Sewell, Cassandra D (Cassandra) wrote: > Then are we saying this is possibly a problem in Linux in general, > and has never been fixed. Nope. It's not a general Linux problem, it's a standard UNIX behaviour. > The strange thing about it is that we had no applications running, > and the space was not freed-up/reclaimed until a system reboot was > done (re-reading of the info from the drive). You just didn't look hard enough. lsof is your friend. > We have used other journaled fs's and not seen this problem. We are > currently considering going to EXT3 for production level machines, > and this is considered a critical problem because it incorrectly > blocks us from new file creation. You will see this "problem" also with other file systems, not only journaled ones. I've seen this also on Solaris with UFS. > My assumption right now is that the local copy of the disk > information somehow was corrupted, and the info on drive was fine. Nope, see above. > Evidently the df command uses the local copy and the du gets the info > directly off disk. Is that a correct assumption? I'm not quite sure where du and df get their infos from, but they can report different usage. > I'm trying to get a handle on what the problem may be, so that I can > resolve it. So any help that you or others can offer in helping me > to resolve would be a big plus. The problem is some application on your box. Regards, Juri _______________________________________________ Ext3-users@redhat.com https://listman.redhat.com/mailman/listinfo/ext3-users
