Hi,

I am facing an issue where one of my site collections (out of around 50 of those) is responding slow on every first visit to the site on a daily basis. Sometimes this happens while trying to access the site after a time gap. This behavior is strange as we've places wake up scripts in our WFEs for all the sites. 

Out of all, only this SC is creating issues. What all point can I check.

Just to mention:-

1. there's no major customization on the site.

2. Site coll is around 900mb in size.

3. A single doc lib has around 30 files, which are ranging around 30mb to 50 mb each.

4. Even if I browse settings page or site content page, I see the same behavior, which removes the possibility of a bulky home page.

Any help is welcome.

The warmup script is only effective for roughly 30 minutes. If you leave the sites without any attempt to access it for 30 mins it'll fall out of memory again.

Thanks for the reply Alex. However, if this is the case than why is only that site slowing down? Other are not. Any pointer I can check.

Im assuming the warm up scripts you are referring to are for the app pools. How many web apps do you have? Is that site collection on its own web app?

Hey Jason,

The Site collection is on a shared web app. Which means there are around 70 site collections in the same web app.

I see that there are around 30 excel files in a document library of this site coll. all of which are > 30 MB in size. Can this be an issue?

No, that is highly unlikely.

Time the site load, time the site load for another site. Time the second load time for both. Post those numbers up here.

Is there more no of SharePont groups ?

It could cause the issue if you explicitly given permission to more no of users and groups.

Thanks Alex Here are the details:

While almost all the other site collections respond in 1-3 seconds, this ONE Site Collection takes around 1.5 mins to load the first time. Immediately after the first load, site works well.

If I leave the site for lets say an hr or so (sometimes even 30 mins), and browse within the site, First load is quick (1-3 sec) but the second load again takes 1-2 mins to load.

Lemmi know if you need any more info.

Dear Sriram,

There are only three SharePoint grps which are created OOB - Visitors, members and Owners.

Are there any other users of the sites? 1-3 seconds means that it's still in memory on the WFE, 90 seconds means that it's being spun up again from disk which is what you'd expect if it's been untouched for more than 25 mins. It's very odd that the first load is quick after an hour of down time, it should take the normal startup time for the first load after that time.

Check for CAPI2 errors, 90 seconds is unusually long and is often caused by CRL checking errors.

• Edited by Alex BrassingtonMicrosoft community contributor Friday, June 26, 2015 9:45 AM

Are there any other users of the sites? 1-3 seconds means that it's still in memory on the WFE, 90 seconds means that it's being spun up again from disk which is what you'd expect if it's been untouched for more than 25 mins. It's very odd that the first load is quick after an hour of down time, it should take the normal startup time for the first load after that time.

Check for CAPI2 errors, 90 seconds is unusually long and is often caused by CRL checking errors.

@Alex; could you provide a source for the 25 minutes drop out? To my best knowledge, sites are only recompiled after doing an application pool recycle, which by default on IIS server happens once a day.

I do admit, I do not know the lifetime of the different caches. (SharePoint cache, asp.net cache) but I would expect it to be more than 25 minutes.

@IMSunny; This might be stupid question, but the default daily application pool recycles are matched with the time the Warm up scripts run, correct? Also, are your warmup scripts hitting all webs or just the top level?

It's the underlying ASP.NET worker process behaviour, if it's inactive for 20 mins (apparently, i can't find any in depth articles to back this up but Google/Bing bring back a fair few mentions) then it gets shut down. This isn't the same as a recompliation.

It's the underlying ASP.NET worker process behaviour, if it's inactive for 20 mins (apparently, i can't find any in depth articles to back this up but Google/Bing bring back a fair few mentions) then it gets shut down. This isn't the same as a recompliation.

@Alex: Nice one Alex! The Idle Time-out setting on the application pool.

Edit: Default value on my IIS 8.5 dev machine is Zero (0)

@IMSummer; What is your current settings for this?

https://technet.microsoft.com/en-us/library/cc771956%28v=ws.10%29.aspx

• Edited by Jesper Arnecke 22 hours 53 minutes ago

Unless im missing something he said they are all on the same web app, which means ANY IIS/App Pool properties are shared. I can't see a scenario where this is an IIS/App Pool configuration issue.

What about Content Databases? Are they all in one?

What does the ULS log say during this 1-2 minutes its loading?

Unless im missing something he said they are all on the same web app, which means ANY IIS/App Pool properties are shared. I can't see a scenario where this is an IIS/App Pool configuration issue.

What about Content Databases? Are they all in one?

What does the ULS log say during this 1-2 minutes its loading?

You might warm up the web application, but not the webs. This does make a difference. Hence a top level warm up script only warms up the frontpage, while more advanced warm up scripts will target each of the webs.

I am a little confused. Had the wake script not been working, how come rest of the Site collections are performing well. Wake up script is for all the site collections under the Web App. 

So is the case with App pool settings.

One more point - The site is opening fine from within the WFE.

• Edited by IMSunny 13 minutes ago