UNIX for Dummies Questions & Answers

Just a noob here trying to get some help

I am in need of certain information from multiple servers. These servers are Dell PowerEdges with OpenSuSE 10.2 to 11.3.

Here are some lines from /usr/sbin/hwinfo and information I need from those lines:

OEM id: "DELL"
Product: "PowerEdge 850"
Serial: "5GTKL94"

Here are some lines from /proc/cpuinfo and information I need from those lines:

I need total number of cpu cores (16 core)
cpu cores : 4
cpu cores : 4
cpu cores : 4
cpu cores : 4

Here are some lines from /proc/meminfo and information I need from those lines:

If possible display RAM as GB (8 GB)
MemTotal: 8185084 kB

Here are some lines from "df -h" command and information I need from those lines:

Filesystem Size Used Avail Use% Mounted on
/dev/md0 72G 63G 9.2G 88% /



I also need this information but do not have a command to show it yet:

Any help is appreciated!

Cheers

"grep" will not help you (at least not "grep" alone), because grep filters lines with some patterns in it from a text - nothing less, nothing more. It can either display these filtered lines or count them, but that's it.

But as luck will have it, there are more tools in the Unix toolbox than just "grep" and solving your problem is easy:

Code:

OEM id: "DELL"
Product: "PowerEdge 850"
Serial: "5GTKL94"

In this case you have to filter certain parts of the lines. You can use sed, awk or similar tools for this, but i prefer pure shell code, as this runs by far fastest (i simulate your utility by a simple "cat" and a here-document, which will give the output. Replace this with your original command):

Code:

(cat <<EOF
OEM id: "DELL"
Product: "PowerEdge 850"
Serial: "5GTKL94"
EOF) |\
while read LINE ; do
     LINE="${LINE#*\"}"     # chop off everything up to the first double quote
     LINE="${LINE%\"*}"     # chop off everything after the last double quote
     print - "$LINE"
done

The CPU count is a little more tricky, because you have to sum. Use "awk" for such things:

Code:

(cat <<EOF
cpu cores : 4
cpu cores : 4
cpu cores : 4
cpu cores : 4
EOF) |\
awk 'BEGIN {sum=0;} {sum+=$4;} END {printf( "%2d\n", sum);}'

The next is a combination of line filter and cutting some text from the line. You can use "sed" for such things:

Code:

(cat <<EOF
If possible display RAM as GB (8 GB)
MemTotal: 8185084 kB
EOF) |\
sed -n 's/^MemTotal: //p'

The last one is of the same form (line filter plus text cutting), hence the same approach:

Code:

(cat <<EOF
Filesystem Size Used Avail Use% Mounted on
/dev/md0 72G 63G 9.2G 88% /
EOF) |\
sed '1d;s/^.[^ ]* //;s/ .*$//'

To get the number of hdisks is a problem without knowing which hdisks you have: IDE disks are usually named differently than SCSI (SAS) disks, LUNs using one of the many FC-technologies (SDD, Multipath, Powerpath, and so on) often have even several names for the same disk (each way to the fabric gets its own name), etc..

Tell us, which disks you have or expect and we can tell you how to scan for these.

I hope this helps.

bakunin