Helpful one-liner to help fix a problem we ran into the other day.<\/p>\n
The thinking is:<\/p>\n Use lsof to get all the open filehandles which conveniently also shows who has it open.<\/p>\n Loop through them, using the Splitting on whitespace. \u00c2\u00a0The input to each iteration of the Since we just care about the count, only use the 3rd column by forcing the output of the split into an array and using a slice.<\/p>\n Now, we just want the count for those users so we increment a hash with the user name as they key.<\/p>\n Of course, the split is returning the name so that gives us the user name and hash key.<\/p>\n And increment that. \u00c2\u00a0Unlike python, for example, you can just increment it.<\/p>\n It will do that for every iteration of the perl -e 'map{$c{(split(\/\\s+\/))[2]}++} `lsof`;print \"$_ $c{$_}\\n\" for (keys %c);'<\/p>\n`lsof`<\/code><\/p>\n `<\/code> as a cheat that it inputs an array<\/p>\nmap { \u00c2\u00a0 } `lsof`;<\/code><\/p>\nmap{ }<\/code> defaults to $_, and if you don't put anything to split in a perl split, it uses $_. \u00c2\u00a0Neat.<\/p>\nsplit(\/\\s+\/)<\/code><\/p>\n (split(\/\\s+\/))[2] <\/code><\/p>\n $c{ }++ <\/code><\/p>\n $c{(split(\/\\s+\/))[2]} <\/code><\/p>\n$c{(split(\/\\s+\/))[2]}++<\/code><\/p>\nmap{ }<\/code>. \u00c2\u00a0i.e. every line in the output of the `lsof`<\/code>.<\/p>\n