Re: About TLF... quick ideas
* Tomi Manninen OH2BNS <tpm@prkele.tky.hut.fi> [2002 Apr 23 07:17 -0500]:
> No, AFAIK journaling only makes recovery from a crash quicker by making
> sure that the disk file system can never enter an inconsistent state. That
> is not the same as not loosing data.
Thanks for that tip, Tomi.
I keep hearing of journalling as a panecea and am willing to try it
myself.
> But there are standard ways of making sure the data is written to the
> disk: fsync(), sync() etc. to force a flush or open() with O_SYNC to make
> the opened file such that all writes wait until the data is written to
> disk.
This crossed my mind later in the day. I figured the logging program
should flush the buffer to disk immediately after each log entry to aid
in data safety.
CT manages quick duping and call checking due to the fact it maintains
the log data in RAM as well as on disk. Now, I would say it's debatable
whether that is necessary in a Linux system as it maintains its own disk
cache. Most often the dupe check would probably hit the cache anyway,
but...
I guess that's down the road.
I'm going to keep tabs on this project and help out if I can.
73, de Nate >>
--
Wireless | Amateur Radio Station N0NB | "We have awakened a
Internet | n0nb@networksplus.net | sleeping giant and
Location | Bremen, Kansas USA EM19ov | have instilled in him
Amateur radio exams; ham radio; Linux info @ | a terrible resolve".
http://www.qsl.net/n0nb/ | - Admiral Yamomoto
--
To UNSUBSCRIBE, email to debian-hams-request@lists.debian.org
with a subject of "unsubscribe". Troble? Contact listmaster@lists.debian.org
Reply to: