Framework.StartupShutdown

#1

Hello,

Our Dundas BI prod instance is on a Windows Server 2016 The server has 28GB of memory.

We find that the application shuts down a few times throughout the day. I was wondering if anyone else experienced this and what troubleshooting steps you went though to solve it? Thanks for any insights!

I’m using this article as a reference to ensure we are developing with best practice in mind. https://www.dundas.com/resources/blogs/best-practices/best-practices-for-performance-optimization-of-dashboards-in-dundas-bi

See log error below:

image
image.png897×84 4.53 KB

#2

Hi Derek,
Do you see any error in the application log? And do you see a pattern in which this issue is taking place for eg. maybe it is happening when a specific job is running or when the server is too loaded ?

1 Like
#3

If it is the same time every day, check the app pool recycling settings in IIS on the server itself. It could be scheduled to recycle every 8 hours or on a rolling schedule. This stopping and starting will show like that when the app pool gets recycled.

2 Likes
#4

Thank you Christopher, I should’ve mentioned above, there are random times it occurs but the majority of the occurrences are every 4 hours daily up until a certain time in the day. I will ask the IT team to look into this to see if they see anything in the app pool recycling settings in IIS on the server.

#5

Thank you Shashank. The issue occurs every once in awhile randomly but also every 4 hours throughout the day up until a certain time. I haven’t been able to determine a pattern on the random occurrences but will ask the IT team to first look at what Christopher is recommending for the reoccurring shut downs.

#6

It seems to be some kind of pattern, let me know if you get any updates from the IT team.

#7

Thanks @christopher.simpson we did check the app pool recycling settings in the IIS on the server and it was set to recycle on a 4 hour increment. We have since changed it.

For the random occurrences we are still troubleshooting the issue.

Thanks!

2 Likes
#8

I wonder if it could be to do with the server getting over a certain threshhold of CPU/RAM and therefore killing the process?
We did once have such an issue

#9

Thanks @david.glickman it does appear as though the CPU is spiking when a few warehouse jobs with many transforms run. It’s my understanding that I should try to do as many of these transforms in the query whenever possible for better performance.

#10

I don’t really know the details, but we had jobs that were abnormally terminated when the server gave up. We solved that by throttling the process and then offloading the jobs to another server. At that point, I noticed that the framework stopping and restarting was reduced.

Not sure if that helps, just chipping in with my 2cents.