On Thu, Jun 12, 2014 at 6:45 AM, Timothy Murphy <gayleard at eircom.net> wrote: > >>> Does XFS have any advantages over ext4 for normal users, eg with laptops? >>> I've only seen it touted for machines with enormous disks, 200TB plus. > >> It is generally better at handling a lot of files - faster >> creation/deletion when there are a large number in the same directory. > > I'm wondering if, for the home user, BackupPC would be a good test of that? > Otherwise I can't think of a case where I would have a very large number > of files in the same directory. There are users on the backuppc list that recommend XFS - but for 'home' size systems it probably doesn't matter that much. >> The only down side for a long time has been on 32bit machines where >> the RH default 4k kernel stacks were too small. > > Do you mean that that is a down side of XFS, or ext4? XFS - it needs more working space.. RedHat's choice to configure the kernel for 4k stacks on 32bit systems is probably the reason XFS wasn't the default filesystem in earlier versions. And now that I think of it, this may be an issue again if CentOS revives 32bit support. >>> Does XFS have the same problems that LVM has if there are disk faults? > >> You can't really expect any file system to work if the disk underneath >> is bad. Raid is your friend there. > > In my meagre experience, when a disk shows signs of going bad > I have been able to copy most of ext3/ext4 disks before compete failure, > while LVM disks have been beyond (my) rescue. > Actually, this was in the time of SCSI disks, > which seemed quite good at giving advance warning of failure. I'm not sure what controls the number of soft retries before giving up at the hardware layer. My only experience is that with RAID1 pairs a mirror drive seems to get kicked out at the first hint of an error but the last remaining drive will try much harder before giving up. -- Les Mikesell lesmikesell at gmail.com