[patchset] quota fixes, cleanups and sysquotas for AIX
uri at samba.org
Thu Feb 14 12:03:11 UTC 2019
On 2/14/19 1:36 PM, Björn JACKE wrote:
> On 2019-02-14 at 01:57 +0100 Björn JACKE via samba-technical sent off:
>> On 2019-02-13 at 23:24 +0200 Uri Simchoni via samba-technical sent off:
>>> 6/13 - A consequence of this change is that setting the
>>> storage-quota/inode-limit ratio depends on compile-time platform, which
>>> is probably not desirable (see vfs_set_ntquota() and how inode limit is
>>> calculated). Perhaps the ratio should have its own constant (I'm not
>>> sure what's the rationale of allowing 2K per inode in the original code).
>> I think the we should remove this artifical inode ratio limit entirely. Windows
>> has no inode quota limit settings here and we should not introduce such limits
>> out of nothing.
> to address this point I propose attached patch, review appreciated again.
That looks reasonable. We should give people some time to respond
though, and maybe also mention that in WHATSNEW.txt.
I wonder whether the patch would remove the inode limit or set it to
zero (no new inodes allowed). For example in the Linux driver, we always
set the dqb_valid to QIF_LIMITS, which is (QIF_BLIMITS | QIF_ILIMITS),
so if I understand the docs, it means "set the inode quota to zero
inodes". Haven't tested though...
More information about the samba-technical