitoring & backing up you can perform.  I have ~800 switches, routers, and firewalls, so you'd think I could get by with an NPM license for 1000 elements.  But that's not how it works.

An "element" is something you poll or want information about.  My chassis switches have 384 physical ports on each of them, and each of those physical ports has two additional virtual/logical items ("controlled" and "uncontrolled" security interface items).  That means one chassis switch's physical ports total up to 384 "elements" to poll.  If I wanted to also poll each of the two additional virtual/logical entries (controlled and uncontrolled) for each physical port, I'd add 768 elements to the count, for a total (so far) of 1152 elements.

Then there are other items I want to monitor on the switch:

  • CPU
  • Fan
  • Approximately 30 temperature sensors
  • Virtual interfaces (like SVI's--routing interfaces on VLAN's, for example)
  • Memory
  • Power

Each of these will take up multiple "elements".

Then there are volumes in servers (or in switches or PC's) to monitor--each of them is an "element".  You can see the element count increases easily.

I ended up purchases four SLX / "Unlimited Element" licenses to cover my polling needs.

There IS a physical limit of elements a single Solarwinds poller can handle:  12,000.  Which means I could poll up to 48,000 elements without running into problems.

Believe it or not, that still wasn't enough--if I wanted to poll everything.  So, I swapped out my licenses to the NAM package, which allows me to install and license up to 25 pollers, but is limited to 100,000 elements.  That's a good size for me, and I end up saving support contract costs using NAM instead of the individual licenses.  It's only $1,000/year, but I get a lot more elements, diverse pollers, and some other packages I didn't have (including VNQM, IPAM, HA, and UDT) for a one-time upcharge.

Figure out what you want to poll, and what you don't want to poll.  For example, I don't poll any interface that is "controlled" or "uncontrolled", and I save a LOT of element count!

I poll ALL physical ports--even if they're not active today. They'll become active in the future, and I want to do right by my users by knowing about problems their port may have once they start using it.

I don't manage all L2 VLAN's, but I do manage all CPU, Memory, Power, Fans, and Temperature, as well as all virtual ports (port-channels, VSS interfaces, VPC's, etc.)

When you figure this out, you may discover you have the right amount of Additional Polling Engines (remember, 12,000 elements per APE), or you may need more APE's.

2.  As far as the Polling servers, is it a separate server or is it a server that has the Agent installed on it?  Could we use our Domain Controllers as Polling servers?

Polling servers are separate, and they tie into the main NPM poller, to share information and load.  Install them after NPM is going, and install the right number of them to do all the polling you want/need.  APE's do the NPM polling, the NCM tasks, the NTA work, for the nodes they monitor, and they lighten the load of the main NPM server.  They report their information to the Solarwinds Database, which the main NPM server displays on the NPM front page.

Given the demand on polling servers (Each of mine uses 8 CPU's running 2800 MHz and up to 32 GB of RAM), I would not install an APE on a Domain Controller.  You don't want ANYTHING to possibly slow down or freeze your Domain Controllers.

By the way, my environment is all VM.  I can add CPU's or Memory at will, but I had to buy dedicated VM blades to give me that flexibility.  At least I know I'm not impacting any other applications or virtual servers when I have changes or need more resources.

3.  What documentation would you recommend for me to start out with?  I am looking on the Solar Winds Customer Portal at the NPM documentation.  I also quickly looked at the System Application Monitor documentation in the Portal.

I recommend starting with the NPM documentation.  If your NPM is already installed, you can skim over the requirements and just verify your installation has what's recommended based on the elements you'll be polling.  Once you have NPM going, you've got a LOT of great information at your fingertips, and you and your team can know when PC's, Servers, Switches, Routers, Firewalls, AP's, UPS's, etc. are down, or when they're in trouble.

After getting NPM going, I recommend setting up NCM.  It does so much to cover your back when there are questions, compliance issues, restorations, and daily configuration changes to monitor.  It's pulled my bacon out of the fire several times, and it lets me see every change made on the network.

I next installed / configured NTA so I knew who was using all the bandwidth between sites and on the Internet.  It's saved my organization a lot of money by ensuring we only purchase the required bandwidth for WAN and Internet, instead of guessing, or letting any particular group of users go crazy and eat up all the bandwidth.

After that, I'd suggest installing DPA and SAM together.  The one won't work without the other.  Get your System Admins and DBA's working with you on this, since you'll need WMI and database credentials to monitor these systems.

Last, install WPM.  After you have expertise on the other modules, WPM is the icing on top of the cake, and gets you in touch with your users' experiences with your servers' applications.

4.  Is there an ERD/Diagram of the entire Orion Platform?

I recall one in the Admin Guide.  Start here:  NPM_12_0_Administrator_Guide.pdf

As I reread your question and environment, there are a few other items you may (one-day) consider.  One is the Additional Web Server.  When you have more things reporting to NPM than it can easily handle, and Additional Web Server can share the load.  The other thing is Enterprise Operations Console.  The EOC is used by very large organizations who have multiple Additional Web Servers, or several dedicated NPM instances (as opposed to simple Additional Polling Engines).  You may discover EOC could be a way for your top tier of Network Analysts and Network Architects to get a good high-level view of all your Solarwinds environment.  Cross that bridge when you come to it, though.

I recommend you talk with your Solarwinds SE and describe your environment, then see if they think you have the appropriate licenses and modules.  Use common sense, since these are salespeople, who will be paid to get you to buy more products and licenses.  They're not evil, but they ARE human, with financial needs served by financial incentives for achieving sales goals--by selling you products.  Take their advice and bounce it off a few of your respected peers and resources to see if it makes good sense, or if it's sales for the sake of dollars.

If you have a Network Architect, even an external one paid for with hours of support contract fees pre-purchased, leverage those hours and get a solid set of advice going so you have the right direction set in front of you.  Going down the wrong path now will result in . . . let's just call them "career-limiting decisions."   ;^)

Good luck!  Don't hesitate to ask very specific questions in this forum.  At the same time, it's not wasted time opening a support ticket with Solarwinds online to ensure you get the help you pay for.  If an issue is urgent, do NOT use the online support solution except to start the ticket.  Use the telephone.  Build the case online, submit it, but call the Support phone number after you do, and reference the incident ID number.  You may spend ten or twenty minutes on hold getting to the first person, but if you have that ticket built, it'll speed you through the system and get an expert working on your issue faster than just letting a web ticket sit in cyberspace for a week or two.

Swift packets!

Rick Schroeder

View solution in original post

8 Replies

Sending you a PM. 

0 Kudos
Level 13

Hi sabtampa, welcome to THWACK and to SolarWinds! I work in the User Experience (UX) team, and our job is to help make products easier to use, more useful and more efficient. One of the ways we do this is chatting with new users over WebEx, hearing what seems straightforward--and what doesn't (especially that!). If and when you have time, I'd love to have a call with you to learn about your newb experiences. And, in exchange, you'd get 3000 thwack points which can be used for all kinds of swag. Radioteacher​ and other users on thwack can vouch that UX feedback sessions are fun and painless, and that we do listen!  Feel free to reach out privately via thwack message or via email (kellie.mecham@solarwinds.com).

Eager to hear how things are going!  --Meech

Thanks Kellie!  I want to get the first few servers built and then the installation of NPM first.  After that, I am sure I will need more help.

0 Kudos

You're in the right area, sabtampa​.  Either you can find the information here, or someone can share it with you, do it for you, or point you to good resources & links.

1.  Do each of the modules (NPM, SAM, WPM) need their own application server and corresponding DB server or can all the modules reside on one main server and one DB server (w/ multiple DBs)?  What would be the recommendation for an organization of our size?

A:  Lots of Solarwinds can share application space on the main instance of NPM.  For example, NCM, NTA, VNQM, IPAM, UDT can all be installed on the main NMP server.  Some, like NTA, have their own separate database, while others, like NPM and NCM, share a different database.  You can click on each link from the Thwack Community Product forum and read their requirements from up above this page:

pastedImage_0.png

Regarding Engineer's Toolset For Web, I think you can install that for no charge on your main NPM server, and also on your APE's (Additional Polling Engines).

The Engineer's Toolset For Desktop should be installed on your PC or Laptop.  I purchase an additional license of this for each of my Network Analysts so they have a local copy on their laptop to use for troubleshooting wherever they go.

The NPM and NTA and NCM licensing you purchased will determine how much monitoring & backing up you can perform.  I have ~800 switches, routers, and firewalls, so you'd think I could get by with an NPM license for 1000 elements.  But that's not how it works.

An "element" is something you poll or want information about.  My chassis switches have 384 physical ports on each of them, and each of those physical ports has two additional virtual/logical items ("controlled" and "uncontrolled" security interface items).  That means one chassis switch's physical ports total up to 384 "elements" to poll.  If I wanted to also poll each of the two additional virtual/logical entries (controlled and uncontrolled) for each physical port, I'd add 768 elements to the count, for a total (so far) of 1152 elements.

Then there are other items I want to monitor on the switch:

  • CPU
  • Fan
  • Approximately 30 temperature sensors
  • Virtual interfaces (like SVI's--routing interfaces on VLAN's, for example)
  • Memory
  • Power

Each of these will take up multiple "elements".

Then there are volumes in servers (or in switches or PC's) to monitor--each of them is an "element".  You can see the element count increases easily.

I ended up purchases four SLX / "Unlimited Element" licenses to cover my polling needs.

There IS a physical limit of elements a single Solarwinds poller can handle:  12,000.  Which means I could poll up to 48,000 elements without running into problems.

Believe it or not, that still wasn't enough--if I wanted to poll everything.  So, I swapped out my licenses to the NAM package, which allows me to install and license up to 25 pollers, but is limited to 100,000 elements.  That's a good size for me, and I end up saving support contract costs using NAM instead of the individual licenses.  It's only $1,000/year, but I get a lot more elements, diverse pollers, and some other packages I didn't have (including VNQM, IPAM, HA, and UDT) for a one-time upcharge.

Figure out what you want to poll, and what you don't want to poll.  For example, I don't poll any interface that is "controlled" or "uncontrolled", and I save a LOT of element count!

I poll ALL physical ports--even if they're not active today. They'll become active in the future, and I want to do right by my users by knowing about problems their port may have once they start using it.

I don't manage all L2 VLAN's, but I do manage all CPU, Memory, Power, Fans, and Temperature, as well as all virtual ports (port-channels, VSS interfaces, VPC's, etc.)

When you figure this out, you may discover you have the right amount of Additional Polling Engines (remember, 12,000 elements per APE), or you may need more APE's.

2.  As far as the Polling servers, is it a separate server or is it a server that has the Agent installed on it?  Could we use our Domain Controllers as Polling servers?

Polling servers are separate, and they tie into the main NPM poller, to share information and load.  Install them after NPM is going, and install the right number of them to do all the polling you want/need.  APE's do the NPM polling, the NCM tasks, the NTA work, for the nodes they monitor, and they lighten the load of the main NPM server.  They report their information to the Solarwinds Database, which the main NPM server displays on the NPM front page.

Given the demand on polling servers (Each of mine uses 8 CPU's running 2800 MHz and up to 32 GB of RAM), I would not install an APE on a Domain Controller.  You don't want ANYTHING to possibly slow down or freeze your Domain Controllers.

By the way, my environment is all VM.  I can add CPU's or Memory at will, but I had to buy dedicated VM blades to give me that flexibility.  At least I know I'm not impacting any other applications or virtual servers when I have changes or need more resources.

3.  What documentation would you recommend for me to start out with?  I am looking on the Solar Winds Customer Portal at the NPM documentation.  I also quickly looked at the System Application Monitor documentation in the Portal.

I recommend starting with the NPM documentation.  If your NPM is already installed, you can skim over the requirements and just verify your installation has what's recommended based on the elements you'll be polling.  Once you have NPM going, you've got a LOT of great information at your fingertips, and you and your team can know when PC's, Servers, Switches, Routers, Firewalls, AP's, UPS's, etc. are down, or when they're in trouble.

After getting NPM going, I recommend setting up NCM.  It does so much to cover your back when there are questions, compliance issues, restorations, and daily configuration changes to monitor.  It's pulled my bacon out of the fire several times, and it lets me see every change made on the network.

I next installed / configured NTA so I knew who was using all the bandwidth between sites and on the Internet.  It's saved my organization a lot of money by ensuring we only purchase the required bandwidth for WAN and Internet, instead of guessing, or letting any particular group of users go crazy and eat up all the bandwidth.

After that, I'd suggest installing DPA and SAM together.  The one won't work without the other.  Get your System Admins and DBA's working with you on this, since you'll need WMI and database credentials to monitor these systems.

Last, install WPM.  After you have expertise on the other modules, WPM is the icing on top of the cake, and gets you in touch with your users' experiences with your servers' applications.

4.  Is there an ERD/Diagram of the entire Orion Platform?

I recall one in the Admin Guide.  Start here:  NPM_12_0_Administrator_Guide.pdf

As I reread your question and environment, there are a few other items you may (one-day) consider.  One is the Additional Web Server.  When you have more things reporting to NPM than it can easily handle, and Additional Web Server can share the load.  The other thing is Enterprise Operations Console.  The EOC is used by very large organizations who have multiple Additional Web Servers, or several dedicated NPM instances (as opposed to simple Additional Polling Engines).  You may discover EOC could be a way for your top tier of Network Analysts and Network Architects to get a good high-level view of all your Solarwinds environment.  Cross that bridge when you come to it, though.

I recommend you talk with your Solarwinds SE and describe your environment, then see if they think you have the appropriate licenses and modules.  Use common sense, since these are salespeople, who will be paid to get you to buy more products and licenses.  They're not evil, but they ARE human, with financial needs served by financial incentives for achieving sales goals--by selling you products.  Take their advice and bounce it off a few of your respected peers and resources to see if it makes good sense, or if it's sales for the sake of dollars.

If you have a Network Architect, even an external one paid for with hours of support contract fees pre-purchased, leverage those hours and get a solid set of advice going so you have the right direction set in front of you.  Going down the wrong path now will result in . . . let's just call them "career-limiting decisions."   ;^)

Good luck!  Don't hesitate to ask very specific questions in this forum.  At the same time, it's not wasted time opening a support ticket with Solarwinds online to ensure you get the help you pay for.  If an issue is urgent, do NOT use the online support solution except to start the ticket.  Use the telephone.  Build the case online, submit it, but call the Support phone number after you do, and reference the incident ID number.  You may spend ten or twenty minutes on hold getting to the first person, but if you have that ticket built, it'll speed you through the system and get an expert working on your issue faster than just letting a web ticket sit in cyberspace for a week or two.

Swift packets!

Rick Schroeder

View solution in original post

Thank you Rick!  You gave me a lot to digest, but very helpful!!!  I really appreciate it.

Thank you

You are the man!

0 Kudos

sabtampa

Welcome to THWACK! I'm one of the Community Moderators here on THWACK. If you ever have any questions, please do not hesitate to post on here or feel free to DM me

I am happy to help you out here! I know CourtesyIT would definitely be able to help you out with this question. He is one of our MVP's and knows a ton about our products!

0 Kudos

Thanks Brett.

0 Kudos