Couple of useful Dtrace articles

Using DTrace to understand mpstat and vmstat output

Observing I/O Behavior With The DTraceToolkit

How to get the latest file from the directory

#
way 1:
$ ls -lrt | awk '{ f=$NF }; END{ print f }'
#
#The above awk is to print the last argument of the last line

#
way 2:
$ ls -t1 | head -n

REFERENCE: bash - parameter expansion, i.e. http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_06_02

file=data.txt
  echo ${file%.*}
  echo ${file#*.}

outputs the file base and extension respectively ("data" and "txt").

${variable#pattern} is similar to "echo $variable |sed 's/^pattern//'"
but the pattern matching isn't "greedy" unless you double the #.
Likewise ${variable%pattern} is similar to
"echo $variable | sed 's/pattern$//'"

Read file writting strings into array

declare -a ARRAY
count=0

while read line; do
echo $line
ARRAY[$count]=$line
let count=count+1
echo $count
done < ips.tmp

echo ${ARRAY[1]}
echo ${#ARRAY[@]}

File System workload generator

In addition to the previous post:

FileBench looks like a perfect tool for the load generation.

Here is the article explaining why dd is not the best way to measure HDD performance. The main point is no multithreading in it. vdbench is recommended.

How to find out the symlink source (without readlink)

$ cat basher
#!/usr/local/bin/bash
string="lrwxr-xr-x 1 mmmm users 66 Sep 4 09:58 LINK_SEND -> /apps/aaa/home/mmmm/FFFFF/Apps/RRRRRR/unit_tests/TARGET_SEND"
echo $string
string=${string##* }
echo $string
string=${string%/*}
echo $string
exit 0
$
$
$ ./basher
lrwxr-xr-x 1 mmmm users 66 Sep 4 09:58 LINK_SEND -> /apps/aaa/home/mmmm/FFFFF/Apps/RRRRRR/unit_tests/TARGET_SEND
/apps/aaa/home/mmmm/FFFFF/Apps/RRRRRR/unit_tests/TARGET_SEND
/apps/aaa/home/mmmm/FFFFF/Apps/RRRRRR/unit_tests
$

source

How to check Kernel default values

echo "vx_bc_bufhwm/D" | adb -k /dev/ksyms /dev/mem
echo "ncsize/D" | adb -k /dev/ksyms /dev/mem
echo "vxfs_ninode/D" | adb -k /dev/ksyms /dev/mem

http://seer.entsupport.symantec.com/docs/196721.htm

How to see the full command line in ps?

[root@hostname ~]#/usr/ucb/ps -auxwww | grep -i java
nobody 4589 0.1 4.228480842528 ? S 14:26:35 0:21 /usr/java1.4/bin/java -Djava.awt.headless=true -Xms200m -Xmx300m org.apache.jserv.JServ /jserv/etc/jserv.properties

Get application pid listening on a network port in Solaris (aka netstat -npl in Solaris?)

Source link


#!/bin/bash

# Get the process which listens on port

# $1 is the port we are looking for

if [ $# -lt 1 ]; then
echo "Please provide a port number parameter for this script"
echo "e.g. $0 22"
exit
fi

echo "Greping for your port, please be patient (CTRL+C breaks) ... "

for i in `ls /proc` ; do
pfiles $i | grep AF_INET | grep $1
if [ $? -eq 0 ] ; then
echo Is owned by pid $i
fi
done


To research: In comments to the original post somebody offers to use lsof.

How do I know my global zone nodename from a non-global zone ?

That's what I found on the Web.

For Solaris Zones, the method I use is the one devised by Menno Lageman (http://blogs.sun.com/menno). Mount the global zone's /etc/nodename using a read-only LOFS mount on /etc/chassis in the non-global zone. It automatically changes whenever the nodename of the global zone changes, whether it is because the name of the global zone is modified or because the zone is moved to another physical host using zoneadm detach/attach. Just do this:


# zonecfg -z zone

zonecfg:zone> add fs

zonecfg:zone:fs> set dir=/etc/chassis

zonecfg:zone:fs> set special=/etc/nodename

zonecfg:zone:fs> set type=lofs

zonecfg:zone:fs> add options ro

zonecfg:zone:fs> end

Taken from here:
http://blogs.sun.com/apatoki/entry/how_do_i_know_my

I did not try this yet, I have no appropriate Solaris box by hand. But that could be useful.