Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 22 Jun 2006 09:59:41 -0500
From:      Eric Anderson <anderson@centtech.com>
To:        freebsd-fs@freebsd.org, pjd@freebsd.org
Subject:   Re: Journaling UFS with gjournal.
Message-ID:  <449AB05D.3030402@centtech.com>
In-Reply-To: <200606221444.k5MEiwHu074696@lurza.secnetix.de>
References:  <200606221444.k5MEiwHu074696@lurza.secnetix.de>

next in thread | previous in thread | raw e-mail | index | archive | help
Oliver Fromme wrote:
> Pawel Jakub Dawidek <pjd@freebsd.org> wrote:
>  > There is no manual page for gjournal(8) yet. There is an option you can
>  > use to specify journal size:
>  > 
>  >         # gjournal label -s 2147483648 ad0
>  > 
>  > This will create 2GB journal and the rest will be for the data.
> 
> Cool.  I guess you can also specify the journal switch
> interval as an option, right?
> 
> Oh, by the way:  Are both UFS1 and UFS2 supported?
> 
>  > +> Another question:  What happens if you (accidentally) make
>  > +> the journals provider too small, so it hits the end before
>  > +> the next regular switch?
>  > 
>  > Currently it will panic, but I'm working on it.
> 
> OK.  Actually I already guessed that it would panic.  :-)
> It probably would be better to block all write requests
> and wait until the other journal has been written out, so
> it can be re-used.  And print(9) a message so the admin
> knows what kind of problem there is.
> 
> However, if the size of the journals provider is twice the
> maximum raw transfer rate of the disk multiplied by the
> switch interval time (as you recommended), it should be
> impossible to get a journal overrun.

What if the journal had to write out little bits across the entire 
journaled area, so the throughput was greatly reduced, while the 
journaled area had massive heavy random reads keeping the disk heads 
very busy?


Eric



-- 
------------------------------------------------------------------------
Eric Anderson        Sr. Systems Administrator        Centaur Technology
Anything that works is better than anything that doesn't.
------------------------------------------------------------------------



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?449AB05D.3030402>