We still have constant complaints that the web interface is still painfully slow. I've experienced myself watching the spinning chaser go on and on for ~30 - 60 seconds. Watching the server while this is happening shows no high CPU (~12 - 20%), barely any usage on the SM service and IIS worker, our primary SM installation drive (an NVME RAID) is doing nothing but twiddling its thumbs and our main storage array is also not really doing much.

 

I watched and logged a period over an hour, the disk queue on mail storage never went above 2 (24x 1TB SAS RAID 10), the queue on the primary installation drive never goes above 0, CPU stayed below 39%, there is just nothing physically on the server that seems to be holding SM back. I even went as far as contacting BroadCom (LSI) and had them review the controller to make sure it was not bottlenecked and they stated that the controller processor is only at max over a 24 hour period at 39% (and that was during an intensive backup operation).

 

This is doing silly things like opening a folder that has no emails in it, somehow takes ~30 seconds, switching between folders is a nightmare, have numerous complaints that sending emails (with no attachments) takes minutes sometimes. Even customers with decently powerful machines are complaining of load times now and across all browsers.

 

I am still not sure that a single worker process is the way to go, it seems that having webmail, EWS, EAS, WebDav for CalDAV etc. (possibly other items as well?) running over a single worker process can hurt a single worker? Back when I tested 2 workers, it did "feel" like it improved services, numerous customers reported it felt snappier and that the spinny ball of death nearly went away.

 

Are there any SM 16 users still experiencing slowness?