Sorry for wading in here, but I have about 5 minutes to spare, ... Some places use the atime field to archive (or remove) files that are unused. I've worked at several shops where they have large "reports" that just accumulate daily, and they manage their disk space using the atime feature - if none of these reports (typically large text files over 150MB in size) have been accessed within 7 days (or whatever their policy is), they are deleted - or archived. (Of course, it might take them a full week to print one of these files out, but that logic seems to escape the PHBs.)
Keep in mind that their definition of access means "read" - not "modified". But Ben does have a point about the design of the Unix file system - the atime info is considered "file data" and not "file metadata" - if you even sniff the file, then Unix updates the atime field. A little too zealous for most people's needs. Also, its concievable that NFS could use this field to determine when to flush a file from its cache (but NFS cacheing has had so many problems that I don't want to open this can of worms). --Bruce Quoting Benjamin Scott <[EMAIL PROTECTED]>: > On 26 Mar 2002, at 8:03am, Kevin D. Clark wrote: > > I would encourage people to not mess around with [the noatime mount] > > option. > > I would encourage people to do what meets their needs. > > In many situations, the atime field is either not useful, or actively > bad > for performance. > > In general, *I* don't find the "atime" field to be particularly > useful. > Too many things (backups, file searches, etc.) will touch the atime > field. > For example, some of our systems do nightly full backups with GNU tar. > Most files on such systems will have an atime field no older than 24 > hours. > > But that is here. Things will likely be different where you are. :-) > > -- > Ben Scott <[EMAIL PROTECTED]> > | The opinions expressed in this message are those of the author and do > not | > | necessarily represent the views or policy of any other person, entity > or | > | organization. All information is provided without warranty of any > kind. | > > > > > > ***************************************************************** > To unsubscribe from this list, send mail to [EMAIL PROTECTED] > with the text 'unsubscribe gnhlug' in the message body. > ***************************************************************** > ------------------------------------------------- This mail sent through IMP: www.milessmithfarm.net ***************************************************************** To unsubscribe from this list, send mail to [EMAIL PROTECTED] with the text 'unsubscribe gnhlug' in the message body. *****************************************************************
