In our company we have installed Epicor 10.1.500.14, we have around 300 users online in an environment with 3 terminal servers (Virtualized), 1 of applications (physical), 1 of DB (physical) and one of Reports (virtualized). The first week of the live outing was going well, but it took two weeks and we noticed that the processes and reports take a long time to process.
I can do to improve the environment.
I appreciate all the help you can give me.
1 appserver for 300 people seems like a stretch
José C Gomez
Senior Software Engineer
If you were to restart the task agent configuration would performance improve? If not
If you were to restart the appserver process that the task agent configuration is pointing to would performance improve? If not
If you were to restart the reporting services Windows service would performance improve? If not
If you were to restart the entire reporting Windows server would performance improve?
By restarting one item at a time it helps isolate where the problem may be.
If performance doesn’t improve you could install and run the performance and diagnostic tool config check to determine if there are any best practices that may need to be implemented.
Dear William@
We are having E10.1 we had lot of issue in the performance Epicor doesn’t
have load balancing feature we have did manually load balance like we have
created multiple virtual machine we have kept 40 users on one virtual
machine configure should be
4 x vCPU
16GB RAM
These VM’s should be on the fastest drives possible, ideally SSD’s, but 15K
SAS will be OK
I am curious about this comment. I don’t want to derail the main tread but we are completely compatible with ARR and F5 type devices and let you choose your preference of networking or Windows Load Balancing approaches. What do you think we are lacking? Genuinely interested
Dear bart@
Have you tried F5 or windows load when end user need E10.1 access using smart client instance when i was having performance i had approach to epicor tech team they have suggested those configuration and number of user on one vm .
The same number of users were maintained in our previous environment 905.702 and we did not present these drawbacks. Our current environment is much better than the previous one as far as hardware is concerned. When we restart all the servers, we have noticed the improvement, but as the days go by we are slow again. I leave a datum. I have a scheduled task to start on Sundays, but it is starting on Saturdays. And so others I have.
Yes every day we have to restart application server and DB server performance will increase .
We run behind an F5 for the SaaS offering so I know it well. My team also did the development and testing on the ARR support. Those are valid deployment options that we discuss in the architecture and Installation guides.
I am more curious your answers to Nathan above and also any results from PDT as to your environment.
Hi Bart@
My environment is different we are using direct smart client application ARR can be useful when we are access E10.1 using http protocol .
<omitted - Brain was not populated with enough caffeine yet>
That still does not answer the results of PDT. I’m curious as to what it reports.
Dear Bart@
Could you please share ARR configuration for net.tcp if you have any guide or document
it helps for me if you don’t mine please
That was me moving too fast. I was mixing up net.tcp on F5 and ARR, my apologies.
The real concern here is the performance and I still have not seen you discuss what PDT recommends.
