Solaris

Is there a command or file I can look at that tells me how much real memory a machine has? A little background. In my shop we run a bunch of java programs, sometimes some of these jobs have config definitions that call for 2G. I would like to know how many I can run before I exhaust rescources. Any feed back would be most helpful

Thanks

sysinfo() often has page size and RAM page count. However, in the age of VM, you can push that a bit, but not so hard that you start the system thrashing.

You exhaust VM when swap is consumed, which df can tell you of. Sometimes swap and /tmp drink from the same disk!

However, if you mmap() files, you can use RAM and address space without exhausting swap. If you munmap() and later mmap() the same pages, they can still be in RAM, so a 32 bit program can use this to access more than 4GB still at RAM speed. A 64 bit program can mmap64() every file on the machine and wander in them as if in RAM.

Dynamic libraries are loaded using mmap/mmap64(), so everyone is executing the same RAM pages. This can sometimes be true of the JAVA JVM and application VM footprint.

If by real memory, you mean RAM, then the simpler way would be:

Code:

prtconf|head

If you have "top" utility. you will have similar output ( rest output omitted).

Greetings, I tried to execut the command prtdiag but received the following error:
prtdiag can only be run in the global zone

The prtconf worked very well. Any ideas?

You are running that command inside non-global zone. That is why you can't use prtdiag.

I am not sure I know the difference between the two. I am not an adminer, i monitor batch and machine resources and stuff like that. Can you point me to some reference on the topic?

Thanks.