#How do you many files from many forders quick

1 messages · Page 1 of 1 (latest)

rose saffron
#

Currently we use a find an mtime to delete thousands and More files from thousands of forders which needs on cloud storage very long time. Is there any fester Option to delete files with mtime from a folder but not all files or the complete folders, which makes asynchron delete useless

brittle sable
#

you can use client-side async delete which basically means to move files or folders into a "trashcan" directory which then gets deleted by ONTAP in the background, or directly delete them on the cluster. But since you still need to scan through all directories to find the files to delete, this is only faster if you move complete folders into the trashcan (since you then don't need to recurse into them). Whether that helps in you case or not I cannot decide

steady carbon
#

I have a powershell script I put together to remove anything older than 60 days from a shared temp directory (NFS/CIFS) (commented out directory part bc no subdirs in dir)

`$limit = (Get-Date).AddDays(-60)
$path = "C:\temp"

Delete files older than the $limit.

Get-ChildItem -Path $path -Recurse -Force | Where-Object { !$.PSIsContainer -and $.CreationTime -lt $limit } | Remove-Item -WhatIf

Delete any empty directories left behind after deleting the old files.

#Get-ChildItem -Path $path -Recurse -Force | Where-Object { $.PSIsContainer -and (Get-ChildItem -Path $.FullName -Recurse -Force | Where-Object { !$_.PSIsContainer }) -eq $null } | Remove-Item -Force -Recurse`

raven schooner
#

Another option may be to look into XCP and utilize the analytics of the tool

brittle sable
#

yeah that's also a good way, since XCP is quite a bit faster than the OS's NFS client (never tried the CIFS version though)

gusty remnant
late arch
#

Want to also call out what you mean by "faster" Async delete will be faster from the client side and dramatically reduce load on the client but we don't actually reclaim the space back any faster.

rose saffron
#

With an azure cvo ha with Page Blobs and high load - sometimes with throttling searching through 5 to 7 mill files in thousand of folders, delete from nfs clients need 1-2 days, Performance via cifs is also slow, currently i Player with tiering and fabricpool hot tier to Split the load on Performance and capacity storage tier which after a General cleanup brought a signifikant Increase of Speer. Never thought that tierisch could bring so much speed