We are running a .Net 1.1 based Windows Service (not an ASP.Net application), and we are getting System.OutOfMemoryException errors under heavy load.
The service basically hosts an in memory cache, consisting of an Asset hashtable, nested within that is an account hashtable, and within that is a class that stores values for a given time period (for the Asset+Account combination). The service serves up aggregates of this data to clients, as well as accepts updates to the data. The total number of nodes remains constant throughout the service lifetime.
In machine.Config, we see things such as:
<processModel
enable="true"
timeout="Infinite"
idleTimeout="Infinite"
shutdownTimeout="00:00:05"
requestLimit="Infinite"
requestQueueLimit="5000"
restartQueueLimit="10"
memoryLimit="60"
webGarden="false"
cpuMask="0xffffffff"
userName="machine"
password="AutoGenerate"
/>
These all seem to be related to ASP.Net/IIS applications, but our OutOfMemoryException is not occurring under ASP.Net, and there seems to be no equivalent configuration setting for non ASP applications.
Does this section perhaps apply to all .Net based applications, not just ASP.Net?
I ask because, our service was getting up around 1.2 GB of memory consumption (we are storing a large database in memory, yes, with good reason) when the error occurred, which is coincidentally roughly equal to 60% of 2GB (the memory "limit" of 32 bit applications). Could this apparent IIS config setting be causing our windows service memory to be capped as well?
To complicate matters a bit further, we are running this on .Net 1.1 32 Bit, under 64 Bit Windows Server 2003 (yes, we do have a reason for this unusual configuration), with 12 GB of RAM. From what I understand, each 32 Bit process should be able to address up to 4GB of RAM, should it not? Or, does this require changes to either the registry or a .Net config file?
NOTE: I am aware of the /3GB windows startup switch, but since we are on 64 Bit windows, I don't think that should apply (but feel free to correct me if I'm wrong).
Update 1
People seem to agree that processModel configuration is specific to ASP.Net applications only.
One answer says that 32 bit apps on 64 bit OS still have a 2GB per process limit, but most any reference I have been able to find says that each 32 bit process has access to 4GB on a 64 Bit OS. (But, perhaps this only only enabled through setting the IMAGEFILELARGEADDRESSAWARE bit?)
Some relevant links
How to set the IMAGE_FILE_LARGE_ADDRESS_AWARE bit for C# apps:
http://bytes.com/groups/net-c/569198-net-4gt
IIS6 Available Memory for 32-Bit Application with Web Garden on x64 OS (32Gb Ram):
IIS6 Available Memory for 32-Bit Application with Web Garden on x64 OS (32Gb Ram)
.NET Debugging Demos Lab 3: Memory:
http://blogs.msdn.com/tess/archive/2008/02/15/net-debugging-demos-lab-3-memory.aspx
Should be useful to find the specifics of the OutOfMemoryException?
Pushing the Limits - Virtual Memory:
http://blogs.technet.com/markrussinovich/archive/2008/11/17/3155406.aspx
Read this to understand concepts, and use testlimit to rule out machine/config issues. Once convinced it's your app's fault, read & re-read the articles from Tess' blog.
Final Update
Well, for our situation, this turned out to apparently be missing an .Net Service Pack....apparently there was an issue with remoting getting this exception, after the service pack it cleared up entirely!
The processModel configuration element is specific to ASP.NET processes and is not applicable to other .NET processes.
If you are running a 32-bit process on a 64-bit OS, your still subject to the process limit of a 32-bit process, which is 2GB. The practical limit is actually closer to 1.5 to 1.8GB, depending on your application characteristics - in other words, its very unlikely you will ever actually reach the 2GB process barrier.
In order for your Windows service to take advantage of the full 4GB of process space your expecting you will need to:
mark your process as LARGE_ADDRESS_AWARE. Believe this can be done using editbin.exe, but I've never done it! It also might open up a new can of worms... :) I'll see if I can't validate.
add /3GB in boot.ini
reboot server
Also consider the memory allocation profile of your application. If you are allocating objects greater than 85K in size, then these objects will be allocated in the large object heap. The large object heap is swept, but not compacted like other heaps, meaning that you could be experiencing fragmentation which will eventually keep the .net memory manager from allocating a continuous block of memory to satisfy the request.
You likely want to take snaps of the process and review what objects are in what heaps to get a better idea of whats going on within your process memory space.
Also, check the size of the page file on the server. An inadequately sized page file can also cause problems considering its shared across all processes, though that tends to error with system exceptions with some verbiage around 'virtual memory'.
Good luck!
Z
References:
Memory Limits for Windows Releases
Tess Ferrandez, .NET Debugging: Memory
The ProcessModel key is only used for ASP.NET, and even then, on Windows Server 2003 it's pretty much useless because you have the Application Pool configuration.
The /3GB switch does not apply to 64-bit Windows.
As I understand it, you CAN get OutOfMemoryExceptions if you have objects pinned in memory preventing the GC from effectively defragmenting the memory stack when a collection occurs.
You should strongly consider moving your service to a 64-bit application if you know you are jamming gigabytes of data into it. IMO you're playing with fire if you are coming this close to the limit.
Related
Wrote a C# scraper. And analyzed the markup of 30K URLs to pull certain metrics from them.
Run the same code on two machines:
my dev box with 4 core CPU, 8 logical processors and 32GB or RAM. It used up to 300MB of RAM to the end. As I display the WorkingSet size, I could even see the GC kick in and lower memory use, then growing back again.
on a EC2 instance, same data but an instance with only 2 processors and 1.7GB of RAM. Here it used 1.1GB or RAM and, when all threads concluded work, it went down to 300MB just like my local test.
RAM usage was checked with both Environment.WorkingSet and Task Manager. My NET speed is not negligible so I don't think it could affect things even if that Amazon instance might be a little faster. (EC2 net performance differs per instance and this one is on the affordable side hence slower side.)
Why this memory use discrepancy? And can I somehow estimate before hand the memory use in C#?
My guess is that having a slower CPU in the cloud, the GC preferred to keep allocating more than cleaning up what was already used. But this is just my theory to excuse it's unexpected behavior, based on wishful thinking. Still, on my 32GB of RAM it could have used way more but it behaved. On 1.7GB of RAM it went all crazy using 1.1GB of it... I don't get it.
In C++ I just think of how many URLs I fetch at the same time, I think of 256KB average size + size of extracted data and I can tell, before hand, how much memory will be used quite precise. But this C# test left me wondering.
As I plan to release this tool in the wild... I don't feel comfortable taking over half the RAM especially on a lesser machine.
UPDATE: Forgot to mention both machines are Windows 8. Actually one is 8.1 (local) and one Server 2012 (EC2 cloud) both with .NET 4.5.2.
Im a little confused with regards to the memory limitations of an application. As far as i can see, if i write a c# application, targeting x64, my program will have access to 8TB of Virtual address space = space on the HD?
OS >= Windows 7 professional supports 192gigs of RAM. So if i had 192gig system (unfortunately i dont), i could load just over 8.1TB of data into memory (assuming no other processes were running)?
Is virtual memory only used when i have run out available ram? Im assuming there is a performance implication associated with virtual memory vs using RAM?
Apologies if these appear stupid questions, but when it comes to memory management, im rather green.
Your question is actually several related question, taking each individually:
OS >= Windows 7 professional supports 192gigs of RAM. So if i had 192gig system (unfortunately i dont), i could load just over 8.1TB of data into memory (assuming no other processes were running)?
No, it would still be 8 TB. That is the maximum amount of addressable space, whether it is in RAM or elsewhere.
However you could never have 8 TB in use, even if you some how unloaded Windows itself, as the OS needs to keep track of the space being used. In total, you could probably get to 7 TB approximately.
is virtual memory only used when i have run out available ram?
No, if you have virtual memory turned on the entirety of RAM is typically preloaded onto your HDD (give or take a few seconds). This allows the OS to unload something to make room if it feels the need, without having to persist the data. Note that the OS keep thorough track so will know if this is the case or not.
Im assuming there is a performance implication associated with virtual memory vs using RAM?
Depends on your context. Every seek on the hard drive takes a computational eternity, however it is still a fraction of a second. Assuming your process isn't thrashing and repeatedly accessing virtual memory, you should not notice a significant performance hit outside high performance computing.
Apologies if these appear stupid questions, but when it comes to memory management, im rather green.
Your main problem is you have some preconceived notions about how memory works that don't line up with reality. If you are really interested you should look into how memory is used in a modern system.
For instance, most people conceptualize that a pointer points to a location in memory, since it is the fundamental structure. This isn't quite true. In fact the pointer contains a piece of information that can be decoded into a location in the addressable space of the system, which isn't always in RAM. This decoding process uses quite a few tricks that are interesting, but beyond the scope of this question.
Normally, you should write applications targeting Any CPU. The .NET loader then decides (depending on the platform it is running on) which version of the run-time environment will execute the application and into what kind of native code it will be compiled. There is no need to specify the platform, unless you are using custom native components which will be loaded into the process created for your application. This process is then associated with some virtual address space - how that is mapped to physical memory is managed by the OS...
We have a 64bit C#/.Net3.0 application that runs on a 64bit Windows server. From time to time the app can use large amount of memory which is available. In some instances the application stops allocating additional memory and slows down significantly (500+ times slower).When I check the memory from the task manager the amount of the memory used barely changes. The application keeps on running very slowly and never gives an out of memory exception.
Any ideas? Let me know if more data is needed.
You might try enabling server mode for the Garbage Collector. By default, all .NET apps run in Workstation Mode, where the GC tries to do its sweeps while keeping the application running. If you turn on server mode, it temporarily stops the application so that it can free up memory (much) faster, and it also uses different heaps for each processor/core.
Most server apps will see a performance improvement using the GC server mode, especially if they allocate a lot of memory. The downside is that your app will basically stall when it starts to run out of memory (until the GC is finished).
* To enable this mode, insert the following into your app.config or web.config:
<configuration>
<runtime>
<gcServer enabled="true"/>
</runtime>
</configuration>
The moment you are hitting the physical memory limit, the OS will start paging (that is, write memory to disk). This will indeed cause the kind of slowdown you are seeing.
Solutions?
Add more memory - this will only help until you hit the new memory limit
Rewrite your app to use less memory
Figure out if you have a memory leak and fix it
If memory is not the issue, perhaps your application is hitting CPU very hard? Do you see the CPU hitting close to 100%? If so, check for large collections that are being iterated over and over.
As with 32-bit Windows operating systems, there is a 2GB limit on the size of an object you can create while running a 64-bit managed application on a 64-bit Windows operating system.
Investigating Memory Issues (MSDN article)
There is an awful lot of good stuff mentioned in the other answers. However, I'm going to chip in my two pence (or cents - depending on where you're from!) anyway.
Assuming that this is indeed a 64-bit process as you have stated, here's a few avenues of investigation...
Which memory usage are you checking? Mem Usage or VMem Size? VMem size is the one that actually matters, since that applies to both paged and non-paged memory. If the two numbers are far out of whack, then the memory usage is indeed the cause of the slow-down.
What's the actual memory usage across the whole server when things start to slow down? Does the slow down also apply to other apps? If so, then you may have a kernel memory issue - which can be due to huge amounts of disk accessing and low-level resource usage (for example, create 20000 mutexes, or load a few thousand bitmaps via code that uses Win32 HBitmaps). You can get some indication of this on the Task Manager (although Windows 2003's version is more informative directly on this than 2008's).
When you say that the app gets significantly slower, how do you know? Are you using vast dictionaries or lists? Could it not just be that the internal data structures are getting so big so as to complicate the work any internal algorithms are performing? When you get to huge numbers some algorithms can start to become slower by orders of magnitude.
What's the CPU load of the application when it's running at full-pelt? Is actually the same as when the slow-down occurs? If the CPU usage decreases as the memory usage goes up, then that means that whatever it's doing is taking the OS longer to fulfill, meaning that it's probably putting too much load on the OS. If there's no difference in CPU load, then my guess is it's internal data structures getting so big as to slow down your algos.
I would certainly be looking at running a Perfmon on the application - starting off with some .Net and native memory counters, Cache hits and misses, and Disk Queue length. Run it over the course of the application from startup to when it starts to run like an asthmatic tortoise, and you might just get a clue from that as well.
Having skimmed through the other answers, I'd say there's a lot of good ideas. Here's one I didn't see:
Get a memory profiler, such as SciTech's MemProfiler. It will tell you what's being allocated, by what, and it will show you the whole slice n dice.
It also has video tutorials in case you don't know how to use it. In my case, I discovered I had IDisposable instances that I wasn't Using(...)
is there anyway i can have my application tell how much memory the user has and if the application is getting close to taking up a high percentage of that.
also, how do you know how much memory the machine gives to OS, video cards, etc . .
for example, if you have 4gb of memory, how much actual memory is given to applications, can you configure this.
is there anyway i can have my application tell how much memory the user has and if the application is getting close to taking up a high percentage of that.
Yes, it's possible (see some of the other answers), but it's going to be very unlikely that your application really needs to care. What is it that you're doing where you think you need to be this sensitive to memory pressure?
also, how do you know how much memory the machine gives to OS, video cards, etc . .
Again, this should be possible using WMI calls, but the bigger question is why do you need to do this?
for example, if you have 4gb of memory, how much actual memory is given to applications, can you configure this.
No, this isn't a configurable value. When a .NET application starts up the operating system allocates a block of memory for it to use. This is handled by the OS and there is no way to configure the algorithms used to determine the amount of memory to allocate. Likewise, there is no way to configure how much of that memory the .NET runtime uses for the managed heap, stack, large object heap, etc.
I think I read the question a little differently, so hopefully this response isn't too off topic!
You can get a good overview of how much memory your application is consuming by using Windows Task Manager, or even better, Sysinternals Process Monitor. This is a quick way to review your processes at their peaks to see how they are behaving.
Out of the box, an x86 process will only be able to address 2GB of RAM. This means any single process on your machine can only consume up to 2GB. In reality, your likely to be able to consume only 1.5-1.8 before getting out of memory exceptions.
How much RAM your copy of Windows can actually address will depend on the Windows version and cpu architecture.
Using your example of 4GB RAM, the OS is going to give your applications up to 2GB of RAM to play in (which all processes share) and it will reserve 2GB for itself.
Depending on the operating system your running, you can tweak this, using the /3GB switch in the boot.ini, will adjust that ratio to 3GB for applications and 1GB for the OS. This has some impact to the OS, so I'd review that impact first and see if you can live with tradeoff (YMMV).
For a single application to be able to address greater than /3GB, your going to need to set a particular bit in the PE image header. This question/answer has good info on this subject already.
The game changes under x64 architecture. :)
Some good reference information:
Memory Limits for Windows Releases
Virtual Address Space
I think you can use WMI to get all that information
If you don't wish to use WMI, you could use GlobalMemoryStatusEx():
Function Call:
http://www.pinvoke.net/default.aspx/kernel32/GlobalMemoryStatusEx.html
Return Data:
http://www.pinvoke.net/default.aspx/Structures/MEMORYSTATUSEX.html
MemoryLoad will give you a number between 0 and 100 that represents the ~ percentage of physical memory in use and TotalPhys will tell you total total amount of physical memory in bytes.
Memory is tricky because usable memory is a blend of physical (ram) and virtual (page file) types. The specific blend, and what goes where, is determined by the operating system. Luckily, this is somewhat configurable as Windows allows you to stipulate how much virtual memory to use, if any.
Take note that not all of the memory in 32-bit Windows (XP & Vista) is available for use. Windows may report up to 4GB installed but only 3.1-3.2GB is available for actual use by the operating system and applications. This has to do with legacy addressing issues IIRC.
Good Luck
What is the largest heap you have personally used in a managed environment such as Java or .NET? What were some of the performance issues you ran into, and did you end up getting a diminishing returns the larger the heap was?
I work on a 64-bit .Net system that typically uses 9-12 GB, and sometimes as much as 20GB. I have not seen any performance problems even while garbage collecting, and I have been looking hard as I was not expecting it to work so well.
An earlier version hung on to some objects for too long resulting in occasional GCs that freed up 3GB+. Even then, there was no noticeable impact on performance. The system is running on a 16-core server with 32GB RAM, which probably helps...
In .Net, on Windows 32-bit, You can only really get to about 1.4 GB of memory usage before things start getting really screwy (out of memory exceptions). This is due to a limitation in 32 bit windows that limits a single process to using more than 2 GB of RAM. There is /3GB switch you can put in your boot.ini, but that will only bring you a little bit further. If you want to use lots of memory, you should seriously consider running on a 64 bit version of windows.
I currently have a production application with 6 GB of memory. You'll need a 64-bit box as well for the JVM to be able to address that much.
The garbage collector is really the only thing (that I've found so far) where performance degrades with size, and then only if you manually kick off a System.GC, which forces the JVM to bring everything to a screeching halt as it traverses 6 GB worth of objects. Takes a good 20 seconds, too. The default GC behavior does not do this, BTW, you have to be dumb enough to make it do that. Also worth researching JVM tuning at this size.
You can also find things like distributed and clustered JVMs, sorry, don't have any good references as I didn't look into this option too closely, although I did find references to larger installations.
I am unsure what you mean by heap, but if you mean memory used, I have used quite a bit, 2GB+. I have a web app that does image processing and it requires loading 2 large scan files into memory to do analysis.
There were performance issues. Windows would swap out lots of ram, and then that would create a lot of page faults. There was never any need for anymore than 2 images at a time as all requests were gainst those images (I only allowed 1 session per image set at a time)
For instance, to setup the files for initial viewing would take about 5 seconds. Doing simple analysis and zooming would be fairly fast once in memory, in the order of .1 to .5 seconds.
I still had to optimize, so I ended up preparsung the files and chopping into smaller peices and worked only with the peices that were required by the user at the time.
I have used from 2GB to 5GB of memory in java, but usually when I get to more than 2GB I really start thinking about memory optimization. Diminishing returns can vary from not optimizing when it's necessary because you have a lot of memory, to not having memory available for the OS/Disk caches (which can help your application overall).
For Java, I recommend watching your memory usage per generation over time. Do you create a lot of temporary objects or have long-lasting objects that consume a lot of memory? A lot of optimization of memory can be done when knowing those things.