Vault reuse

[BenLoosli]

We have our vaults set to 25K while the default is 50K and Microsoft says the 50K per folder is an okay limit.
If we go back and change the limit to 50K and make the existing vaults writeable again, will Windchill start putting the files in vault 1 or continue with the last folder used and then move to the next new folder?

[avillanueva]

I think if the folders are auto-managed the order is greyed out so it will likely just write to the same one. If you want to reset things, you can create a new vault, change mapping rules to point to it and then let the system revault and it should move all the content the new vault, creating new folders under the 50k limit rule. Eventually, all the old folders and vault will have unreferenced files and can be unmounted. 

I had manually managed vaults a while back and used this process to shift over to a vault with automatic folder creation.

[mmeadows-3]

At the end of the day it doesn't matter because the total number of files on disk won't be impacted.  But my obsessive compulsive kicks in and I go back and set the folders to read/write again at the DB level.  As they fill up, they will be set back to read-only again.  It follows the vault folders sequence order.

update FvFolder set readOnly=0;

Doing a new vault frequently leads to cleanup frustration at the end of the revaulting process.  There are a few files that just won't move or prove challenging to clean up.  It is simpler to toggle the read-only switch and let the existing folders auto-fill again.

FYI: Not all hardware is equal.  Use caution upping the limit on poorly performing hardware.  50k is fine in most environments today.

[RandyJones]

---

@mmeadows-3 wrote:

FYI: Not all hardware is equal.  Use caution upping the limit on poorly performing hardware.  50k is fine in most environments today.

---

We use 100k here with RHEL and zfs.

[avillanueva]

I feel like I should get a prize. 1M here.

[mmeadows-3]

Sorry, I was talking Windows.  Linux is the clear winner in this category.  Windows has slowly increased to 50k files per folder without a serious performance impact while Unix/Linux has been 100k+ per folder for a very long time.

Without Microsoft's MFT concept, is there any files per folder upper limit on Linux?

[RandyJones]

---

@mmeadows-3 wrote:

Without Microsoft's MFT concept, is there any files per folder upper limit on Linux?

---

According to wikipedia zfs limit of files per directory is: 2⁴⁸: number of entries in any individual directory

https://en.wikipedia.org/wiki/ZFS#:~:text=Some%20theoretical%20limits%20in%20ZFS,maximum%20size%20of%20any%20attribute