A fun little awk help. Legacy device getter.

[Steven E. Protter]

Ouptput:

disk 29 0/6/2/0.151.15.23.0.0.4 sdisk CLAIMED DEVICE IBM 2145
/dev/dsk/c19t0d4 /dev/rdsk/c19t0d4


That is standard 11.23 ioscan.

I have the -H hardware device.

0/0/1/1.15.0

Which I got from setboot

I am needing the legacy device. Can't use ioscan -m because that useful little tool does not exist.

It will look kind of like:

I can't get the awk working today.


ioscan -kfnC fc | awk ‘/fcd/ {getline;fcd=$NF;print fcd,$2}’ | while read -r fdev
do

fcmsutil ${fdev} | awk ‘/Hardware / {print $5};/World / { print $7}’

done


Something like:
ioscan -kfnCdisk | awk '/$pboot/ {getline;fcd=$pboot; print fcd,$2}'

Except maybe something that works.

SEP

Was thinking:

[James R. Ferguson]

Hi SEP:

> ioscan -kfnCdisk | awk '/$pboot/ {getline;fcd=$pboot; print fcd,$2}'

If you are attempting to pass 'pboot' as a variable, this isn't gong to work. You should do something like:

# ioscan -kfnCdisk | awk -v pboot=${pboot} '$0~pboot {getline print pboot,$2}'

Regards!

...JRF...

[Steven E. Protter]

ioscan -kfnCdisk | awk -v pboot=${pboot} '$0~pboot {getline print pboot,$2}'

I get the concept.

awk: The statement cannot be correctly parsed.
The source line is 1.

I'm hacking away here.

This looks almost done.

Steven

[Steven E. Protter]

Fixed it:

ioscan -kfnCdisk | awk -v pboot=${pboot} '$0~pboot {getline;print pboot,$2}'

Chime in JRF for full pointage.

SEP

[James R. Ferguson · Accepted solution]

Hi (again) SEP:

Oops, i forgot the semicolon :-) I meant:

# ioscan -kfnCdisk | awk -v pboot=${pboot} '$0~pboot {getline; print pboot,$2}'

...just as you deduced.

Regards!

...JRF...

[Viktor Balogh]

good to know how to import shell variables inside awk. I always used a much ugly way by alternating quotes and escape characters. :)

# ioscan -kfnCdisk | awk '/'$pboot'/ {getline;fcd='$pboot'; print fcd,$2}'

[Steven E. Protter]

It was still kind of wrong.

abootdisk=$(ioscan -kfnCdisk | awk -v aboot=${aboot} '/aboot/ $0~aboot {getline;print aboot,$2}' | awk '{ print $2 }');

That is the final version.

Done, assigning points.

awk is awkful
awk is fun, more than enough for anyone.

SEP

[Viktor Balogh]

I've got two things what's not so clear to me:

why did you use both /aboot/ and $0~aboot ? both says "if the line contains the value of the variable aboot, do this..."

and why did you give the output to a second awk statement? (which could be done inside the first awk) that seems to me a performance hog.

[Steven E. Protter]

I stopped working on this when it worked.

I will award points for working more efficient code.

The statement executes twice as part of a process that chooses alternate boot disk to be broken out of a two disk root/boot mirror set as a target for a DRD clone.

If it were executing hundreds of times, I would work harder on efficiency.

Its not, so I stopped. I am not averse to a theoretical exercise to improve the code.

The current code is the top new post at http://hpux.ws

I would love to tune it up and make it more awkful.

SEP

[Dennis Handly]

>Viktor: why did you use both /aboot/ and $0 ~ aboot?

I'm not sure what both really do together?

>both says "if the line contains the value of the variable aboot"

No, /aboot/ says look for the constant RE containing "aboot", not the variable. That's why JRF switched $0 ~ aboot

>why did you give the output to a second awk statement?

(I assumed the extra token was to make sure it worked?)
Not really done, just changed: :-)
... | awk -v aboot=${aboot} '$0 ~ aboot {getline; print $2}')