[6107] in Release_7.7_team
Re: Reworking liblocker
daemon@ATHENA.MIT.EDU (Jacob Morzinski)
Wed Dec 3 21:42:18 2008
To: release-team@mit.edu, athena10@mit.edu
From: Jacob Morzinski <morzinski@MIT.EDU>
Date: Wed, 03 Dec 2008 21:41:31 -0500
In-Reply-To: <Release_7.7_team:6106@unknown-discuss-server> (Ken Raeburn's
message of "Wed, 3 Dec 2008 17:57:57 -0500")
Message-ID: <w6my6yw8xwk.fsf@horobi.mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Spam-Flag: NO
X-Spam-Score: 0.00
The question of whether lockers universally attach at
"/mit/lockername" piqued my interest.
I did a cursory crawl through Moira. Excluding IMAP filesystems,
I found about 275 names which do not attach at "/mit/lockername".
A lot of them attach on "/os", "/srvd", or "/install". Removing
anything that attaches in those three places reduces the list
down to about 120 filesystems.
Some of the filesys entries do attach NFS mounts at non-"mit"
locations ("achilles:u1" will attach at "/achilles/u1").
Some of the entries are aliases of some form. x11r{2,3,4} all
attach at "/mit/x11", "oracle8" attaches at "/mit/oracle", etc.
Some of the aliases are subtle, and I'd need to think carefully
about breaking them ("saltzer" is an alias that will attach at
"/mit/Saltzer" - this is the only capitalized username in the
Athena system. Similarly, a capitalized student group has an
alias for a lowercase version of their name.)
Some of the filesys entries are interesting ideas that didn't catch on
("xess_rw" will attach "/afs/.athena/software/xess" to "/mit/xess")
Some of the entries are for people who have explicit pointers to
personal home directories in various afs cells: a filesys entry
named username-ath or username-net or username-sipb, each
referring to a directory in the indicated cell, each attaching at
/mit/username, usually linked into an FSGROUP.
No-one has locker names with "/" in them, but I'll echo Ken's
statement that restricting a lockername to mount at /mit/lockername
"is a functional change that should be considered carefully."
--
Jacob Morzinski jmorzins@mit.edu
Ken Raeburn <raeburn@MIT.EDU> writes:
> On Dec 3, 2008, at 16:53, Evan Broder wrote:
>> I think I'm in favor of forcing lockername to mount at /mit/lockername
>> always, but I could possibly be swayed.
>
> It's probably reasonable, but it is a functional change that
> should be considered carefully.
>
> I hope no one has created locker names with "/" in them.
[...snip...]
> I got those through the moira client and hesiod and knowing some
> of our history, but I don't have the ability to check the
> database for other cases.
>
> Anything of type MUL (like athena-sun4sys) probably won't be
> handled well by an automounter. I don't know if effectively
> desupporting that type is likely to be a real problem.
