23 Replies Latest reply on Aug 8, 2018 10:25 AM by serena

    VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller

    jonathanswift

      I welcome the integration of VMAN with Solarwinds, but now it seems the CentOS appliance is being discontinued and the functionality (or some of it) is being incorporated in to the main poller.   We paid a fortune for the appliance to add functionality to Solarwinds and provide extra graphs and reporting, and maintenance is 60% higher than our SAM ALX license.  Is this really good value for money?  We wanted to keep the load off our main poller but Solarwinds are saying to install VMAN right there on the main poller, and transfer the database in to the Solarwinds database.  My problem is;

       

      - the main poller is heavily used (it's a separate physical machine)

      - the additional poller is heavily used (VMs now but will become a physical soon)

      - the database server is a separate physical, HP Blade with integral SSD.

       

      This keeps the load off our production ESX systems, it keeps the database I/O (which is intense) off the SAN, and doesn't interfere with production systems.  It doesn't slow them down, and if they go down it does not, it's independent.

       

      Now we're being told we need to install VMAN 8 (isn't that a quite big product) on to our main poller, loading it up even further, and making the Solarwinds database very big.  I don't know how big the data is in the VMAN system - and technical support didn't answer my question either.   We'll need to review whether we keep VMAN since I'm not sure it is worth the hasstle of trying to push the main poller to its limits, or causing I/O issues with the SQL database. 

       

      Who has been brave and done it?  I can't find any VMAN videos newer than 2 years ago.  Normally as soon as a new release is there I'm deploying it, but this is such a radical change that I'm putting the intended update on hold, and we were just about to renew our maintenance - that will have to go on hold too.   What happens if we just keep VMAN as it is?   At some point I'm guessing we won't be able to update Solarwinds because it will complain VMAN is too out-of-date.

        • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
          jonathanswift

          I've decided I'm going to have a go.  From what I have learned much of the VIM data is already in the Solarwinds database - it has been fed in over time.   The new polling is streamlined so should be more efficient.  I'm already running on a physical machine with 24 cores and 72gb ram, so performance is fairly good.

            • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
              p.darila

              Hi,

              only some infos:

              - VMAN Appliance as a separate poller continues - so you don't have to migrate, and you can use CentOS VMAN as before

              - VMAN appliance has own database. VMAN appliance is collecting data from hypervisors, store it in own database and then send results from latest polls to VIM/Orion. There is another processing and data are stored in VIM/Orion database -> Orion database will be approximately same also with VMAN Orion poller

              - with VMAN Orion poller - all colletions/polls will be executed directly from Orion machine, but it shouldn't rapidly increase poller utilization (I guess it will be approximately Orion machine until now + VMAN appliance utilization)

              - if utilization on main poller is too high, you can deploy new Free Poller (something like additional poller) - it is working on new VMAN Orion poller -> so you will spread virtualization polling utilization to more machines

              1 of 1 people found this helpful
                • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                  jonathanswift

                  Found a couple more things.  You have to put polling in to basic mode if you have an agent on Hyper-V host.  It then warns that some metrics won't be available.

                   

                  Also, the appliance OS doesn't get upgraded when you upgrade the appliance.  You have to build a new VMAN appliance then migrate all the data.  I have something like 400-600GB of data in the appliance.

                   

                  There see to be no real details of this FREE (nothing is for free) poller.   I had wanted to build a new machine, based on Windows 2016, and put the VMAN polling on there, then perhaps one at another UK datacenter, and one in the US datacenter.  At the moment I have the VMAN appliance in London, a collector in another UK site, and one collector in the US.  That's sort of what I want to replicate.  

                   

                  More and more of VMAN has been incorporated in to the main poller anyway - like recommendations.  I can envisage three Windows 2016 servers, 3 small VMs to act as collectors, which will put the data in the Orion database - still don't know how I'd get the historical data in to the Orion database - somebody said there was a configuration wizard which might do that, but in discussions with Solarwinds they suggested leaving the data on the appliance, keeping it, and swapped primary/secondary licenses, then using the old VMAN appliance for some reporting.

                   

                  Not getting very far - at home I can install VMAN 8 (for Hyper-V) and the machine can be configured later because it gets a DHCP address, but at work the VLAN doesn't have DHCP so I try to use the script to configure the IP and I just get errors everywhere and it doesn't set anything.   The VMware VMAN 8 appliance might work more easily because the IP is set as part of the OVF installation.

                    • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                      p.darila

                      - unfortunately, agent polling is not supported by new VMAN Orion poller. Just be careful, basic polling is really basic.

                      - upgrading VMAN 7.1 to 8.0 is possible with hotfix 2, but you will stay on old CentOS. Full upgrade between VMAN Appliance (to get new centOS) is not possible. But its possible to just connect disk from old VMAN to new one.

                      - FREE poller - it means, that additional poller license is not needed, so its free. You just need to have virtualization license in main orion poller. It should be replacement for VMAN federated collector. (data-sources polled by FP are counted to main license)

                      - I don't think there is a way, how to migrate old data from VMAN appliance to Orion db

                • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                  goodzhere

                  I am having issues with this as well.  I keep getting license issues.  The licensed socket amount keeps bouncing around on the Orion side.  It also would like to know more about this free poller.  I would like to turn the old appliance into this.  And I want to keep the data from the old appliance.  I want to still use the old appliance for some nodes and move others the the new integrated setup.  I also ran into the static IP issue with the brand new 8.0 Hyper-V appliance.  I had to configure it by cmd line in the appliance.  Needless to say, this has not gone very smooth.  I have not been happy with this at all.

                  • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                    serena

                    Hey Jonathan, I just wanted to follow up with you, after you moved off of the appliance do you have any feedback you'd like to share about the experience? Anything that you wish had been better? I'd love to jump on a call with you and get more details.

                      • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                        jonathanswift

                        Hi Serena,

                         

                        In the end I consider the upgrade to be so immature I dare not try it - so I bottled it.   As it happens all future upgrades since have been worthless since I can't get to a position where I am comfortable with what Solarwinds have done.   It is very disappointing as you bring in new features we can't even access because of this botched upgrade.

                         

                        Lots of functionality is missing from the very expensive Virtualization Manager we purchased, so the replacement isn't in a mature enough state to use.   Plus, I have tested quite a bit and I don't want to put excessive load on our main poller - I want an option to host the Virtualization Manager java code on another box - I don't believe you offer that option.  Lastly, the included (not FREE) collectors you had for Virtualization Manager aren't replicated in the new offering - I believe the collectors are part of additional / main pollers, not separate as they are with the existing appliance.  I have a collector in the US, but no additional poller, so it isn't a good equivilent offering, therefore I don't feel able to take any upgrades.

                         

                        In addition, we currently monitors hosts as SNMP - how else would anyone monitor them ?  How else can you get local disk information?  The new offering is highly restrictive on the type of monitoring that can be done.   I raised this at the time but nobody has come back except to confirm that Virtualization Manager -- and this is extraordinarily bizarre -- doesn't actually collect metrics for local drives on the host.  I mean, how exactly are you supposed to watch out for a drive filling up on an ESX host?  Solarwinds doesn't seem to even support this monitoring.

                         

                        Normally I take upgrades to new software within weeks and always stay current, but since the Virtualization Manager. ... what I will call fiasco ... I've been forced to stay on old software.   I don't believe you can load all that processing on to the main poller without serious performance degradation.

                         

                        In addition we purchased an additional web poller for performance but have never been able to use - so a waste of money which I pushed for.  It is incompatible with our current setup -- for some reason -- but we can't upgrade anything now because of the problems with the replacement for Virtualization Manager.   At the moment we will have to carefully consider whether we freeze everything at this level and discontinue maintenance for Solarwinds products as they don't come with the level of support I expect.   We did it with IPAM - we stripped it out and I put it on another machine and I transferred the part of the database I needed, and it just runs fine, doing what we want, with no maintenance on it.  There are no new features we could possibly conceive that we would want.

                         

                        At present I am quite unhappy with the software.  I know that people work hard to add new features but these are now a slap in the face since we can't access them - despite having full maintenance.  In most respects the software is going in the right direction, but for with Virtualization Manager.  We need this to be on a separate physical machine or VM.  Two of our pollers are physical machines - they could not cope as VMs.  The SQL server is a physical blade with an SSD-style plug-in card - it could not cope as a VM.  I'm trying to keep this usable - people will not use it if it is slow - hence why I tried to speed it up by buying the separate web interface.

                         

                        I hope this is enough detail for you to see the problems we are having. 

                        1 of 1 people found this helpful
                          • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                            serena

                            Jonathan, thank you for your response! I'd like to talk through some of your bullet point here:

                             

                            "Lots of functionality is missing from the very expensive Virtualization Manager we purchased, so the replacement isn't in a mature enough state to use.   Plus, I have tested quite a bit and I don't want to put excessive load on our main poller - I want an option to host the Virtualization Manager java code on another box - I don't believe you offer that option.  Lastly, the included (not FREE) collectors you had for Virtualization Manager aren't replicated in the new offering - I believe the collectors are part of additional / main pollers, not separate as they are with the existing appliance.  I have a collector in the US, but no additional poller, so it isn't a good equivilent offering, therefore I don't feel able to take any upgrades."

                             

                            VMAN allows you to use free additional polling engines as a replacement for federated collectors. Grab any Orion installer, and select "Scalability Engine" and it will allow you to specify the free VMAN polling engine. This will allow you to scale and take the load off of your main poller.

                            This is called out in in our documentation here: Features Available in VMAN 8.2.1 (Orion)  &  Install an Additional Polling Engine, Additional Web Server, or HA server - SolarWinds Worldwide, LLC. Help and Support

                            "In addition, we currently monitors hosts as SNMP - how else would anyone monitor them ?  How else can you get local disk information?  The new offering is highly restrictive on the type of monitoring that can be done.   I raised this at the time but nobody has come back except to confirm that Virtualization Manager -- and this is extraordinarily bizarre -- doesn't actually collect metrics for local drives on the host.  I mean, how exactly are you supposed to watch out for a drive filling up on an ESX host?  Solarwinds doesn't seem to even support this monitoring." You can absolutely monitor the host via SNMP, the restriction is that VMware has locked down SNMP monitoring and you would have to manually enable it. Here's our note on the subject Enable SNMP on VMware ESXi and ESX servers - SolarWinds Worldwide, LLC. Help and Support and here's VMware's documentation on how to enable it Configure SNMP for ESXi 

                             

                            Once you've enabled it, simply go through the add a node wizard, select SNMP and you're off to the races and monitoring the host.

                             

                            If you'd like to talk more about this in depth, I'd love to get on a call with you!

                             

                             

                            2 of 2 people found this helpful
                        • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                          laurencetanclearly

                          Hi, VMAN appliance officially discontinued? I do not see any official announcement on Solarwinds KB. Can anyone direct me an official Solarwinds site that VMAN appliance is officially discontinued?

                            • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                              serena

                              laurencetanclearly  wrote:

                               

                              Hi, VMAN appliance officially discontinued? I do not see any official announcement on Solarwinds KB. Can anyone direct me an official Solarwinds site that VMAN appliance is officially discontinued?

                              The appliance is not officially discontinued. You can still use it and integrate it with your Orion installation. However, I absolutely recommend that you migrate, because all of the new features that we have been building are on the Orion side. When/if I deprecate the appliance, you will have plenty of communication from SolarWinds to let you know, and it will be reflected Currently supported software versions - SolarWinds Worldwide, LLC. Help and Support  and will follow our end of life policy End of Life Policy - SolarWinds Worldwide, LLC. Help and Support  and will also be reflected on the relevant release notes as to when the start of such a process would occur.

                                • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                                  jonathanswift

                                  Yes, because VMAN Orion didn't work at all for me.

                                  Case # 00115308, ongoing.   Put back whilst I dealt with a whole load of instability problems.

                                   

                                  Case # 00116803 - Application Dependency monitoring (the connections inbetween nodes/applications) doesn't work at all for me.

                                  Got it working on my test system but it's not deploying NPCAP or any other software, and even deploying that manually didn't work.

                                   

                                  The instability problems seem (touches large 4x6" plank of wood) to have gone away after a lot of upgrading, re-configuring, etc.

                                  Now, as it happens, when I upgraded to HF2 the other day on the main poller it updated Core only, but on the two additional pollers it updated VMAN to a later hotfix (8.2.1 HF1), so there appeared to be some inconsistency.

                                   

                                  I definitely want to move forward and replace the VMAN appliance with the VMAN Orion component, because some features are only available in the latter.

                                  Unfortunately when we could not (with Solarwinds assistance) get VMAN Orion working I had to turn back to the appliance.

                                   

                                  Application Dependency has been totally turned off, I've even removed it from the screens, as it didn't work.  It's really helpful so I want to get that working.

                                  2 of 2 people found this helpful
                                    • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                                      serena

                                      jonathanswift  wrote:

                                       

                                      Yes, because VMAN Orion didn't work at all for me.

                                      Case # 00115308, ongoing.   Put back whilst I dealt with a whole load of instability problems.

                                       

                                      Case # 00116803 - Application Dependency monitoring (the connections inbetween nodes/applications) doesn't work at all for me.

                                      Got it working on my test system but it's not deploying NPCAP or any other software, and even deploying that manually didn't work.

                                       

                                      The instability problems seem (touches large 4x6" plank of wood) to have gone away after a lot of upgrading, re-configuring, etc.

                                      Now, as it happens, when I upgraded to HF2 the other day on the main poller it updated Core only, but on the two additional pollers it updated VMAN to a later hotfix (8.2.1 HF1), so there appeared to be some inconsistency.

                                       

                                      I definitely want to move forward and replace the VMAN appliance with the VMAN Orion component, because some features are only available in the latter.

                                      Unfortunately when we could not (with Solarwinds assistance) get VMAN Orion working I had to turn back to the appliance.

                                       

                                      Application Dependency has been totally turned off, I've even removed it from the screens, as it didn't work.  It's really helpful so I want to get that working.

                                      I've been talking to your AE in your case about your item. We'll be following your case very closely for updates.

                                        • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                                          jonathanswift

                                          In terms of Application Discovery not working, it fails because it could not install the NPCAP software.  It downloads this from the NMAP.ORG website, as per their licensing arrangement - a commercial license for this product allows the OEM to embedd the installer within their application.   I've found it is being blocked by our Checkpoint firewalls, and it's rated as a site which can be used for hacking / nefarious activities.  The product itself isn't a hacking tool, but could be used for that.   So our security team aren't happy unblocking the site so all our servers can see it.  You will run in to this problem in the real world.   We now have a problem because they will download one copy for distribution from a share or web site in our organisation, but there seems little possibility they will allow machines to download it from external web sites on-demand. 

                                           

                                          Solarwinds needs to buy a license to embedd it, or, better solution perhaps, make it available for download from the main poller.

                                          The installer appears to be here Index of /npcap/dist   and the installer, probably, takes latest-npcap-installer.exe or may take a specific (known working) version.

                                           

                                          I've tried downloading a version and installing it manually but that doesn't work - there are no instructions about which options to use when installing it, or any steps needed to integrate it, and whether to do the same with WinpCap (for latency monitoring). 

                                           

                                          Simple test to see if this is bocked is to just visit Index of /npcap/dist from a machine with the agent installed - if you can't get there, if you just don't see the listing, then you are probably being blocked.   Many, many organisations run Checkpoint firewalls (other firewalls are available!) and many like us are heavily audited and will have to get exception forms signed for access to sites like NMAP if, as Checkpoint has done, it has been labelled undesirable.   You engineers must be developing in an environment with 'free' Internet access, i.e. not going through firewalls like Checkpoint.  We have not labelled NMAP as a site which can be used by hackers, Checkpoint has, and with reasons we can understand.

                                           

                                          Hope this helps understand. I'm willing to work to try our a solution which involves hosting it internally if that helps.

                                          • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                                            jonathanswift

                                            So just to update, running the migration wizard this time went through ok, and it's been running good for a week or so.   Most hosts were absolutely fine - some polling VMware ESX hosts as ICMP, some as SMTP, but VMAN Orion is pulling ESX metrics just fine via the vCenters - BUT, I had to change the vCenters from agent polling to WMI - that way I keep vCenter polling, i.e. disk volumes, service monitoring, etc., and polling works for vCenter/ESX hosts.   Other ESX hosts which are standalone are monitoring just fine (say those embedded in Riverbed appliances).

                                             

                                            If you are not careful it will look like vCenter polling is ok but some things don't work when agent polling on the vCenter - one of those is IOPs monitoring on LUNs.  So you must use WMI.  You don't see a message in Virtualization Settings, but when you go to the vCenter you do see a warning about polling.   Once vCenters are polling via WMI they will display IOPs info.  We use that info to find out how many IOPs a machine is doing - if we're debugging performance problems.

                                             

                                            For Hyper-V (we have a small collection right now) I had to change polling to WMI  (was agent) and then it worked just fine.   Just remembering to go back in to List Resources and select Hyper-V.

                                             

                                            We've kept the appliance but I expect to turn it off shortly.   Performance hit on the primary poller appears to have been negligable - and I admit to being surprised.  So there you are, all done and now happy.

                                              • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                                                serena

                                                jonathanswift  wrote:

                                                 

                                                So just to update, running the migration wizard this time went through ok, and it's been running good for a week or so.   Most hosts were absolutely fine - some polling VMware ESX hosts as ICMP, some as SMTP, but VMAN Orion is pulling ESX metrics just fine via the vCenters - BUT, I had to change the vCenters from agent polling to WMI - that way I keep vCenter polling, i.e. disk volumes, service monitoring, etc., and polling works for vCenter/ESX hosts.   Other ESX hosts which are standalone are monitoring just fine (say those embedded in Riverbed appliances).

                                                 

                                                If you are not careful it will look like vCenter polling is ok but some things don't work when agent polling on the vCenter - one of those is IOPs monitoring on LUNs.  So you must use WMI.  You don't see a message in Virtualization Settings, but when you go to the vCenter you do see a warning about polling.   Once vCenters are polling via WMI they will display IOPs info.  We use that info to find out how many IOPs a machine is doing - if we're debugging performance problems.

                                                 

                                                For Hyper-V (we have a small collection right now) I had to change polling to WMI  (was agent) and then it worked just fine.   Just remembering to go back in to List Resources and select Hyper-V.

                                                 

                                                We've kept the appliance but I expect to turn it off shortly.   Performance hit on the primary poller appears to have been negligable - and I admit to being surprised.  So there you are, all done and now happy.

                                                This is fantastic information I'm going to bring this over to our documentation team to see if we can identify areas to surface this information in better key areas.  I'm super happy you're off of the appliance, as that will position you well for upcoming features that we add.

                                                 

                                                Also, since you and Tom below have noticed that the polling status can be misleading - I'm curious where you would expect to find information about vCenter polling status?

                                            • Re: VMAN 8 - appliance discontinued, VMAN becomes incorporated in to main poller
                                              tomaddox

                                              I feel your pain. Orion has been unable to poll our production vCenter servers for over a month, and Solarwinds support seems to be at a loss. Upgrading Orion is a little bit of a house of cards, and apparently Windows Server 2012 R2 is no longer supported, so I guess we'll be planning a migration to 2016. This is really looking like Storage Manager redux, where a really good product was bought and then destroyed by the Orion integration.

                                              1 of 1 people found this helpful