I have problem in one linux application server where 3 instaces is running of java, each instaces is alocated 1 gb memory but my server is using my 21 gb memory , i am surprised for this.

Can anyone help me , i checked with many command but nothing found….

Thanx in advance….

That’s rather crude. You don’t need to program anything to get these stats nor do you need to write RRD scripts. One method that has been used by many system administrators for a long time is through sysstat though this is all console based. If you prefer a web based GUI with RRD graphs then cacti is already programmed to monitor all of this for you. There are also probably several dozen other ways to do this so you probably shouldn’t use terms like “Then, you have to use RRD”. Because you only know one way to do it doesn’t mean it’s the only way and you shouldn’t be dismissive.

P.S.: You’re explanation seems to be lacking lacking a lot and the first thing I notice is how often you update the RRD is based on how often you have configured and RRD to be updated. An RRD database designed to be updated every 5 minutes would be pretty much useless if you update it every hour based on the way RRD is designed to implement this data. A better approach is to log this data (at least) twice as often you want it to be available, for example, if you want a daily bar graph of cpu usage with 48 points of input from data collected every 1/2 hour then log the data at least every 15 minutes. When you are you want to view this data then you dynamically create the RRD based on the the specifications you supply at the time and this way you are not bound to one poorly designed RRD component which you may have implemented ‘a long time ago” and now you are forced to use it even though you need statistics at an entirely different rate then what you had originally planned “way back when”

2) Update the RRD with cpu and memory details periodically (on every 5-mints or hour)

3) Use rrdtool xport / rrdtool graph command to get details of a specific period (hour/day/week/month/year)

any commands ?

We can take the values from /proc/meminfo and calculate this;
and can verify with gnome-system-monitor output.

free -m

that all helped a lot – even if the information is more than 5 years old!

Itzekocke

Yes and no. Generally speaking, Linux the kernel doesn’t do any real auditing / logging of memory usage but then again it shouldn’t either since Linux really is only a kernel so you will need to install another tool to deal with this. The most common / popular / recommended tool for this would be sysstat / sar and on quite a few distros (Debian, Ubuntu, CentOS), sysstat is the literal name of the package you want to install. Once installed, it will start logging the data but their really is no way for it to record data from before it was installed so basically, if you don’t have a audting / logging application like sysstat already installed then you can’t really view the past history until after it is.