4 Replies Latest reply on Sep 15, 2017 1:51 AM by mesverrum

    How to monitor old systems?


      In our environment we go back to even windows NT. The only reason we have systems that are so old is because of the application requirements we are running old applications. Checking with business requirements we can't nor are teams willing to upgrade meaning have a need to monitor old software and hardware too. What I am finding is that as solarwinds evolves their software they are removing ability to monitor old devices and software. I'll give an example, SQL version 2005. We have sql express and version 2005 in our infrastructure that we need to monitor. We want to use SAM sql app insight but it returns a message saying "This version of SQL is not supported", But when you scroll down a majority of items are polled perfectly fine. Making it confusing when the software says not supported yet it shows it's polling several things just fine.


      1. We have old HP box's with old HP-UX version 11.11, and we've exhausted every single possible option and have been unable to find a way to get snmp installed for monitoring. 2. Windows NT can't be monitored, That's ok. But Windows 2003 and above should have full monitoring. Yet when we try templates, etc. IT doesn't work on 2003.

      3. Other features such as Agent monitoring etc, have limitations on OS supported.


      In other words Solarwinds from being a monitoring all solution, to a monitor selectively solution. The problem with that is that if it can't do what we want it to do than we question what is the value of the product to us. But then again maybe we aren't looking at the right place to make things happens. OR are unaware of how to get things monitored properly.


      To make it simple I'll list our pain points.... Azure system with Oracle OS. I understand SNMP is frowned upon for Azure, And Oracle os is not supported for agent monitoring, which basically left us scratching our heads of how to monitor these servers. Next in line is sql 2005, we can monitor the hardware, but since app insight doesn't support sql 2005 we don't have in-depth monitoring for that version. Issues with setting up snmp v3 against unix and linux systems. Unable to load balance our pollers. Half of that is on our side as the configuration on device is not complete thus moving device to a poller not in configuration will cause device to be unsupported. but other half is the software defaulting to one poller every time you run discovery. for those not aware they can change it they add all device to one single poller overwhelming it.


      I can take any suggestions and any help. We've grown tired of searching thwack and opening tickets to no resolution.



        • Re: How to monitor old systems?

          Regarding the SQL 2005, alterego addressed that question here in 2014. 

          Does anyone know why AppInsight isn't Supported on SQL 2005 SP4?

          Essentially SQL 2005 was already out of support before Solarwinds ever came out with the appinsight concept so they never wrote it intending to be used on software that was already past it's EOL.  They have to draw the line somewhere and the ROI in developing new monitoring tools for dead software is really not there and it raises headaches on the support side.  As you saw you can apply it, any of the items that show as unknown don't exist in that version.


          On the HP-UX I found these threads, you may have seen them before but hopefully it gives you a lead, it appears that at least in the past people have been able to get it working, but there are many posts asking about it so I suspect there is something fussy about it.

          HPUX SNMP

          net-snmp / Mailing Lists


          For the SAM templates on server 2003 you typically need to change the fetching from WMI to RPC to monitor, WMI was really not mature back then.


          Haven't ran into the oracle in azure situation myself, hopefully someone else can weigh in on that. 


          As far as being tired of searching, all I can say is any tool with these many knobs and buttons available to you is going to have a learning curve. Hope you guys are able to find what you need to cover your environment.


          -Marc Netterfield

              Loop1 Systems: SolarWinds Training and Professional Services

            • Re: How to monitor old systems?





              Sql team made things easy for me. I gave them the information we can pull and explained the situation. They came back and said they can accept the information given lifting a weight off my shoulder.



              The hp-ux issue. We found that the snmp engine on the os was built to work specifically with hp's monitoring software. So when we messed with the conf file and tried to get it to work it would just shut itself off completely. And we couldn't install net-snmp as my unix guy told me there is no net-snmp engine compatible with such an ancient os. it's prehistoric . Since it's only 3 servers with that os and unix as a whole is making plans to move away from hp-ux in exchange for aix they have told me we can leave these three servers monitored by their home grown monitoring and leave it as up down monitoring in solarwinds. again lifting a weight off my shoulders.



              As far as windows 2003, we found some strange behavior. RPC for some reason won't work at all. Even though the service is running and is used. And wmi works on selected 2003 servers. Most were changed to snmp . and we are experimenting with agents on 2003 and finding that setting them up is proving to be a headache. 2003 has been the problem child servers of our infrastructure. They have a project to remove or upgrade but saying this will take time as owners of the server explain that the tools ran on the server are as old if not older than the os and not supported by new os. Meaning they are kind of stuck with 2003.




              We are having to do alot of customization's due to the fact that we don't have a normal infrastructure  built. And our systems cannot go not monitored because the tool  developers decided to drop support for old systems and software.




              That's why I've been relying on a combination of thwack, internet, and solarwinds support. And upper management has already said that if we say solarwinds can't monitor something they are ready to drop solarwinds for a solution that can monitor it. Technically we don't want to drop it. But we've hit so much brick walls that's been a nightmare trying to build a system according to our needs.




              I appreciate your reply. thank you

                • Re: How to monitor old systems?

                  I think the crux of the issue is that last part,  I'm not aware of any tool that would seamlessly monitor all of those old OS scenarios, but I would be very interested to hear which ones you have in mind.  If you are stuck using 15 year old software then you would also likely be stuck using the "state of the art" monitoring from that same vintage, and most of the vendors still around from that era will all have the same lack of official support for deprecated OS, as in you can try but nobody is going to make you any promises.



                  These systems you describe are perfect examples of the concept of accumulating technical debt.  The more an organization puts off changes like clinging to an EOL OS the more hassles they build up until eventually everything becomes too much of a headache.  I've been there before and it's not a fun task to clean up.

                  1 of 1 people found this helpful
              • Re: How to monitor old systems?

                No response? Any help? Please?