Sponsored Content
Full Discussion: Strange memory behavior
Operating Systems AIX Strange memory behavior Post 302725337 by bakunin on Friday 2nd of November 2012 06:06:53 AM
Old 11-02-2012
Thanks for posting the update.

Quote:
Originally Posted by -=XrAy=-
This system has a lot of filesystems with a lot of small files (cobol sources, etc.).
Maybe the insufficent Inode-cache prevent the System to use the whole FS-cache?
I noticed in another project and under different circumstances that the speed with which to acquire file metadata (basically the contents of inodes) can dramatically speed up file operations:

A network relied heavily on NFS-shares and was redesigned to operate from one huge GPFS-datapool (~500TB). The first thing noticeable was that KDE had to be removed from all the clients, because the damn thing tries to create a hidden file-DB at startup. This is fine when you have a local disk with 20k files, but not when you see several millions of them. (It might be possible to tweak KDE somehow to stop that, but nobody bothered to do so. Desktops are a waste of resources anyway.)

The second observable phenomenon was that backup/restore times could be dramatically improved by moving the metadata onto a SSD. It didn't even have to be big in size: 200-300GB would suffice.

Now this fits in well with what you say about cache sizes and metadata caching. Probably AIX file I/O can be improved by tweaking the resources set aside specifically for dealing with file metadata.

Thanks again for sharing.

bakunin
 

9 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

strange sed behavior

I have a file called products.kp which contains, for example, 12345678,1^M 87654321,2^M 13579123,3 when I run the command cat products.kp| sed -f kp.sed where kp.sed contains s,^M,, I get the output 12345678,1 87654321,2 13579123,3 (5 Replies)
Discussion started by: Kevin Pryke
5 Replies

2. UNIX for Dummies Questions & Answers

Strange Behavior on COM2

Hi, I have a problem with a new touch screen controller that I am trying to use on a SCO 3.0 system. THe touch screen controller only wants to talk at 9600baud. I have updated /etc/inittab per the manual and also edited /usr/lib/event/devices to use 9600 baud. The only way I can get the... (0 Replies)
Discussion started by: Elwood51
0 Replies

3. AIX

Strange Real Memory usage behavior

We have 2 Rs6000 S-85 Each initially had/have 6 processors and 8 GIG of RAM each Node 1 we added 14 processors and 32 GIG of RAM On May 19. (11 days ago) My memory utilization reporting for node 1 showed a spike in available memory 25-30GB for May 19 to 25 . I kind of expected this because I... (5 Replies)
Discussion started by: Skyybugg
5 Replies

4. Shell Programming and Scripting

Very Strange Behavior for redirection

I have searched far and wide for an explanation for some odd behavior for output redirection and haven't come up with anything. A co-worker was working on old scripts which have run for years and embedded in their code were output redirects which worked for the script during execution and then... (5 Replies)
Discussion started by: cahook
5 Replies

5. Shell Programming and Scripting

nawk strange behavior

Dear guys; when deleting repeated lines using nawk as below ; Why the below syntax works? nawk ' !a++' infile > outfile and when using the other below syntax the nawk doesn't work? nawk ' { !a++ } ' infile > outfile or nawk ' { !a++ } ' infile > outfile BR (4 Replies)
Discussion started by: ahmad.diab
4 Replies

6. Programming

Strange behavior in C++

I have the following program: int main(int argc, char** argv){ unsigned long int mean=0; for(int i=1;i<10;i++){ mean+=poisson(12); cout<<mean<<endl; } cout<<"Sum of poisson: "<< mean; return 0; } when I run it, I get the... (4 Replies)
Discussion started by: santiagorf
4 Replies

7. Ubuntu

Ubuntu strange behavior

It is so till login screen. I mean that when I boot my computer, Ubuntu shows a splash screen with mouse instead of Ubuntu logo and in the login screen it shows XUbuntu login screen... It began when I upgraded to previous kernel, I suppose, but I'm not sure... I can't say that it annoys me very... (6 Replies)
Discussion started by: Sapfeer
6 Replies

8. AIX

Strange behavior with tar

I am trying to create an archive using tar. I am specifying a list of directories using the -L option. For testing purposes I created a simple directory structure: /backup/test /backup/test/test1 /backup/test/test2 The file specified by the -L option, named files.txt, contains:... (8 Replies)
Discussion started by: judykstra
8 Replies

9. Shell Programming and Scripting

Strange Ctrl+C behavior

Hello All, I have a strange issue. I've created a shell script which connects to RMAN (Oracle Recovery Manager) and executes full DB backup. I then executed this script with nohup and in the background: $ nohup my_script.sh > logfile.log 2>&1 &The issue is that when I tried to take a look into... (6 Replies)
Discussion started by: JackK
6 Replies
SYS(1)							       AFS Command Reference							    SYS(1)

NAME
sys - Reports the compile-time CPU/operating system type SYNOPSIS
sys DESCRIPTION
The sys command displays the string set at compile time that indicates the local machine's CPU/operating system (OS) type, conventionally called the sysname. This string is the default for the value stored in kernel memory. The Cache Manager substitutes this string for the @sys variable which can occur in AFS pathnames; the OpenAFS Quick Start Guide and OpenAFS Administration Guide explain how using @sys can simplify cell configuration. To set a new value in kernel memory, use the fs sysname command. To view the current value set in the kernel, use either fs sysname or livesys. CAUTIONS
You almost always want to use livesys rather than this command. The sys command displays a single value hard-coded at compile time. It does not query the Cache Manager for the current value and it does not report sysname lists. If you have changed the local system type with fs sysname, or if you run a version of sys compiled differently than the Cache Manager running on the system, the value returned will not match the behavior of the Cache Manager. The only reason to use sys is that livesys wasn't available in older versions of AFS. OUTPUT
The machine's system type appears as a text string: I<system_type> EXAMPLES
The following example shows the output produced on a Sun SPARCStation running Solaris 5.7: % sys sun4x_57 PRIVILEGE REQUIRED
None SEE ALSO
fs_sysname(1), livesys(1) The OpenAFS Quick Start Guides at <http://docs.openafs.org/>. The OpenAFS Administration Guide at <http://docs.openafs.org/AdminGuide/>. COPYRIGHT
IBM Corporation 2000. <http://www.ibm.com/> All Rights Reserved. This documentation is covered by the IBM Public License Version 1.0. It was converted from HTML to POD by software written by Chas Williams and Russ Allbery, based on work by Alf Wachsmann and Elizabeth Cassell. OpenAFS 2012-03-26 SYS(1)
All times are GMT -4. The time now is 07:09 PM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy