On Thu, Jun 3, 2010 at 8:09 PM, Martin Wolff <mwolffedu@xxxxxxxxx> wrote: > > Hi, > I thought this might be an interesting discussion. > > Today, for the first time, I ran into a situation where GNU program couldn't > handle me. I had run "rm ./*" on a directory and it failed because it had > too many arguments (more than 100,000 files existed.). > I then realized that I could just delete the directory just by doing rm -r > directorynamehere > It was only 11 GB in size, but it took forever to delete. I was just curious > if there were any obvious things I should have done differently. Perhaps > there is a better way to trash > 138,000 files? Aha, I'd be interested in a good solution as well. I run into this quite frequently with web applications that cache generated output to the filesystem. Usually the /tmp/ dir has thousands of cache_{md5} files. It's a bit trickier when they're in the same dir with other files you don't want to remove. Usually I chunk it into ranges of files that rm can handle using brackets, rm /tmp/cache_[0-3]* or I have a shell script that just loops through all the files and unlinks them. As for the speed, I know that certain filesystems are much faster at handling small file deletion — in particular reiserfs. I know that xfs has very fast large file deletion and small file directory listing, but ymmv. It's an interesting problem to have. -- Brett Bieber ---- Husker Linux Users Group mailing list To unsubscribe, send a message to huskerlug-request@xxxxxxxxxxxxx with a subject of UNSUBSCRIBE