1

This is an oddball, but I'm curious if anyone else ran into this. When we installed Sitecore, we used a name that we didn't want to at the end of the day. We went through and stopped the site and app pool, then did a rename on the site, app pool, folder, etc, then started back up.

As soon as we hit Sitecore, we noticed the memory usage of the IIS worker process spike to 99% almost immediately, just trying to hit the admin login page. We reversed out the changes and things returned to normal, then tried reintroducing them one at a time, and the one change that caused the issue was renaming the IIS site. If it stayed on the original name, no problem, but change it to anything else and the memory spiked and held.

Has anyone else seen this and know what might cause it?

9

The only thing I can think off is the fact that renaming the IIS would change Sitecore's instance name if you haven't specified it explicitly with the setting "InstanceName".

Sitecore has several queues and uses this "instance name" to keep track of which servers have done what.

If you have a lot of elements in the EventQueue and PublishQueue tables, Sitecore should consider that it hasn't executed any of these actions, having to reprocess the whole queue. The spike you are experiencing could be explained by this extra work.

I would truncate these tables on core, master and web databases and try again.

  • Nice catch. That could definitely be the issue. – Mark Cassidy Aug 28 '17 at 14:54
  • +1 great answer! I'm going to leave my answer up, but I definitely think that this should be the first thing to try. I added a reference to this in my answer, as well – Zachary Kniebel Aug 28 '17 at 19:47
3

That doesn't sound like a memory leak to me - just not enough information for me to make that conclusion -, but the described symptoms may still indicate a problem.

Before doing anything, have a look at @VincentGaliana's answer, and try clearing out your EventQueue and PublishQueue tables. If the issue persists, proceed with with the below.

To troubleshoot the issue further, first try the following:

  1. Perform the renaming of your IIS site
  2. Verify that all of your permissions are set correctly for your website folder (this should switch over automatically, but this is just to make sure that nothing weird happened)
  3. Clear out all of your Temporary ASP.NET files
  4. Recycle the App Pool from IIS
  5. Try pinging the site again.

If, at this point, memory spikes for more than a few minutes (longer, if your site typically has longer initialization periods) then try the following analysis steps:

  1. Allow the site to continue initializing (with the high memory usage)
  2. Look in the data folder to see if Sitecore is writing to its logs and report any errors that you see
  3. Check to see how SQL is performing (whether running locally or not). If possible, connect to the SQL Profiler and check to see what kinds of queries (if any) are executing
  4. Test out the connection to and availability of MongoDB from the server (using RoboMongo or another tool), and, if you're running MongoDB locally, check the activity of the MongoDB process

It is likely that one or more of the above will help give you an indication of what is causing your memory to spike. If none of the above pan out, I recommend attaching JetBrains dotMemory to the process to analyze exactly where the memory is going.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.