Curious, I guess I don't understand the BSD disclaimer. The application should
not need to track any of this. The OS should be tracking open FD and locks for
the process, and when it closes a FD on behalf of a process it should also
remove the locks.
-Original Message-
>From: "Marvin Hu
I checked, and I don't see that disclaimer in the latest POSIX documentation.
-Original Message-
>From: Robert Engels <[EMAIL PROTECTED]>
>Sent: Jan 24, 2007 9:04 AM
>To: java-dev@lucene.apache.org
>Subject: Re: [jira] Commented: (LUCENE-710) Implement
>
>Curious, I guess I don't understan
On Jan 24, 2007, at 8:04 AM, Robert Engels wrote:
Curious, I guess I don't understand the BSD disclaimer. The
application should not need to track any of this. The OS should be
tracking open FD and locks for the process, and when it closes a FD
on behalf of a process it should also remove
Where do you see that the some JVMs map to fcntl, from what I can see they map
to flock?
-Original Message-
>From: Marvin Humphrey <[EMAIL PROTECTED]>
>Sent: Jan 24, 2007 10:11 AM
>To: java-dev@lucene.apache.org
>Subject: Re: [jira] Commented: (LUCENE-710) Implement
>
>
>On Jan 24, 2007,
Actually, in the FileLock API it states
On some systems, closing a channel releases all locks held by the Java
* virtual machine on the underlying file regardless of whether the locks were
* acquired via that channel or via another channel open on the same file. It
* is strongly recommended t
I took the NFS issue up on the Linux NFS list. Here's the first
suggestion: hard links.
Thoughts?
Marvin Humphrey
Rectangular Research
http://www.rectangular.com/
Begin forwarded message:
From: Trond Myklebust
Date: January 24, 2007 7:06:58 PM PST
To: Marvin Humphrey <[EMAIL PROTECTED]>
C
> > You could easily fix this by having the reader create a hard link
> > to the
> > index file. e.g.
> >
> > ln foo foo-client.my.org-$$
> > open("foo-client.my.org-$$");
> >
> > read()
> > ...
> > close()
> > rm foo-client.my.org-$$
What happens when a client exits prematurely (without remo
