04-08-2019
Pinned system memory growing constantly
Hi,
Pre data: a server running AIX 6.1 TL9 with 2GB memory and a small amount of CPU, running a very light workload.
I have a server which crashed on lack of memory. After the crash I found - using nmon analyser - that there was something eating up memory. Nmon referred to it as "system".
Now, after 2 days of uptime I see that the same "someting" is eating up memory again. I've been using nmon, vmstat and svmon to see what is happening, but all I see is that
- - all user process is using the same amount of memory
- - vmstat shows a steady decrease on free list
- - svmon -G shows the same steady amount of memory being pinned
- - svmon -P shows that there is no change of pinned memory usage in the listed processes
The rate of memory loss is about 0.5% in every hour.
So I assume that some kernel related object is accumulating the pinned memory.
Can you please help me to find out the problematic part?
Additional info: there was a reported bug in AIX7.2 which have been corrected in an update. But: this machine was running for a while without any problems.
--Trifo
10 More Discussions You Might Find Interesting
1. UNIX for Dummies Questions & Answers
Can you please tell me how to find out System Memory (RAM) for a AIX unix server?
The command prtconf will do for Solaris but I don't know for AIX Unix. (1 Reply)
Discussion started by: anilkumar
1 Replies
2. SCO
In the sco unix ,how can i get the system information ?
such as cpu,memory,interrupter,io ports etc
look forword to quick replay (2 Replies)
Discussion started by: fresh
2 Replies
3. UNIX for Dummies Questions & Answers
Hello all
im working on sunos machine that dont have the top installed and can't be
install , now i need to get information similar to what top gives me about the cpu usage and so
can it be done somehow else where ? (3 Replies)
Discussion started by: umen
3 Replies
4. Programming
Dear all,
When I write the daemon programs it is consuming high memory and processor time. How can I avoid this?
But, the system daemons are not consuming more. How?
Can any one explain how the system daemons are handling the memory consumption and processor time.
Thanks,... (1 Reply)
Discussion started by: nagalenoj
1 Replies
5. UNIX for Advanced & Expert Users
It is a little bit weird to me when i found this on a solaris 9 system with openssh package. Below is the result copied from "top" output:
PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
13491 root 1 59 0 27M 18M sleep 0:06 0.00% sshd -i
20198 root ... (1 Reply)
Discussion started by: sleepy_11
1 Replies
6. HP-UX
Hello, I noticed very high system memory utilization on my new 11.31 Itanium systems. System memory is more than 11GB on 32 GB system. Comparing this to 11.23 it's more than double ... How do I find out what is using it? Is there a way to reduce it?
Thank you,
Kubo (4 Replies)
Discussion started by: trunecm1
4 Replies
7. Solaris
Hi gurus
Im a newbie in solaris..I need to extend file system space in solaris 10 which is using SVM..I have a file system /pin02 which is 93% full n needs to be extended..only 3.6 gb avail space left..the file system is not mirrored...normal ufs file system only..can u please tel me t... (6 Replies)
Discussion started by: madanmeer
6 Replies
8. Shell Programming and Scripting
I am working on a server where the 'root' user ZFS filesystem.
Now when I do Top commands it says only 750M free .But when I count the actual memory utilized it comes only to 12 GB and the total size of the server is 32G.
I think rest of the space is held up by ZFS file system.
Is there a... (5 Replies)
Discussion started by: prasperl
5 Replies
9. AIX
I already checked memory on AIX with "svmon -G" and vmstat (as the figure). I see the value "inuse" nearly=97% use memory, free=3%. However, the value "pin=cache" nearly=26%.
Can I calculate total free memory= free+ pin = 29% ?
please help me!
Thanks so much! (5 Replies)
Discussion started by: xuanthanhnk
5 Replies
10. Linux
Hey everyone. Ok, so I know that from inside of any particular program, it see's through virtualized memory, a full range of available memory. It is given the ability then to place variables, data, user input etc, on the Stack, Heap, BSS, or Code segment of it's range. My question is what does the... (5 Replies)
Discussion started by: Lost in Cyberia
5 Replies
LEARN ABOUT ULTRIX
vmstat
vmstat(1) General Commands Manual vmstat(1)
Name
vmstat - report virtual memory statistics
Syntax
vmstat [ interval [ count ] ]
vmstat -v [ interval [ count ] ]
vmstat -fKSsz
vmstat -Kks namelist [ corefile ]
Description
The command reports statistics on processes, virtual memory, disk, trap, and cpu activity.
If is specified without arguments, this command summarizes the virtual memory activity since the system was last booted. If the interval
argument is specified, then successive lines are summaries of activity over the last interval seconds. Because many statistics are sampled
in the system every five seconds, five is a good specification for interval; other statistics vary every second. If the count argument is
provided, the statistics are repeated count times.
When you run the format fields are as follows:
Procs: information about numbers of processes in various states.
r in run queue
b blocked for resources (i/o, paging, and so on.)
w runnable or short sleeper (< 20 seconds) but swapped
faults: trap/interrupt rate averages per second over the last 5 seconds.
in (non clock) device interrupts per second
sy system calls per second
cs cpu context switch rate (switches/second)
cpu: breakdown of percentage usage of cpu time
us user time for normal and low priority processes
sy system time
id cpu idle time
Memory: information about the use of virtual and real memory. Virtual pages are considered active if they belong to processes which are
running or have run in the last 20 seconds.
avm active virtual pages
fre size of the free list
Pages are reported in units of 1024 bytes.
If the number of pages exceeds 9999, it is shown in a scaled representation. The suffix k indicates multiplication by 1000 and the suffix
m indicates multiplication by 1000000. For example, the value 12345 appears as 12k.
page: information about page faults and paging activity. These are averaged every five seconds, and given in units per second. The size
of a unit is always 1024 bytes and is independent of the actual page size on a machine.
re page reclaims (simulating reference bits)
at pages attached (found in free list not swapdev or filesystem)
pi pages paged in
po pages paged out
fr pages freed per second
de anticipated short term memory shortfall
sr pages scanned by clock algorithm, per-second
disk: s0, s1 ...sn: Paging/swapping disk sector transfers per second (this field is system dependent). Typically paging is split across
several of the available drives. This will print for each paging/swapping device configured into the kernel.
Options
-f Provides reports on the number of forks and vforks since system startup and the number of pages of virtual memory involved in each
kind of fork.
-K Displays usage statistics of the kernel memory allocator.
-k Allows a dump to be interrogated to print the contents of the sum structure when specified with a namelist and corefile. This is
the default.
-S Replaces the page reclaim (re) and pages attached (at) fields with processes swapped in (si) and processes swapped out (so).
-s Prints the contents of the sum structure, giving the total number of several kinds of paging related events that have occurred since
boot.
-v Prints an expanded form of the virtual memory statistics.
-z Zeroes out the sum structure if the UID indicates root privilege.
Examples
The following command prints what the system is doing every five seconds:
vmstat 5
To find the status after a core dump use the following:
cd /usr/adm/crash
vmstat -k vmunix.? vmcore.?
Files
Kernel memory
System namelist
vmstat(1)