On Fri 05 Mar 2004, Mike Fedyk wrote: > Paul Slootman wrote: > >On Fri 05 Mar 2004, Mike Fedyk wrote: > > > >>The error message[1] below wouldn't be seen unless you browsed through the > >>entire output of tmpreaper mailed to you daily. > > > > > >Hmm, tmpreaper is always silent by default. Do you have --verbose and/or > >--showdeleted as additional options? Perhaps consider not doing that > >then... > > Yes, that's exactly what I wanted. > > It sure is a lot darker with this brown paper bag over my head... :-/ > > > > > > >>I have a large (90GB in 500K files :-/) directory structure managed with > >>tmpreaper (not in /tmp), and the error message could be anywhere in the > >>996KB email I noticed this in. > >> > >>[1] > >>error: run time exceeded! > >>This may be indicative of an attack to use tmpreaper to remove critical > >>files; > >>or the directories to clean up are excessive large and/or messed up. > >>Please investigate. > > > > > >The error message should be at the end, as that's the last tmpreaper > >prints before exiting. Hmm, perhaps stdio buffering of the normal output > >is involved; in that case, it should never be more than 8K from the end > >(at least, I believe stdio's normal buffer is 8K). > > > > But it still continues traversing the rest of the tree. I saw several > pids, so it looks like tmpreaper forks itself before following a > directory from the root or something similar... > > >Perhaps I should add an option to override the maximum allowed elapsed > >time, I never envisoned tmpreaper being used to prune such a large > >tree... > > As long as the max is per fork as mentioned above, it should be good for > me. Let me have a few days run with --showdeleted instead of -v to > check the output to be sure.
OK, you've had a few days :-) Any update? Paul Slootman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
