(SOLVED) Good/Bad to fill disks to full with XFS?

[je82]

I read in the wiki (which may or may not be very outdated) that:

 

Quote

 

Avoid filling drives >99%

All filesystems can run into fragmentation and resulting performance problems when drives fill all the way up. With ReiserFS in unRAID, it can result in extremely slow write speeds and/or 10 to 30 second pauses where the drive and network share becomes unavailable (and can cause a windows machine trying to write a file to become somewhat unresponsive during that period). These pauses are believed to be the kernel searching the filesystem tree for free blocks or perhaps it is the filesystem adding area to the superblock (Reference). Depending on sizes of the files stored, it it proposed to stay below about 95 to 98% on each drive.

 

 

Is this true for XFS and modern unraid installs? I like to use "fill up" setting because the data i am filling is never going to move so i prefer just filling all the disks asap and move to the next, but is this a mistake?

 

If i should not fill drives to 100% where is the setting to limit it from filling up? That setting would be nesessary when using the "fill up" setting in disk settings if this is indeed bad practice to do.

 

Thanks.

Edited November 13, 2019 by je82

[JonathanM]

Minimum free space in share settings is what you are looking for.

 

XFS is more forgiving than ReiserFS, but it still can be an issue in certain circumstances. If you are moving files on and off of a drive continually, I'd keep a much larger margin than if the drive is pretty much a WORM area. I'd still leave a few GB free in any circumstance, I've seen some file system upgrades that added space requirements for metadata management and there could be space needed for file system health checks and repairs.

 

The exact numbers aren't clear, so you are pretty much on your own to decide what you are comfortable with. Personally I fill WORM drives to 99.9%, my reasoning is that an empty XFS volume reserves roughly 1GB per 1TB capacity for metadata right away, and I figure a file system repair may need that much space again.

[je82]

43 minutes ago, jonathanm said:

Minimum free space in share settings is what you are looking for.

 

XFS is more forgiving than ReiserFS, but it still can be an issue in certain circumstances. If you are moving files on and off of a drive continually, I'd keep a much larger margin than if the drive is pretty much a WORM area. I'd still leave a few GB free in any circumstance, I've seen some file system upgrades that added space requirements for metadata management and there could be space needed for file system health checks and repairs.

 

The exact numbers aren't clear, so you are pretty much on your own to decide what you are comfortable with. Personally I fill WORM drives to 99.9%, my reasoning is that an empty XFS volume reserves roughly 1GB per 1TB capacity for metadata right away, and I figure a file system repair may need that much space again.

Okey problem is now that i have already filled 1 disk it only has 28kb, how do i move files from it to another drive?

 

I see i can simply ssh in and do a mv /mnt/disk#/ /mnt/disk# command, but will this trigger the parity drives to do their magic and write correct data when doing moves commands from a shell between disks?

Edited November 13, 2019 by je82

[itimpi]

10 minutes ago, je82 said:

I see i can simply ssh in and do a mv /mnt/disk#/ /mnt/disk# command, but will this trigger the parity drives to do their magic and write correct data when doing moves commands from a shell between disks?

Working at the /mnt/diskX level WILL update parity appropriately as files/folders are moved around.

[je82]

3 minutes ago, itimpi said:

Working at the /mnt/diskX level WILL update parity appropriately as files/folders are moved around.

cool thanks!