==> Regarding Re: [autofs] automount never umount; Farkas Levente <[EMAIL 
PROTECTED]> adds:

lfarkas> Ian Kent wrote:
>> That wasn't what I had in mind.
>> 
>> I was thinking more along the lines of the definition of what constitues
>> "use". For example any access, such as a directory listing, will update
>> the "last used" value whereas if I change the definition of "last used"
>> to "currently in use" then to be busy there would have to be an open
>> file or a process working directory set. This may solve the GUI scanning
>> problem we see but wouldn't solve immediate remounting that occurs due
>> to the scanning as well.

lfarkas> that would be useful and may be solve our problem. without it the
lfarkas> current implementation is not usable to umount. think about just
lfarkas> the daily process like slocate database update, tripwire and may
lfarkas> be others.

Things that run infrequently are not at fault, here.  I'd really like to
know precisely what is going on, in your case.

Ian, I think it would be easy enough to instrument the autofs4 module.
Simple things, like figuring out which process is triggerring the lookup.
We can either log it, or just keep a ring buffer.

I don't think I'll have time to put together any code this week, but may
next week.

-Jeff

_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs

Reply via email to