We've got a 3 server environment. 1 development server, 1 test server & 1 (actually 2, db is on a seperate server here) production server. All environment run on the same umbraco version & dll's / plugins etc are all on the same version. The only difference is content. Umbraco Courier & Contour are installed. Concierge gave problems (couldn't edit the partial views / partial view macros anymore, which reminds me i should still report it as a bug...) so we've uninstalled that one again.
All went fine for about 2,5 weeks. After that the application pool started to crash. Sometimes all was going fine for hours & then suddenly the application pool crashes 5 times in a row. I'm not entirely sure but I believe it started after they've used Courier for the first time (nothing was changed for start to that point, Courier was already installed from the beginning).
I've ran a DebugDiag on it & had 6 crashdumps. Analysed them with WinDbg but couldn't find anything weird. (But i haven't used those programs before i must say... i'm not experienced in that area :( )
I also updated the following to the most recent versions: Url Tracker, Mentor WebBlock MVC & CropUp (with ImageResizer paid version).
Also using these packages, but have not updated them yet: Config Tree, uSiteBuilderAdmin (supressSync is on), FamFamFam, Media Icons, Robots.txt Editor.
The server his windows updates have all been installed & server was rebooted.
UmbracoTraceLog nor UmbracoLog nor the application log is showing any real errors other then "Couldn't find any page with nodeId=xxx. This is most likely caused by the page not being published." & "Hibernate.Cfg.Environment - hibernate-configuration section not found in application configuration file"
I also found some Word 'emdashes' in the nodenames but have removed all of them & republished the entire site. That was this morning. The crashes still continued...
Any advice on how I can tackle this issue further? One very important thing, the crash only happens on the PROD server, not sure yet if that is due to people actually visiting the site... ;-)
I've currently uninstalled Courier to see if the crash still happens, I'm really out of ideas here & I think i will probably just start from a fresh server but would really love to know the issue, rather then finding out in 2 or less weeks/days that the same problem is back again :(
Do you still get the crashes now that Courier has been uninstalled? I've heard a lot of complaints about Courier and somehow I think the problem is related to that too.
It wasn't Courier. Not at all. Still having the problem. :(
Performed a load test with loadimpact with 5 different pages. 100 users in 5 min (they normally have about 200 users / hour). App pool didn't crash. But... I'm not quite confident i can compare the loadimpact test with real users. Since loadimpact for example uses the direct link of the CropUp images & not uses the detection scripts to determine the right images-size to use.
I've also added caching on all partials yesterday night (300 secs, caching per page mostly) but saw this morning that cfr the app pool crashes it didn't make any difference.
Analyzing the crash dumps with WinDbg gave us the page causing the crash and the xslt macro involved.
Although we couldn't narrow down the piece of code generating the stack overflow, we managed to optimise the macro a little bit and avoid the crash until we are able to upgrade to v6 or v7.
Application pool crashes
Hello,
We've got a 3 server environment. 1 development server, 1 test server & 1 (actually 2, db is on a seperate server here) production server. All environment run on the same umbraco version & dll's / plugins etc are all on the same version. The only difference is content. Umbraco Courier & Contour are installed. Concierge gave problems (couldn't edit the partial views / partial view macros anymore, which reminds me i should still report it as a bug...) so we've uninstalled that one again.
All went fine for about 2,5 weeks. After that the application pool started to crash. Sometimes all was going fine for hours & then suddenly the application pool crashes 5 times in a row.
I'm not entirely sure but I believe it started after they've used Courier for the first time (nothing was changed for start to that point, Courier was already installed from the beginning).
The error message is the following :
Faulting application name: w3wp.exe, version: 7.5.7601.17514, time stamp: 0x4ce7afa2
Faulting module name: clr.dll, version: 4.0.30319.18408, time stamp: 0x52310752
Exception code: 0xc00000fd
Fault offset: 0x0000000000227a20
Faulting process id: 0xd48
Faulting application start time: 0x01cf13834858e7d4
Faulting application path: c:\windows\system32\inetsrv\w3wp.exe
Faulting module path: C:\Windows\Microsoft.NET\Framework64\v4.0.30319\clr.dll
Report Id: d5371ab1-7f76-11e3-9446-005056a808e9
I've ran a DebugDiag on it & had 6 crashdumps. Analysed them with WinDbg but couldn't find anything weird. (But i haven't used those programs before i must say... i'm not experienced in that area :( )
I also updated the following to the most recent versions: Url Tracker, Mentor WebBlock MVC & CropUp (with ImageResizer paid version).
Also using these packages, but have not updated them yet: Config Tree, uSiteBuilderAdmin (supressSync is on), FamFamFam, Media Icons, Robots.txt Editor.
The server his windows updates have all been installed & server was rebooted.
UmbracoTraceLog nor UmbracoLog nor the application log is showing any real errors other then "Couldn't find any page with nodeId=xxx. This is most likely caused by the page not being published." & "Hibernate.Cfg.Environment - hibernate-configuration section not found in application configuration file"
I also found some Word 'emdashes' in the nodenames but have removed all of them & republished the entire site. That was this morning. The crashes still continued...
Any advice on how I can tackle this issue further? One very important thing, the crash only happens on the PROD server, not sure yet if that is due to people actually visiting the site... ;-)
I've currently uninstalled Courier to see if the crash still happens, I'm really out of ideas here & I think i will probably just start from a fresh server but would really love to know the issue, rather then finding out in 2 or less weeks/days that the same problem is back again :(
thanks,
Gerty
Hello,
Do you still get the crashes now that Courier has been uninstalled? I've heard a lot of complaints about Courier and somehow I think the problem is related to that too.
Jeroen
It wasn't Courier. Not at all. Still having the problem. :(
Performed a load test with loadimpact with 5 different pages. 100 users in 5 min (they normally have about 200 users / hour). App pool didn't crash. But... I'm not quite confident i can compare the loadimpact test with real users. Since loadimpact for example uses the direct link of the CropUp images & not uses the detection scripts to determine the right images-size to use.
I've also added caching on all partials yesterday night (300 secs, caching per page mostly) but saw this morning that cfr the app pool crashes it didn't make any difference.
Hi Gerty,
We are having a similar problem with an old 4.7.2 install, Win Server 2012 R2, IIS 8.5. Which Umbraco version was yours ?
Did you finally found a solution or the root of the problem ?
Thanks in advance
Hi Jose,
From the information in your post, I would recommnd you to check if your are affected by this.
http://issues.umbraco.org/issueMobile/U4-6338
Hope this helps, to solve your issue.
/Dennis
Hi Dennis,
I'm not quite sure it applies to our issue.
Anyway, I checked and the hotfix was already installed.
The problem still there :(
.jac
Hi
Just in case anybody ends up here.
Analyzing the crash dumps with WinDbg gave us the page causing the crash and the xslt macro involved.
Although we couldn't narrow down the piece of code generating the stack overflow, we managed to optimise the macro a little bit and avoid the crash until we are able to upgrade to v6 or v7.
Gerty, Dennis thanks a lot for your help
is working on a reply...