I've got a client with a virtuel server where we are running 5 Umbraco installations. The oldest one (2.1.7) is lightning fast and very stable. There's also another installation running 4.0.3 which runs quite well. The newest installation (4.7.1) is schedueled to replace the oldest from Jan. 1st, but we're currently experienced major problems with it randomly timing out or just hanging for some seconds.
Server specs are Windows 2008 R2 SP1 (64 bit) on an Intel Xeon @ 2 GHz with 1 GB RAM. SQL Server 2008.
I realize too little memory could be the issue, but shouldn't this affect all sites? As of now the process tree normally shows memory usage in the area of 900 MB with peaks up to 950 MB, but I've never seen it higher.
All sites are running within their own application pools.
Any ideas what to do in order to find out what's causing these issues? Also what's recommended for load testing Umbraco (will be next step - as of now only me and a few test users are accessing the site)?
We also have performance problems with Umbraco 4.7.1 due to macros. We have around 12/14 macros per page and in the current version each time you call a macro there is a DB access to load its details. Niels is working on it and there is already a fix (not yet for production). You may have a look at this page: Severe Issue with macro caching under heavy load.
Concerning the load test I use the jMeter tool which is really powerful and free.
I've already got the server upgraded with 0,5 extra RAM, but it does not seem to help. I'll be keeping fingers crossed for 4.7.1.1 to fix the problem.
What I discovered while looking at the Resource Monitor is that when the problems occur the monitor reports as many as 50-60 hard fault/sec for the application pool (w3wp.exe). My ISP sent me the following definition regarding this:
Hard Fault vs. Page FaultFirst of all, a "hard fault" was previously called a "page fault" in earlier versions of Windows. Perhaps page faults were more easily understood from the name, too. A hard fault happens when the address in memory of part of a program is no longer in main memory, but has been instead swapped out to the paging file, making the system go looking for it on the hard disk. When this happens a lot, it causes slowdowns and increased hard disk activity. When it happens an awful lot, the possibility of hard disk thrashing arises. That's when a program stops responding, but the hard drive continues to run for an extended period. This has historically been referred to as "getting into the page file." I assume this could be related to the performance issues and the fact that there are a lot of DB access going on. After adding more RAM the SQL server instance has been eating up most of it (I've allowed a maximum of 512 MB).
Performance issues with 4.7.1
I've got a client with a virtuel server where we are running 5 Umbraco installations. The oldest one (2.1.7) is lightning fast and very stable. There's also another installation running 4.0.3 which runs quite well. The newest installation (4.7.1) is schedueled to replace the oldest from Jan. 1st, but we're currently experienced major problems with it randomly timing out or just hanging for some seconds.
Server specs are Windows 2008 R2 SP1 (64 bit) on an Intel Xeon @ 2 GHz with 1 GB RAM. SQL Server 2008.
I realize too little memory could be the issue, but shouldn't this affect all sites? As of now the process tree normally shows memory usage in the area of 900 MB with peaks up to 950 MB, but I've never seen it higher.
All sites are running within their own application pools.
Any ideas what to do in order to find out what's causing these issues? Also what's recommended for load testing Umbraco (will be next step - as of now only me and a few test users are accessing the site)?
Hello Nikolas,
We also have performance problems with Umbraco 4.7.1 due to macros. We have around 12/14 macros per page and in the current version each time you call a macro there is a DB access to load its details. Niels is working on it and there is already a fix (not yet for production). You may have a look at this page: Severe Issue with macro caching under heavy load.
Concerning the load test I use the jMeter tool which is really powerful and free.
Thanks Christophe,
I've already got the server upgraded with 0,5 extra RAM, but it does not seem to help. I'll be keeping fingers crossed for 4.7.1.1 to fix the problem.
What I discovered while looking at the Resource Monitor is that when the problems occur the monitor reports as many as 50-60 hard fault/sec for the application pool (w3wp.exe). My ISP sent me the following definition regarding this:
Hard Fault vs. Page FaultFirst of all, a "hard fault" was previously called a "page fault" in earlier versions of Windows. Perhaps page faults were more easily understood from the name, too. A hard fault happens when the address in memory of part of a program is no longer in main memory, but has been instead swapped out to the paging file, making the system go looking for it on the hard disk. When this happens a lot, it causes slowdowns and increased hard disk activity. When it happens an awful lot, the possibility of hard disk thrashing arises. That's when a program stops responding, but the hard drive continues to run for an extended period. This has historically been referred to as "getting into the page file." I assume this could be related to the performance issues and the fact that there are a lot of DB access going on. After adding more RAM the SQL server instance has been eating up most of it (I've allowed a maximum of 512 MB).is working on a reply...