Specifying patterns to egrep or awk

Shell Programming and Scripting
1

given this information:

^\w{3} [ :0-9]{11} [._[:alnum:]-]+ kernel:( \[[[:digit:]]+\.[[:digit:]]+\])? [[:alnum:]]+: media error \(bad sector\): status=0x[[:xdigit:]]+ { DriveReady SeekComplete Error }$
^\w{3} [ :0-9]{11} [._[:alnum:]-]+ kernel:( \[[[:digit:]]+\.[[:digit:]]+\])? end_request: I/O error, dev [[:alnum:]]+, sector [[:digit:]]+

i need to set up monitoring for these strings. but as you can see, they contain characters that some applications (i.e. Nagios ) will have issues processing as it considers them potentially risky.

so, will the following changes suffice?

kernel:.*media error.*bad sector.*status=0x.*DriveReady.*SeekComplete.*Error 
kernel:.*end_request: I/O error, dev.*sector

i basically just got rid of any character i think Nagios may complain about. and i replaced it with ".*". note, each of these strings will be passed to either egrep or awk to scan a log.

can any one of you experts please help point out what may be wrong with what i just did?

2

Hello SkySmart,

I have made a test file as follows just an example, if there is any change or additional information please let me know.
Input_file:

cat test144
kernel 1234teg media %%4$3@#@gdeyfceb error 1wiyd78t2f52fd)*& bad sector 13693hedh73 status=0x 398r98hfhvr3f DriveReady h3fyg3v SeekComplete @426t3d Error
kernel 243987r2ufh4fh end_request $#:@~! I/O khdwc87e6f7fg error 3513671grftwqhg!@#$vhdc dev e321r2f2fv35 sector

Now following code will look for all strings not the alpha numeric characters in between the words which you need to search as follows.

awk -vs1="kernel media error bad sector status=0x DriveReady SeekComplete Error" 'BEGIN{split(s1, A);s2=length(A)} {for(i=1;i<=NF;i++){S=$i;for(j in A){if(S ~ A[j]){++f;delete A[j]}}};} END{;if(f==s2){print "ERROR LINE FOUND COMPLETLY"}}'  test144

Output will be as follows.

ERROR LINE FOUND COMPLETLY

Similarly you can add more lines for monitoring either they are present or not into the script.

Thanks,
R. Singh

3

I'm having severe difficulties understanding the request. Is that regexes that you apply and want modified? Or strings that need to be matched? What chars would be considered risky?