On 10 Feb 2022, at 10:21, Chuck Lever III wrote:
On Feb 10, 2022, at 8:28 AM, Benjamin Coddington
<bcodding@xxxxxxxxxx> wrote:
On 8 Feb 2022, at 17:39, Steve Dickson wrote:
On 2/8/22 4:18 PM, Chuck Lever III wrote:
On Feb 8, 2022, at 2:29 PM, Steve Dickson <steved@xxxxxxxxxx>
wrote:
On 2/8/22 11:21 AM, Chuck Lever III wrote:
On Feb 8, 2022, at 11:04 AM, Steve Dickson <steved@xxxxxxxxxx>
wrote:
Hello,
On 2/4/22 7:56 AM, Benjamin Coddington wrote:
The nfs4id program will either create a new UUID from a random
source or
derive it from /etc/machine-id, else it returns a UUID that has
already
been written to /etc/nfs4-id. This small, lightweight tool is
suitable for
execution by systemd-udev in rules to populate the nfs4 client
uniquifier.
Signed-off-by: Benjamin Coddington <bcodding@xxxxxxxxxx>
---
.gitignore | 1 +
configure.ac | 4 +
tools/Makefile.am | 1 +
tools/nfs4id/Makefile.am | 8 ++
tools/nfs4id/nfs4id.c | 184
+++++++++++++++++++++++++++++++++++++++
tools/nfs4id/nfs4id.man | 29 ++++++
6 files changed, 227 insertions(+)
create mode 100644 tools/nfs4id/Makefile.am
create mode 100644 tools/nfs4id/nfs4id.c
create mode 100644 tools/nfs4id/nfs4id.man
Just a nit... naming convention... In the past
we never put the protocol version in the name.
Do a ls tools and utils directory and you
see what I mean....
Would it be a problem to change the name from
nfs4id to nfsid?
nfs4id is pretty generic, too.
Can we go with nfs-client-id ?
I'm never been big with putting '-'
in command names... nfscltid would
be better IMHO... if we actually
need the 'clt' in the name.
We have nfsidmap already. IMO we need some distinction
with user ID mapping tools... and some day we might
want to manage server IDs too (see EXCHANGE_ID).
Hmm... So we could not use the same tool to do
both the server and client, via flags?
nfsclientid then?
You like to type more than I do... You always have... :-)
But like I started the conversation... the naming is
a nit... but I would like to see one tool to set the
ids for both the server and client... how about
nfsid -s and nfsid -c
The tricky thing here is that this little binary isn't going to set
anything, and we probably never want people to run it from the
command line.
A 'nfsid -s' and 'nfsid -c' seem to want to do much more. I feel
they are
out of scope for the problem I'm trying to solve: I need something
that
will generate a unique value, and persist it, suitable for execution
in a
udevd rule.
Perhaps we can stop worrying so much about the name of this as I
don't think
it should be a first-class nfs-utils command, rather just a helper
for udev.
And maybe the name can reflect that - "nfsuuid" ?
The client ID can be an arbitrary string, so I think not.
I feel like we might all be missing the fact that this tool doesn't
create
client IDs. The tool only creates uuids, and returns what may have
already
been set by something somewhere else. It's not supposed to ever get
typed
out or run by people. Any other suggestions? Here's where we are:
nfs4id - no: we dislike the number 4
nfsuuid - no: it doesn't have to be a uuid
nfsid - no: too ambiguous
nfscltid - no: also too ambiguous
nfsclientid - no: too much typing
Since I've already re-written it, I'm going to send it again as nfsuuid
-
and let's bikeshed on it again over there, and see if we can make
suggestions that might make everyone happy.
Ben
