Hello,
Apologies if this has been answered elsewhere. I did a quick search and did not find it.
Is it possible to get a total flow count that is being received by Orion from within NTA? I really need this data.
That is a fair amount. To give you an idea, in our stress testing, we begin our lowest tests with 9000 PDUs per second and go up from there. Of course it all depends, mainly on your SQL hardware and if this is just for one node, it wouldn't take many before you would get into a "stressed" situation. (Assuming your SQL hardware is like our lab machines)
You may want to monitor the Performance Counters on the SQL machine to get an idea of how it is handling it as you add more nodes for the NTA service to collect from.
Thanks,
David
Thanks David. The primary polling engine seems to be handling those numbers fine. The only problem is that the NetFlowService.exe service on the polling engine continues to consume more and more memory until it eventually crashes (I suspect this is because it is throttling how fast it writes to the SQL db based on what that server load is). I was told in a different thread to open a case about this and will be doing so later today.
I've known that our SQL server has needed an upgrade since working with you last time. Work is progressing on that front. I may do a test with one of our SANs or simply order a new server with better disks. :-)
Yes, disks have time and time again been shown to be the bottleneck of performance for our customers (as well as in our tests). You mentioned that I have worked with you before, but I haven't associated you with your Thwack name. Did I have you throttle back the checkpointing that SQL server does? It would have been where we added a startup parameter on the SQL Server service of -k. It is something that has continued to help customers who have a SQL Server that is stressed in its disk I/O.
Anyhow, yes, still persue upgrading the hardware and do open a support ticket. Then we'll see what we can do to fix your problem.
BTW, we know that we have a problem here with the NTA service's memory growing in excess and we're testing some fixes and look for other causes as well.
Thanks,
David
Hi David. It may have been a different David that I worked with before. It was on NPM issues that were related to my SQL db disks being the bottleneck. Nothing to do with NTA.
I will open the case in a bit. Thanks again to you all for taking the time to answer questions here.
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 195,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process.