[prev in list] [next in list] [prev in thread] [next in thread] 

List:       jfs-discussion
Subject:    Re: [Jfs-discussion] nointegrity
From:       pg_jfs () jfs ! for ! sabi ! co ! UK (Peter Grandi)
Date:       2007-08-30 18:46:56
Message-ID: 18135.4256.122955.20382 () base ! ty ! sabi ! co ! UK
[Download RAW message or body]

>>> On Wed, 29 Aug 2007 17:05:32 -0400, "Justin Bronder"
>>> <jsbronder@gmail.com> said:

[ ... ]

jsbronder> By "lose anything" I meant to suggest that assuming
jsbronder> the data has been written to disk, will JFS have any
jsbronder> problems later on.

Fine, but in general it is not easy to know when the data has
been fully written to disk, unless special precautions are taken.

jsbronder> I'm not overly worried about the metadata for, as
jsbronder> mentioned, the filesystem only gets one large file at
jsbronder> a time, which is then moved off later.

Good, but the metadata includes the pointers to where the data
(and even metadata) extents are. If a file contains many such
extents there can be quite a bit of metadata ''in flight''. In
your case probably this is not the case.

jsbronder> [ ... ] So if I'm getting a performance increase, and
jsbronder> my only worry is a power loss before the data exists on
jsbronder> the hard drive, then I'm pretty happy using nointegrity.

That seems the case, but again I feel like re-restating that
"before the data exists on the hard drive" is a fuzzy concept
unless one is careful. I'd consider using 'nointegrity' with
'fsync' just before 'close'.

[ ... ]

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jfs-discussion mailing list
Jfs-discussion@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jfs-discussion
[prev in list] [next in list] [prev in thread] [next in thread]