memory usage

Alex_Systems_1041 · November 7, 2005, 4:14pm

What is the quickest way to determine what processes
are consuming what amount of RAM?
pidin mem’s output is a bit “busy”

The objective is to try & save memory on a system
which is beginning to become unstable due to lack
of memory. Putting in more RAM is a bit expensive
& so would like to postpone that option.

eg.

pidin in

CPU:X86 Processors:1 FreeMem:1460Kb/125Mb BootTime:Nov 02 11:43:01 CAT 2005
Processor1: 586 Cyrix 586 F5M4S0 300MHz FPU
pidin mem
pid tid name prio STATE code data
stack
1 1 ./boot/sys/procnto 0f READY 1460K 12K
0(320)*
1 2 ./boot/sys/procnto 10r RUNNING 1460K 12K
0(4096)
1 3 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 4 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 5 ./boot/sys/procnto 63r RECEIVE 1460K 12K
0(4096)
1 6 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 7 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 8 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 9 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 10 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 11 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 12 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 13 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
procnto @cfbf7000 12K 12K
8194 1 sbin/tinit 10r REPLY 8192 20K
8192(516K)*
ldqnx.so.2 @b0300000 312K 16K
3 1 proc/boot/slogger 10o RECEIVE 8192 72K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
8196 1 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
8196 2 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
8196 3 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
8196 4 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
ldqnx.so.2 @b0300000 312K 16K
5 1 proc/boot/pci-bios 10o RECEIVE 36K 24K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
6 1 roc/boot/devb-eide 10o SIGWAITINFO 48K 12M
8192(516K)*
6 2 roc/boot/devb-eide 21r RECEIVE 48K 12M
4096(12K)
6 3 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
6 4 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
6 6 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
6 7 roc/boot/devb-eide 10o RECEIVE 48K 12M
4096(16K)
6 8 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
ldqnx.so.2 @b0300000 312K 16K
libcam.so.2 @b8200000 44K 4096
cam-disk.so @b820c000 12K 4096
io-blk.so @b8210000 108K 8192
fs-qnx4.so @b822d000 40K 8192
7 1 proc/boot/devc-con 15o RECEIVE 48K 88K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
/dev/mem @40100000 ( 0) 4096
/dev/mem @40101000 ( b8000) 32K
45064 1 sbin/mqueue 15r RECEIVE 12K 40K
8192(516K)*
ldqnx.so.2 @b0300000 312K 16K
217097 1 usr/sbin/random 10r SIGWAITINFO 20K 340K
8192(516K)*
217097 2 usr/sbin/random 10r RECEIVE 20K 340K
4096(132K)
217097 3 usr/sbin/random 10r NANOSLEEP 20K 340K
4096(132K)
ldqnx.so.2 @b0300000 312K 16K
libm.so.2 @b8200000 60K 8192
libz.so.2 @b8211000 44K 8192
135178 1 sbin/devc-pty 10r RECEIVE 32K 120K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
167947 1 sbin/io-net 10r SIGWAITINFO 60K 716K
8192(516K)*
167947 2 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 3 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 4 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 5 sbin/io-net 10r RECEIVE 60K 716K
4096(132K)
167947 6 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 7 sbin/io-net 21r RECEIVE 60K 716K
4096(132K)
167947 8 sbin/io-net 9r RECEIVE 60K 716K
4096(132K)
167947 10 sbin/io-net 21r CONDVAR 60K 716K
4096(132K)
ldqnx.so.2 @b0300000 312K 16K
npm-tcpip.so @b8200000 236K 56K
devn-rtl.so @b8249000 48K 4096
npm-qnet.so @b8256000 136K 8192
229388 1 usr/sbin/dumper 10r RECEIVE 16K 20K
4096(516K)*

–
Using Opera’s revolutionary e-mail client: http://www.opera.com/m2/

David_Bacon1 · November 7, 2005, 4:30pm

Well, there’s “ps -al” (and try “use ps” for more
refinement…) -

dB

“Alex/Systems 104” wrote ~ Mon, 07 Nov 2005 18:14:16 +0200:

What is the quickest way to determine what processes
are consuming what amount of RAM?
pidin mem’s output is a bit “busy”

The objective is to try & save memory on a system
which is beginning to become unstable due to lack
of memory. Putting in more RAM is a bit expensive
& so would like to postpone that option.

eg.

pidin in

CPU:X86 Processors:1 FreeMem:1460Kb/125Mb BootTime:Nov 02 11:43:01 CAT 2005
Processor1: 586 Cyrix 586 F5M4S0 300MHz FPU
pidin mem
pid tid name prio STATE code data
stack
1 1 ./boot/sys/procnto 0f READY 1460K 12K
0(320)*
1 2 ./boot/sys/procnto 10r RUNNING 1460K 12K
0(4096)
1 3 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 4 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 5 ./boot/sys/procnto 63r RECEIVE 1460K 12K
0(4096)
1 6 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 7 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 8 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 9 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 10 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 11 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 12 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
1 13 ./boot/sys/procnto 10r RECEIVE 1460K 12K
0(4096)
procnto @cfbf7000 12K 12K
8194 1 sbin/tinit 10r REPLY 8192 20K
8192(516K)*
ldqnx.so.2 @b0300000 312K 16K
3 1 proc/boot/slogger 10o RECEIVE 8192 72K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
8196 1 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
8196 2 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
8196 3 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
8196 4 sbin/pipe 10r RECEIVE 16K 52K
4096(132K)
ldqnx.so.2 @b0300000 312K 16K
5 1 proc/boot/pci-bios 10o RECEIVE 36K 24K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
6 1 roc/boot/devb-eide 10o SIGWAITINFO 48K 12M
8192(516K)*
6 2 roc/boot/devb-eide 21r RECEIVE 48K 12M
4096(12K)
6 3 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
6 4 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
6 6 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
6 7 roc/boot/devb-eide 10o RECEIVE 48K 12M
4096(16K)
6 8 roc/boot/devb-eide 15o RECEIVE 48K 12M
4096(16K)
ldqnx.so.2 @b0300000 312K 16K
libcam.so.2 @b8200000 44K 4096
cam-disk.so @b820c000 12K 4096
io-blk.so @b8210000 108K 8192
fs-qnx4.so @b822d000 40K 8192
7 1 proc/boot/devc-con 15o RECEIVE 48K 88K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
/dev/mem @40100000 ( 0) 4096
/dev/mem @40101000 ( b8000) 32K
45064 1 sbin/mqueue 15r RECEIVE 12K 40K
8192(516K)*
ldqnx.so.2 @b0300000 312K 16K
217097 1 usr/sbin/random 10r SIGWAITINFO 20K 340K
8192(516K)*
217097 2 usr/sbin/random 10r RECEIVE 20K 340K
4096(132K)
217097 3 usr/sbin/random 10r NANOSLEEP 20K 340K
4096(132K)
ldqnx.so.2 @b0300000 312K 16K
libm.so.2 @b8200000 60K 8192
libz.so.2 @b8211000 44K 8192
135178 1 sbin/devc-pty 10r RECEIVE 32K 120K
4096(516K)*
ldqnx.so.2 @b0300000 312K 16K
167947 1 sbin/io-net 10r SIGWAITINFO 60K 716K
8192(516K)*
167947 2 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 3 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 4 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 5 sbin/io-net 10r RECEIVE 60K 716K
4096(132K)
167947 6 sbin/io-net 10r RECEIVE 60K 716K
4096(68K)
167947 7 sbin/io-net 21r RECEIVE 60K 716K
4096(132K)
167947 8 sbin/io-net 9r RECEIVE 60K 716K
4096(132K)
167947 10 sbin/io-net 21r CONDVAR 60K 716K
4096(132K)
ldqnx.so.2 @b0300000 312K 16K
npm-tcpip.so @b8200000 236K 56K
devn-rtl.so @b8249000 48K 4096
npm-qnet.so @b8256000 136K 8192
229388 1 usr/sbin/dumper 10r RECEIVE 16K 20K
4096(516K)*

–
Using Opera’s revolutionary e-mail client: > http://www.opera.com/m2/

Alex_Systems_1041 · November 7, 2005, 7:20pm

On Mon, 07 Nov 2005 18:30:24 +0200, David Bacon <dbacon@qnx.com> wrote:

Well, there’s “ps -al” (and try “use ps” for more
refinement…) -

Thanks

David_Gibbs1 · November 11, 2005, 7:16pm

Alex/Systems 104 <acellarius@yah0o.lsd.com> wrote:

What is the quickest way to determine what processes
are consuming what amount of RAM?
pidin mem’s output is a bit “busy”

sin

By default it seems to show: proces code, process data, sum of thread stacks
allocated, sum of thread stacks total.

Or, there is the system resources view in the 6.3.0 SP2 IDE, System
Information perspective.

-David

David Gibbs
QNX Training Services
dagibbs@qnx.com

Alex_Systems_1041 · November 14, 2005, 7:43am

On Fri, 11 Nov 2005 21:16:02 +0200, David Gibbs <dagibbs@qnx.com> wrote:

Alex/Systems 104 <> acellarius@yah0o.lsd.com> > wrote:
What is the quickest way to determine what processes
are consuming what amount of RAM?
pidin mem’s output is a bit “busy”

sin

By default it seems to show: proces code, process data, sum of thread
stacks
allocated, sum of thread stacks total.

Or, there is the system resources view in the 6.3.0 SP2 IDE, System
Information perspective.

Thanks
BTW I never use sin on QNX6, because it seems it’s a stepchild these days
compared to pidin.
-sin takes about 14s to respond
-it has no headers so it’s unclear what the “sin”
output means.
(compare pidin)

This is on a 6.2.1B system the customer is using.

David_Gibbs1 · November 14, 2005, 2:56pm

Alex/Systems 104 <acellarius@yah0o.lsd.com> wrote:

On Fri, 11 Nov 2005 21:16:02 +0200, David Gibbs <> dagibbs@qnx.com> > wrote:

Alex/Systems 104 <> acellarius@yah0o.lsd.com> > wrote:
What is the quickest way to determine what processes
are consuming what amount of RAM?
pidin mem’s output is a bit “busy”

sin

By default it seems to show: proces code, process data, sum of thread
stacks
allocated, sum of thread stacks total.

Or, there is the system resources view in the 6.3.0 SP2 IDE, System
Information perspective.

Thanks
BTW I never use sin on QNX6, because it seems it’s a stepchild these days
compared to pidin.
-sin takes about 14s to respond
-it has no headers so it’s unclear what the “sin”
output means.
(compare pidin)

I agree in the general case. But, in this specific case, the collecting
of the information on memory useage into a “per-process” block seemed to
do exactly what you wanted. The documentation should say what the columns
are.

-David

David Gibbs
QNX Training Services
dagibbs@qnx.com