Question

Photo of Andrew Lee

0

Slow Initial Page Load

3

We've been using Rock for several years on a 3 Essentials Virtual Private Machine with virtually no issues.  We have begun experiencing severe slowness when loading the initial login page (Can take several minutes).  Looking at the server the W3WP process is consuming 80 or more percent of the CPU during the page load.  Once the page loads and you login to Rock, the speed seems fine (Unless you let it sit for a bit and then come back to it in which case you'll experience the same slow load).

We are not aware of this issue relating to a recent update either Rock or Windows but am struggling to find an answer.

Any suggestions on where to look/things to check?

Andrew

  • Kenneth Roach

    Kenneth Roach

    I have also been experiencing the same slowness in loading issues. In addition SQL server does not auto start if the server reboots. I suspect something has changed since their platform problems earlier this year.

    • Photo of Jim Michael

      1

      Hi Andrew-

      Slow page load is normal after a server restart on initial page loads (as things get recompiled), but as you say it should stay fast on subsequent page loads. Can you look at Admin | System Settings | Jobs Admin and see if your "Job Pulse" job is running successfully every minute? This is the mechanism that keeps Rock from "falling asleep" by requesting a special page, which helps when users aren't actively accessing the system.

      That said, by default IIS will recycle the Rock AppPool once a day, so at least once per day (usually in the wee hours) Rock will be slow again on initial load, but this usually goes unnoticed. Unless your system is barely used (not being accessed a lot by staff or public users) it should not be falling asleep like you seem to be seeing, unless the Job Pulse isn't running for some reason... or, perhaps, Winhost is recycling your AppPool way too often.


      • Andrew Lee

        Andrew Lee

        Thanks Jim. IIS looks like it's recycling every 29 hours. The Rock Job Pulse is running every 1 minute successfully.


        I opened a ticket with 3 Essentials but they were unsuccessful in finding a culprit.

        • Jim Michael

          Jim Michael

          How married to 3Essentials are you? I know they are listed in the docs as a recommended provider, but over the past few years I see more and more issues with users using them. If you're a 501c3, you can get $3500/yr free on which to run an Azure VM+SQL, which is a VERY common and solid way to run Rock.

          • Andrew Lee

            Andrew Lee

            Not at all! We are a 501c3 and have begun the process of migrating. We're back up and running for now. The fix appears to have been your suggestion of the app app idle timeout and recycle time ---Thank you for those suggestions!

          • Photo of Jim Michael

            0

            You might also check out this blog post https://weblogs.asp.net/owscott/why-is-the-iis-default-app-pool-recycle-set-to-1740-minutes  The default 29 hr recycle time should really be changed to a specific time, like 4am or something (you'll want to make it happen outside of any important Rock jobs trying to run) because otherwise being 29hr it will constantly move on you, to the point that it can/will recycle right in the middle of check-in (yikes!). 


            There's also an Idle Timeout mentioned in the article that should also be disabled. Not sure if those will help your current situation, but they should be configured properly regardless.

          • Photo of Michael Kea

            0

            Hi guys.  Just adding that I have my Rock site uptime checked by StatusCake (free monitoring tool) every 5 minutes.  Not only will this notify me of any issues, but its ping also keeps my site "awake" so that it is very responsive upon login.  FYI - I've been running on Rock for 3 years this way, (Hosted on Azure) and we've never had issues.

            Thanks,