I was doing a quick experiment to see how my algorithm's memory performance looks like.
The input is about 2 Mb and the algorithm takes about 1 second to run it.
I ran this in a loop for 500 times to be able to look at the memory allocation.
This is how jConsole shows the memory usage:
As you can see heap memory usage increases (kinda exponentially) every two times before GC starts (even though the input is the same).
Does anybody know if this is expected and why it happens? Is it some optimization done by JVM?
Thanks!
Does anybody know if this is expected and why it happens? Is it some optimization done by JVM?
The JVM is trying to minimise the time spent GC-ing. If you use more memory, it doesn't have to GC as often.
a leak?
If you look at memory usage after a GC it is much the same so clearly it doesn't have a memory leak. Or at least not a big one.
You have to look at the memory used after Full GCs to confirm there is a memory leak and I assume these are minor collections.
Related
This question already has answers here:
Java using much more memory than heap size (or size correctly Docker memory limit)
(5 answers)
Growing Resident Size Set in JVM
(1 answer)
Closed 2 years ago.
my java service is running on a 16 GB RAM host with -Xms and -Xmx set to 8GB.
The host is running a few other processes.
I noticed that my service consuming more memory over time.
I ran this command ps aux | awk '{print $6/1024 " MB\t\t" $11}' | sort -n on the host and recorded the memory usage by my java service.
When the service started, it used about 8GB memory (as -Xms and -Xmx set to 8GB) but after a week, it used about 9GB+ memory. It consumed about 100MB more memory per day.
I took a heap dump. I restarted my service and took another heap dump. I compared those two dumps but there were not much difference in the heap usage. The dump shows that the service used about 1.3GB before restart and used about 1.1 GB after restart.
From the process memory usage, my service is consuming more memory over time but that's not reported in the heap dump. How do I identify the increase in the memory usage in my service?
I set the -Xms and -Xmx to 8GB. Host has 16GB RAM. Do I set the min/max heap too high (50% of the total memory on the host)? would that cause any issues?
OK so you have told the JVM that it can use up to 8GB for the heap, and you are observing a total memory usage increasing from 1.1GB to 1.3GB. That's not actually an indication or problem per se. Certainly, the JVM is not using anywhere like as much memory as you have said it can do.
The second thing to note is that it is unclear how you are measuring memory usage. You should be aware that a JVM uses a lot of memory that is NOT Java heap memory. This includes:
The memory used by the java executable itself.
Memory used to hold native libraries.
Memory used to hold bytecodes and JIT compiled native code (in "metaspace")
Thread stacks
Off-heap memory allocations requested by (typically) native code.
Memory mapped files and shared memory segments.
Some of this usage is reported (if you use the right tools).
The third thing is that the actually memory used by the Java heap can vary a lot. The GC typically works by copying live objects from one "space" to another, so it needs a fair amount of free space to do this. Then once it has finished a run the GC looks at how much space is (now) free as a ratio with the space used. If that ratio is too small, it requests more memory from the OS. As a consequence, there can be substantial step increases in total memory usage even though the actual usage (in non-garbage objects) is only increasing gradually. This is quite likely for a JVM that has only started recently, due to various "warmup" effects.
Finally, the evidence you have presented does not say (to me!) that there is no memory leak. I think you need to take the heap dumps further apart. I would suggest taking one dump 2 hours after startup, and the second one 2 or more hours later. That would give you enough "leakage" to show up in a comparison of dumps.
From the process memory usage, my service is consuming more memory over time but that's not reported in the heap dump. How do I identify the increase in the memory usage in my service?
I don't think you need to do that. I think that the increase from 1.1GB to 1.3GB in overall memory usage is a red herring.
Instead, I think you should focus on the memory leak that the other evidence is pointing to. See my suggestion above.
Do I set the min/max heap too high (50% of the total memory on the host)? would that cause any issues?
Well ... a larger heap is going to have more pronounced performance degradation when the heap gets full. The flipside is that a larger heap means that it will take longer to fill up ... assuming that you have a memory leak ... which means it could take longer to diagnose the problem, or be sure that you have fixed it.
But the flipside of the flipside is that this might not be a memory leak at all. It could also be your application or a 3rd-party library caching things. A properly implemented cache could use a lot of memory, but if the heap gets too close to full, it should respond by breaking links1 and evicting cached data. Hence, not a memory leak ... hypothetically.
1 - Or if you use SoftReferences, the GC will break them for you.
My point is, since there's limited amount of heapsize, does the JVM need to run garbage collection more frequently? and practically, is it a performance killer?
The optimal amount of memory to use might be 2-5x the minimum heap to run the program. How often the GC is run is inversely proportional to the amount of free memory after a GC.
practically, is it a performance killer?
That depends on your application, but I would assume it is.
Given RAM is relative cheap compared to the cost of your time, I tend to make sure I have plenty of RAM. You can buy 16 GB for less than $80.
This kind of depends on the algorithm used for the gc and the jdk you are using. The normal gc is a killer as it stops execution of the other threads. If you are on jdk 1.6 or better you can make this visible using e.g. visualVM.
There are different gc algorithms to overcome this. Here I would send you to the docs as they the differences best
Finding the right balance between the memory requirements of your application and the memory allocation you give it (using Xmx) is a key performance tuning activity.
Yes, you want to make heap big enough so that the JVM does not end up thrashing on constant GC, which can absolutely be a performance killer.
What heap size you need is totally application dependent.
Our JBoss 3.2.6 application server is having some performance issues and after turning on the verbose GC logging and analyzing these logs with GCViewer we've noticed that after a while (7 to 35 hours after a server restart) the GC going crazy. It seems that initially the GC is working fine and doing a GC every hour or so but at a certain point it starts going crazy and performing full GC's every minute. As this only happens in our production environment have not been able to try turning off explicit GCs (-XX:-DisableExplicitGC) or modify the RMI GC interval yet but as this happens after a few hours it does not seem to be caused by the know RMI GC issues.
Any ideas?
Update:
I'm not able to post the GCViewer output just yet but it does not seem to be hitting the max heap limitations at all. Before the GC goes crazy it is GC-ing just fine but when the GC goes crazy the heap doesn't get above 2GB (24GB max).
Besides RMI are there any other ways explicit GC can be triggered? (I checked our code and no calls to System.gc() are being made)
Is your heap filling up? Sometimes the VM will get stuck in a 'GC loop' when it can free up just enough memory to prevent a real OutOfMemoryError but not enough to actually keep the application running steadily.
Normally this would trigger an "OutOfMemoryError: GC overhead limit exceeded", but there is a certain threshold that must be crossed before this happens (98% CPU time spent on GC off the top of my head).
Have you tried enlarging heap size? Have you inspected your code / used a profiler to detect memory leaks?
You almost certainly have a memory leak and the if you let the application server continue to run it will eventually crash with an OutOfMemoryException. You need to use a memory analysis tool - one example would be VisualVM - and determine what is the source of the problem. Usually memory leaks are caused by some static or global objects that never release object references that they store.
Good luck!
Update:
Rereading your question it sounds like things are fine and then suddenly you get in this situation where GC is working much harder to reclaim space. That sounds like there is some specific operation that occurs that consumes (and doesn't release) a large amount of heap.
Perhaps, as #Tim suggests, your heap requirements are just at the threshold of max heap size, but in my experience, you'd need to pretty lucky to hit that exactly. At any rate some analysis should determine whether it is a leak or you just need to increase the size of the heap.
Apart from the more likely event of a memory leak in your application, there could be 1-2 other reasons for this.
On a Solaris environment, I've once had such an issue when I allocated almost all of the available 4GB of physical memory to the JVM, leaving only around 200-300MB to the operating system. This lead to the VM process suddenly swapping to the disk whenever the OS had some increased load. The solution was not to exceed 3.2GB. A real corner-case, but maybe it's the same issue as yours?
The reason why this lead to increased GC activity is the fact that heavy swapping slows down the JVM's memory management, which lead to many short-lived objects escaping the survivor space, ending up in the tenured space, which again filled up much more quickly.
I recommend when this happens that you do a stack dump.
More often or not I have seen this happen with a thread population explosion.
Anyway look at the stack dump file and see whats running. You could easily setup some cron jobs or monitoring scripts to run jstack periodically.
You can also compare the size of the stack dump. If it grows really big you have something thats making lots of threads.
If it doesn't get bigger you can at least see which objects (call stacks) are running.
You can use VisualVM or some fancy JMX crap later if that doesn't work but first start with jstack as its easy to use.
Let's say I've got an applciation which has a memory leak. At some point the GC will try very hard to clear memory and will slow down my application. I know that if you set this parameter for the JVM -XX:-UseGCOverheadLimit it will throw an OutOfMemoryException:
if more than 98% of the total time is spent in garbage collection and less than 2% of the heap is recovered.
However this is somehow not good enough for me. Because my application will become very slow even before these numbers hit. The GC will absorb the CPU for some time before the OutOfMemoryException will be thrown. My goal is to somehow recognize very early if there will most likly a problem and then throw the OutOfMemoryexception. After that I have some kind of recovery strategy.
Ok now I've found these two additional parameters GCTimeLimit and GCHeapFreeLimit. With them it is possible to tweak the two quoted constants (98% and 2%).
I've made some tests on my own like a small piece of code which produces a memory leak and played with those settings. But I'm not really sure how to find the correct tradeoff. My hope is that someone else had the same problem and came up with a reasonable solution, or maybe there are some other GC switches which i don't know yet.
I'm feeling a little bit lost since I'm not really an expert on this topic and it seems that there are a lot of thing's which can be considered.
If you are using the Sun/Oracle JVM, this page seems to be a pretty complete GC-tuning primer.
You can use java.lang.management.MemoryUsage to determine the used memory, and total memory available. As it approaches the tunable GC collection threshold then you can throw your error.
Of course doing this is a little ridiculous. If the issue is that you need more memory then increase the heap size. The more likely issue is that you're not releasing memory gracefully when you're done with it.
Side-step the JVM heap and use something like Terra Cotta's Big Memory which uses direct memory management to grow beyond the reach of the garbage collector.
Occasionally, somewhere between once every 2 days to once every 2 weeks, my application crashes in a seemingly random location in the code with: java.lang.OutOfMemoryError: GC overhead limit exceeded. If I google this error I come to this SO question and that lead me to this piece of sun documentation which expains:
The parallel collector will throw an OutOfMemoryError if too much time is
being spent in garbage collection: if more than 98% of the total time is
spent in garbage collection and less than 2% of the heap is recovered, an
OutOfMemoryError will be thrown. This feature is designed to prevent
applications from running for an extended period of time while making
little or no progress because the heap is too small. If necessary, this
feature can be disabled by adding the option -XX:-UseGCOverheadLimit to the
command line.
Which tells me that my application is apparently spending 98% of the total time in garbage collection to recover only 2% of the heap.
But 98% of what time? 98% of the entire two weeks the application has been running? 98% of the last millisecond?
I'm trying to determine a best approach to actually solving this issue rather than just using -XX:-UseGCOverheadLimit but I feel a need to better understand the issue I'm solving.
I'm trying to determine a best approach to actually solving this issue rather than just using -XX:-UseGCOverheadLimit but I feel a need to better understand the issue I'm solving.
Well, you're using too much memory - and from the sound of it, it's probably because of a slow memory leak.
You can try increasing the heap size with -Xmx, which would help if this isn't a memory leak but a sign that your app actually needs a lot of heap and the setting you currently have is slightly to low. If it is a memory leak, this'll just postpone the inevitable.
To investigate if it is a memory leak, instruct the VM to dump heap on OOM using the -XX:+HeapDumpOnOutOfMemoryError switch, and then analyze the heap dump to see if there are more objects of some kind than there should be. http://blogs.oracle.com/alanb/entry/heap_dumps_are_back_with is a pretty good place to start.
Edit: As fate would have it, I happened to run into this problem myself just a day after this question was asked, in a batch-style app. This was not caused by a memory leak, and increasing heap size didn't help, either. What I did was actually to decrease heap size (from 1GB to 256MB) to make full GCs faster (though somewhat more frequent). YMMV, but it's worth a shot.
Edit 2: Not all problems solved by smaller heap... next step was enabling the G1 garbage collector which seems to do a better job than CMS.
The >98% would be measured over the same period in which less than 2% of memory is recovered.
It's quite possible that there is no fixed period for this. For instance, if the OOM check would be done after every 1,000,000 object live checks. The time that takes would be machine-dependent.
You most likely can't "solve" your problem by adding -XX:-UseGCOverheadLimit. The most likely result is that your application will slow to a crawl, use a bit more of memory, and then hit the point where the GC simply does not recover any memory anymore. Instead, fix your memory leaks and then (if still needed) increase your heap size.
But 98% of what time? 98% of the entire two weeks the application has been running? 98% of the last millisecond?
The simple answer is that it is not specified. However, in practice the heuristic "works", so it cannot be either of the two extreme interpretations that you posited.
If you really wanted to find out what the interval over which the measurements are made, you could always read the OpenJDK 6 or 7 source-code. But I wouldn't bother because it wouldn't help you solve your problem.
The "best" approach is to do some reading on tuning (starting with the Oracle / Sun pages), and then carefully "twiddle the tuning knobs". It is not very scientific, but the problem space (accurately predicting application + GC performance) is "too hard" given the tools that are currently available.