Hi all,
In our apps server, we are running 4 JVMs with the following
settings :
JVM_1 : -ms3072M -mx3072M -XX:MaxPermSize=1024M
JVM_2 : -ms3072M -mx3072M -XX:MaxPermSize=1024M
JVM_3 : -ms1024M -mx1536M -XX:MaxPermSize=1024M
JVM_4 : -ms128M -mx128M -XX:MaxPermSize=128M
We are running on 12 GB memory and 2 CPU Xeon dualcore and Redhat 4.
During peak load, when I type free -m command from Linux , the output
is :
Mem total : 12010 used : 11988 free : 21
Swap total : 4096 used : 0 free : 4096
( no matter how long, the swap is never used by the system )
Based on this , am I running short of memory ?
Thank you,
xtanto
|
|
0
|
|
|
|
Reply
|
krislioe (119)
|
2/18/2010 7:59:38 AM |
|
On Wed, 17 Feb 2010 23:59:38 -0800, Krist wrote:
> Hi all,
>
> In our apps server, we are running 4 JVMs with the following settings :
>
> JVM_1 : -ms3072M -mx3072M -XX:MaxPermSize=1024M JVM_2 : -ms3072M
> -mx3072M -XX:MaxPermSize=1024M JVM_3 : -ms1024M -mx1536M
> -XX:MaxPermSize=1024M JVM_4 : -ms128M -mx128M -XX:MaxPermSize=128M
>
> We are running on 12 GB memory and 2 CPU Xeon dualcore and Redhat 4.
>
> During peak load, when I type free -m command from Linux , the output is
> :
>
> Mem total : 12010 used : 11988 free : 21 Swap total : 4096
> used : 0 free : 4096 ( no matter how long, the swap is never
> used by the system )
>
> Based on this , am I running short of memory ?
>
Its not swapping, so its probably happy.
The only thing I notice is that there's only 21 MB left for disk
buffering, etc. Are you noticing slow I/O or reduced network throughput?
Have you used sar or other monitoring tools to examine system operation?
--
martin@ | Martin Gregorie
gregorie. | Essex, UK
org |
|
|
0
|
|
|
|
Reply
|
martin1645 (527)
|
2/18/2010 2:25:49 PM
|
|
According to Martin Gregorie <martin@address-in-sig.invalid>:
> > During peak load, when I type free -m command from Linux , the output is
> > :
> >
> > Mem total : 12010 used : 11988 free : 21 Swap total : 4096
> > used : 0 free : 4096 ( no matter how long, the swap is never
> > used by the system )
> >
> > Based on this , am I running short of memory ?
> >
> Its not swapping, so its probably happy.
>
> The only thing I notice is that there's only 21 MB left for disk
> buffering, etc.
This output of 'free -m' is incomplete. It should look like this:
total used free shared buffers cached
Mem: 7937 1916 6021 0 197 849
-/+ buffers/cache: 869 7068
Swap: 30513 0 30513
Note that there are _two_ 'free' numbers for RAM (not counting swap). On
the first line, this is the totally unused RAM. On the second line, this
is the sum of that unused RAM and the RAM currently used to cache disk
data. Here, my PC is rather freshly rebooted so there are 6 GB
untouched, but already 1 GB used for disk cache. RAM used as a disk
cache can be freed immediately, so it is "kind-of-free".
The amount of unused RAM should gradually lower, down to a small
proportion of the total RAM, while the computer is up (it does not reach
zero because the kernel keeps a bit of unused RAM for some emergency
cases deep within the kernel itself). I guess that the '21' above is the
amount of unused RAM, and it is not a concern.
--Thomas Pornin
|
|
0
|
|
|
|
Reply
|
pornin1 (320)
|
2/18/2010 2:57:37 PM
|
|
On 18 Feb, 21:57, Thomas Pornin <por...@bolet.org> wrote:
> According to Martin Gregorie =A0<mar...@address-in-sig.invalid>:
>
> > > During peak load, when I type free -m command from Linux , the output=
is
> > > :
>
> > > Mem =A0total : =A012010 =A0 =A0 used : 11988 =A0 =A0free : 21 Swap to=
tal : 4096 =A0 =A0
> > > =A0 used : 0 =A0 =A0 =A0 =A0 =A0 free : 4096 ( no matter how long, th=
e swap is never
> > > used by the system )
>
> > > Based on this , am I running short of memory ?
>
> > Its not swapping, so its probably happy.
>
> > The only thing I notice is that there's only 21 MB left for disk
> > buffering, etc.
>
> This output of 'free -m' is incomplete. It should look like this:
>
> =A0 =A0 =A0 =A0 =A0 =A0 =A0total =A0 =A0 =A0 used =A0 =A0 =A0 free =A0 =
=A0 shared =A0 =A0buffers =A0 =A0 cached
> Mem: =A0 =A0 =A0 =A0 =A07937 =A0 =A0 =A0 1916 =A0 =A0 =A0 6021 =A0 =A0 =
=A0 =A0 =A00 =A0 =A0 =A0 =A0197 =A0 =A0 =A0 =A0849
> -/+ buffers/cache: =A0 =A0 =A0 =A0869 =A0 =A0 =A0 7068
> Swap: =A0 =A0 =A0 =A030513 =A0 =A0 =A0 =A0 =A00 =A0 =A0 =A030513
>
> Note that there are _two_ 'free' numbers for RAM (not counting swap). On
> the first line, this is the totally unused RAM. On the second line, this
> is the sum of that unused RAM and the RAM currently used to cache disk
> data. Here, my PC is rather freshly rebooted so there are 6 GB
> untouched, but already 1 GB used for disk cache. RAM used as a disk
> cache can be freed immediately, so it is "kind-of-free".
>
> The amount of unused RAM should gradually lower, down to a small
> proportion of the total RAM, while the computer is up (it does not reach
> zero because the kernel keeps a bit of unused RAM for some emergency
> cases deep within the kernel itself). I guess that the '21' above is the
> amount of unused RAM, and it is not a concern.
>
> =A0 =A0 =A0 =A0 --Thomas Pornin
Hi Thomas,
Thanks for your reply.
The complete output is like this :
total used free shared buffers
cached
Mem: 12010 11980 30 0 149
2326
-/+ buffers/cache: 9504 2505
Swap: 4095 0 4095
Does it mean actually there is still 2505+30 kb Free =3D 2.5 GB ??
Thank you,
Krist
|
|
0
|
|
|
|
Reply
|
krislioe (119)
|
2/20/2010 6:23:42 AM
|
|
According to Krist <krislioe@gmail.com>:
> Does it mean actually there is still 2505+30 kb Free = 2.5 GB ??
Basically, yes. No worry there.
--Thomas Pornin
|
|
0
|
|
|
|
Reply
|
pornin1 (320)
|
2/20/2010 1:53:30 PM
|
|
|
4 Replies
26 Views
(page loaded in 0.093 seconds)
Similiar Articles: How does one find the terminal's width within the JVM? - comp.lang ...... find the terminal's width within the JVM? 6 81 (4/15/2007 7:47:57 AM ... Finding memory leaks through ps command ... Wikipedia, the free encyclopedia A Java virtual machine ... Java memory leak on Solaris - what is "heap inspection tool ...I am trying to investigate a java memory leak, which is reprod... ... due to the scalable architecture of the JVM which has improved in bytecode execution, memory ... Finding memory leaks through ps command - comp.unix.solaris ...It's not impossible for there to be memory leaks in the JVM. It's happened before. ... vsz and rss? - comp.unix.solaris... utilization to check for memory leaks. I am ... Re: proc logistic: 'out of memory' - comp.soft-sys.sas... 80,000 cases and 80,000 controls) - and am receiving >>an 'out of memory' message. here is the code that i am running: >> >>proc logistic data=outf.tendon_short ... Advantage of 64bit Java vs 32bit ? Performance? - comp.lang.java ...I am wondering if there are practically really so much advantages of ... than 1% > I guess the difference is in how much one JVM instance can use RAM memory. V215 error message when upgrading memory - comp.unix.solaris ...dsalt@mbacs.com wrote: > I am attempting to upgrade the memory in a V215 using X8711A 4GB (2GB ... for that "upgrade ... any difference on would be systems so short on memory ... Short String Compression - comp.compressionI am saving considerable memory doing this. At the moment i have ... and only moderately increased memory usage. Unless someone knows of an easily impelmented short string ... DBI (with Oracle) 'out of memory' error - comp.lang.perl.misc ...I am running a select command which returns 355,0... ... In short, you should probably be doing: Prepare SQL ... comp.soft-sys.matlab How to avoid Out of Memory Errors ... I am ... Fast memcpy() ??? - comp.lang.c++Yes I am sure that no one read this February 2010 article in the ... are massively more complicated than a memcpy (e.g. dynamic memory allocation vs embedded short ... how much memory does an empty STL deque occupy? - comp.lang.c++ ...i was wondering whether it pays off in term of memory use ... by an empty container unless I really know that I'm short ... 7/10/2012 2:20:52 AM Java virtual machine - Wikipedia, the free encyclopediaThe Java virtual machine heap is the area of memory used by the JVM, specifically HotSpot, for dynamic ... The young generation stores short-lived objects that are created ... How to avoid 2GB memory limit of JVM in Linux.And it's not just my personal choice, too, of course :) I am not sure my ... I think it is quite a shame that in the current world you can not use JVM memory more ... 7/16/2012 1:56:55 PM
|