RE: Are journal's UB & UP entries guaranteed to be in 1-2 sequenc e?

[Vern Hamberg <vhamberg@xxxxxxxxxxxxxxxxxxxxxxxxx>]

Walden, I don't think I spoke clearly enough. As I understand it, all 
journal entries for a row update are written before anything is written to 
disk. This is probably an atomic operation - I hope so - but I'd need to 
check with my source, who was part of the journal team some years ago. 
There is a very small window of vulnerability, but both journal entries 
would be written before any actual change to the physical file itself.

The reason for writing to the journal first is to minimize loss to the 
actual physical data, AFAIK.

Nonetheless, if Al spoke to someone on the inside and was told that both 
entries will be there, I'll take his word on that.

Vern

At 01:33 PM 9/5/2003 -0400, you wrote:
> Jim,
>
> True. The point of the last post was simply that the logic didn't hold.
>
> -Walden
>
> -----Original Message-----
> From: Jim Franz [mailto:franz400@xxxxxxxxxxxx]
> Sent: Friday, September 05, 2003 12:21 PM
> To: Midrange Systems Technical Discussion
> Subject: Re: Are journal's UB & UP entries guaranteed to be in 1-2 sequenc
> e?
>
> all they would need is to get two journal sequence numbers within
> a single lock. what would be hard about that?
> jim
>
> ----- Original Message -----
> From: "Walden H. Leverich III" <WaldenL@xxxxxxxxxxxxxxx>
>
> > >journal entries are written before the write/update to the file
> > >- this assures that things can be rebuilt from journal in the
> > >event of failure. So this would support Al's information, I
> > >think.
> >
> > Nope, I have to disagree that this supports Al's information. Think of it
> > this way, I have to deposit money into a bank before I withdraw it[1], and
> > so to you. That is:
> >
> > Walden - Deposit
> > Walden - Withdraw
> > Vern - Deposit
> > Vern - Withdraw
> >
> > It can never be:
> > Walden - Withdraw
> > Walden - Deposit
> >
> >
> > However, it can be
> >
> > Walden - Deposit
> > Vern - Deposit
> > Vern - Withdraw
> > Walden - Withdraw
> >
> > And I believe this is the exact scenerio that the original poster was
> > worried about.
> >
> > -Walden
> >
> > [1] - I'm ignoring overdraft protection, monies already in the account,
> > margin accounts, etc. You get the point.
> >
> > ------------
> > Walden H Leverich III
> > President
> > Tech Software
> > (516) 627-3800 x11
> > (208) 692-3308 eFax
> > WaldenL@xxxxxxxxxxxxxxx
> > http://www.TechSoftInc.com
> >
> > Quiquid latine dictum sit altum viditur.
> > (Whatever is said in Latin seems profound.)
> >
> >
> > -----Original Message-----
> > From: Vern Hamberg [mailto:vhamberg@xxxxxxxxxxxxxxxxxxxxxxxxx]
> > Sent: Friday, September 05, 2003 11:28 AM
> > To: Midrange Systems Technical Discussion
> > Subject: Re: Are journal's UB & UP entries guaranteed to be in 1-2
> sequence?
> >
> >
> > We were discussing journals yesterday - journal entries are written before
> > the write/update to the file - this assures that things can be rebuilt
> from
> > journal in the event of failure. So this would support Al's information, I
> > think.
> >
> > Vern
> >
> > At 10:35 AM 9/5/2003 -0400, you wrote:
> > >I asked, and was given the assurance that it was so.
> > >
> > >Al
> > >
> > >Even though you are Al <g>, is that documented anywhere?
> > >
> > >GA
> >
> >
> > _______________________________________________
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
> list
> > To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > To subscribe, unsubscribe, or change list options,
> > visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > Before posting, please take a moment to review the archives
> > at http://archive.midrange.com/midrange-l.
> > _______________________________________________
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
> list
> > To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > To subscribe, unsubscribe, or change list options,
> > visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > Before posting, please take a moment to review the archives
> > at http://archive.midrange.com/midrange-l.
> >
> >
>
>
> _______________________________________________
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> or email: MIDRANGE-L-request@xxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/midrange-l.
> _______________________________________________
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> or email: MIDRANGE-L-request@xxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/midrange-l.