UNIX for Dummies Questions & Answers

this jumping back and forth gets confusing after awhile - a better approach:

Code:

ssh_cmd="ssh $host ps aux|awk -v proc="${process}" '{if( gensub( \".*/\", \"\", \$11 ) == proc) print \$9}'"

Yes, this means that the awk will be reading all of the output records from the ps command, however this is more efficient than searching for the string with grep and then processing the records with awk. For each record that grep matches, there is an extra write and read of that record to get it into the awk process. In addition, there is the extra overhead of starting grep. Since awk can perform the same 'record matching' function that grep does, it makes more sense to do both the matching and the processing of matched records in the same process. If awk were a process that is not capable of selecting the correct records to work on, then you would need to use grep to do the preprocessing.

For something as small as searching output from the ps command, the difference might not be noticeable, but it's good to practice these more efficient habits all of the time.

yes,you're right,if user gives the processx as foo

Code:

host=localhost

process=/usr/sbin/sshd
ssh_cmd="ssh $host ps aux|awk '{if(\$11==\"$process\") print \$9}'" && $ssh_cmd
03:19

process=sshd
ssh_cmd="ssh $host ps aux|awk '{if( gensub( \".*/\", \"\", \$11 ) == \"$process\") print \$9}'" && $ssh_cmd
03:19

will be trying these out... results to follow. thanks!