[Engine-devel] PosixFS: GUI mock-ups have been updated

Thu May 10 17:42:15 UTC 2012

> ----- Original Message -----
> From: "Andrew Cathrow" <acathrow at redhat.com>
> Sent: Thursday, May 10, 2012 8:18:58 PM
> 
> ----- Original Message -----
> > From: "Einav Cohen" <ecohen at redhat.com>
> > To: "Andrew Cathrow" <acathrow at redhat.com>
> > Cc: engine-devel at ovirt.org, "Simon Grinberg" <sgrinber at redhat.com>,
> > "Saggi Mizrahi" <smizrahi at redhat.com>, "Geert
> > Jansen" <gjansen at redhat.com>, "Ori Liel" <oliel at redhat.com>, "Yair
> > Zaslavsky" <yzaslavs at redhat.com>, "Ayal Baron"
> > <abaron at redhat.com>
> > Sent: Thursday, May 10, 2012 1:12:06 PM
> > Subject: Re: [Engine-devel] PosixFS: GUI mock-ups have been updated
> > 
> > > ----- Original Message -----
> > > From: "Andrew Cathrow" <acathrow at redhat.com>
> > > Sent: Thursday, May 10, 2012 8:03:16 PM
> > > 
> > > 
> > > ----- Original Message -----
> > > > From: "Einav Cohen" <ecohen at redhat.com>
> > > > To: "Andrew Cathrow" <acathrow at redhat.com>
> > > > Cc: engine-devel at ovirt.org, "Simon Grinberg"
> > > > <sgrinber at redhat.com>,
> > > > "Saggi Mizrahi" <smizrahi at redhat.com>, "Geert
> > > > Jansen" <gjansen at redhat.com>, "Ori Liel" <oliel at redhat.com>,
> > > > "Yair
> > > > Zaslavsky" <yzaslavs at redhat.com>, "Ayal Baron"
> > > > <abaron at redhat.com>
> > > > Sent: Thursday, May 10, 2012 1:01:20 PM
> > > > Subject: Re: [Engine-devel] PosixFS: GUI mock-ups have been
> > > > updated
> > > > 
> > > > > ----- Original Message -----
> > > > > From: "Andrew Cathrow" <acathrow at redhat.com>
> > > > > Sent: Thursday, May 10, 2012 6:06:23 PM
> > > > > 
> > > > > 
> > > > > ----- Original Message -----
> > > > > > From: "Einav Cohen" <ecohen at redhat.com>
> > > > > > To: "Andrew Cathrow" <acathrow at redhat.com>, "Geert Jansen"
> > > > > > <gjansen at redhat.com>, "Ori Liel" <oliel at redhat.com>, "Yair
> > > > > > Zaslavsky" <yzaslavs at redhat.com>, "Ayal Baron"
> > > > > > <abaron at redhat.com>
> > > > > > Cc: engine-devel at ovirt.org, "Simon Grinberg"
> > > > > > <sgrinber at redhat.com>,
> > > > > > "Saggi Mizrahi" <smizrahi at redhat.com>
> > > > > > Sent: Thursday, May 10, 2012 10:56:09 AM
> > > > > > Subject: Re: [Engine-devel] PosixFS: GUI mock-ups have been
> > > > > > updated
> > > > > > 
> > > > > > > ----- Original Message -----
> > > > > > > From: "Andrew Cathrow" <acathrow at redhat.com>
> > > > > > > Sent: Thursday, May 10, 2012 4:57:44 PM
> > > > > > > 
> > > > > > > ----- Original Message -----
> > > > > > > > From: "Einav Cohen" <ecohen at redhat.com>
> > > > > > > > To: "Saggi Mizrahi" <smizrahi at redhat.com>, "Yair
> > > > > > > > Zaslavsky"
> > > > > > > > <yzaslavs at redhat.com>
> > > > > > > > Cc: "Haim Ateya" <hateya at redhat.com>, "Eldan
> > > > > > > > Hildesheim"
> > > > > > > > <info at eldanet.com>, engine-devel at ovirt.org, "Eldan
> > > > > > > > Hildesheim" <ehildesh at redhat.com>, "Simon Grinberg"
> > > > > > > > <sgrinber at redhat.com>
> > > > > > > > Sent: Thursday, May 10, 2012 9:51:32 AM
> > > > > > > > Subject: Re: [Engine-devel] PosixFS: GUI mock-ups have
> > > > > > > > been
> > > > > > > > updated
> > > > > > > > 
> > > > > > > > > ----- Original Message -----
> > > > > > > > > From: "Saggi Mizrahi" <smizrahi at redhat.com>
> > > > > > > > > Sent: Thursday, May 10, 2012 4:39:49 PM
> > > > > > > > > 
> > > > > > > > > I do express that empty mount options SHOULD NOT send
> > > > > > > > > an
> > > > > > > > > empty
> > > > > > > > > string, rather, omit the whole argument.
> > > > > > > > 
> > > > > > > > Yes, this should be handled on the backend side (Yair -
> > > > > > > > please
> > > > > > > > note,
> > > > > > > > maybe it is already implemented like this - don't
> > > > > > > > know):
> > > > > > > > When
> > > > > > > > getting a null-or-empty "mount options" value from the
> > > > > > > > client,
> > > > > > > > the
> > > > > > > > backend needs to make sure to *not* set the relevant
> > > > > > > > parameter
> > > > > > > > in
> > > > > > > > the vdsm verb at all.
> > > > > > > > 
> > > > > > > > So leaving the "mount options" text-box empty in the
> > > > > > > > GUI
> > > > > > > > is
> > > > > > > > legal,
> > > > > > > > only needs to be handled in a certain way in the
> > > > > > > > backend.
> > > > > > > 
> > > > > > > 
> > > > > > > 
> > > > > > > In theory for a PosixFS file system a user could create
> > > > > > > multiple
> > > > > > > storage domains of different PosixFS types. Perhaps
> > > > > > > that's
> > > > > > > not
> > > > > > > a
> > > > > > > problem, but worth noting.
> > > > > > > 
> > > > > > > Is "Path" the correct term to use for the remote mount? I
> > > > > > > can
> > > > > > > imagine
> > > > > > > customers thinking that is local and messing with fstab.
> > > > > > > Not sure if there's a better term - filesystem URI ?
> > > > > > 
> > > > > > - In the initial mock-up, it was called "Mount Spec". Is it
> > > > > > better?
> > > > > 
> > > > > I don't like any of the options - but have a preference for
> > > > > Filesystem URI, but I'd like others to weigh in here.
> > > > > My concern with path is that it could mean local or remote,
> > > > > so
> > > > > another option is "Remote Path"
> > > > 
> > > > But it *can* be local or remote, so why "Remote Path"? "Path"
> > > > actually sounds like a good term.
> > > > 
> > > 
> > > Can it be local - do we want a user mounting a local filesystem?
> > 
> > If it is possible - I don't see why limiting it. It should be
> > similar
> > to defining a "Local on Host" storage domain IIUC. Even if it is
> > potentially harmful for some reason, we don't want to nanny the
> > user, right? He should know what he is doing.
> 
> I'm not saying we should stop them, we should make it clear how this
> is going to be used.
> When we're setting up a storage domain it's shared storage.

I am probably missing something here, but I assume that "Path" isn't a good term (I apologize for not understanding why).
In any case, we need to finalize this term, since it has implications on the rest-api as well.

Andrew/Geert/Simon/Ayal/Miki/Saggi/others: Please feel free to suggest a new term, or vote for one of the previously-discussed terms ("Remote Path" / "Path" / "Mount Spec" / "File System URI").
If no decision will be made here, the term will remain as-is, i.e. "Path".

> 
> 
> 
> > 
> > > > > 
> > > > > 
> > > > > > 
> > > > > > - Note that the current PosixFS implementation in the
> > > > > > rest-api
> > > > > > utilizes the already-existing "<path>" property within the
> > > > > > "<storage>" tag within the "<storage_domain>" rest-api
> > > > > > business
> > > > > > entity, therefore I put in the mockup the same term.
> > > > > > Do you think that the rest-api should have a different term
> > > > > > as
> > > > > > well?
> > > > > > 
> > > > > > > 
> > > > > > > I presume we are doing just not-null validation for path.
> > > > > > > 
> > > > > > > Obviously we can't validate the mount options but how
> > > > > > > good
> > > > > > > is
> > > > > > > the
> > > > > > > error reporting back going to be - if the mount options
> > > > > > > are
> > > > > > > wrong,
> > > > > > > or if something fails with the mount will we see "error
> > > > > > > 12345"
> > > > > > > in
> > > > > > > the UI and require the user to go digging in vdsm logs or
> > > > > > > are
> > > > > > > we
> > > > > > > going to pull back and display toe complete message.
> > > > > > 
> > > > > > Depends on backend/vdsm; Yair/Ayal?
> > > > > > 
> > > > > > > 
> > > > > > > 
> > > > > > > > 
> > > > > > > > > 
> > > > > > > > > ----- Original Message -----
> > > > > > > > > > From: "Einav Cohen" <ecohen at redhat.com>
> > > > > > > > > > To: "Yair Zaslavsky" <yzaslavs at redhat.com>, "Ayal
> > > > > > > > > > Baron"
> > > > > > > > > > <abaron at redhat.com>
> > > > > > > > > > Cc: "Saggi Mizrahi" <smizrahi at redhat.com>, "Andrew
> > > > > > > > > > Cathrow"
> > > > > > > > > > <acathrow at redhat.com>, "Miki Kenneth"
> > > > > > > > > > <mkenneth at redhat.com>, "Simon Grinberg"
> > > > > > > > > > <sgrinber at redhat.com>,
> > > > > > > > > > "Eldan Hildesheim" <ehildesh at redhat.com>, "Eldan
> > > > > > > > > > Hildesheim" <info at eldanet.com>, "Alexey Chub"
> > > > > > > > > > <achub at redhat.com>,
> > > > > > > > > > engine-devel at ovirt.org, "Haim Ateya"
> > > > > > > > > > <hateya at redhat.com>
> > > > > > > > > > Sent: Thursday, May 10, 2012 9:28:31 AM
> > > > > > > > > > Subject: Re: PosixFS: GUI mock-ups have been
> > > > > > > > > > updated
> > > > > > > > > > 
> > > > > > > > > > > ----- Original Message -----
> > > > > > > > > > > From: "Yair Zaslavsky" <yzaslavs at redhat.com>
> > > > > > > > > > > Sent: Thursday, May 10, 2012 4:21:42 PM
> > > > > > > > > > > 
> > > > > > > > > > > On 05/10/2012 04:16 PM, Einav Cohen wrote:
> > > > > > > > > > > > Please review the mock-ups on the feature page:
> > > > > > > > > > > > http://www.ovirt.org/wiki/Features/PosixFSConnection#Changes_in_GUI
> > > > > > > > > > > > 
> > > > > > > > > > > > Comments are welcome.
> > > > > > > > > > > 
> > > > > > > > > > > From talking to Haim I understood that path
> > > > > > > > > > > should
> > > > > > > > > > > include
> > > > > > > > > > > ":"
> > > > > > > > > > 
> > > > > > > > > > From talking to Ayal, the path can be similar in
> > > > > > > > > > its
> > > > > > > > > > format
> > > > > > > > > > to
> > > > > > > > > > a
> > > > > > > > > > path
> > > > > > > > > > provided when creating an NFS storage domain (e.g.
> > > > > > > > > > "server:/dir1/dir2"), *or* to a path provided when
> > > > > > > > > > creating
> > > > > > > > > > a
> > > > > > > > > > Local
> > > > > > > > > > storage domain (e.g. "/tmp/dir3"), meaning, without
> > > > > > > > > > ":".
> > > > > > > > > > @Ayal - any chance for a clarification here?
> > > > > > > > > > 
> > > > > > > > > > > In addition - if we only support V1, why add the
> > > > > > > > > > > combo
> > > > > > > > > > > box?
> > > > > > > > > > 
> > > > > > > > > > We are always showing the combo-box, even if we
> > > > > > > > > > have
> > > > > > > > > > only
> > > > > > > > > > one
> > > > > > > > > > option
> > > > > > > > > > in it (so the user will know what is the value that
> > > > > > > > > > is
> > > > > > > > > > being
> > > > > > > > > > sent).
> > > > > > > > > > However, we disable it. I updated the mock-up to
> > > > > > > > > > clarify
> > > > > > > > > > this.
> > > > > > > > > > 
> > > > > > > > > > > 
> > > > > > > > > > > Thanks!
> > > > > > > > > > > 
> > > > > > > > > > > > 
> > > > > > > > > > > > ----
> > > > > > > > > > > > Thanks,
> > > > > > > > > > > > Einav
> > > > > > > > > > > 
> > > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > _______________________________________________
> > > > > > > > > Engine-devel mailing list
> > > > > > > > > Engine-devel at ovirt.org
> > > > > > > > > http://lists.ovirt.org/mailman/listinfo/engine-devel
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > _______________________________________________
> > > > > > > > Engine-devel mailing list
> > > > > > > > Engine-devel at ovirt.org
> > > > > > > > http://lists.ovirt.org/mailman/listinfo/engine-devel
> > > > > > > > 
> > > > > > > _______________________________________________
> > > > > > > Engine-devel mailing list
> > > > > > > Engine-devel at ovirt.org
> > > > > > > http://lists.ovirt.org/mailman/listinfo/engine-devel
> > > > > > > 
> > > > > > > 
> > > > > > > 
> > > > > > 
> > > > > 
> > > > 
> > > 
> > 
> 