Thread: 8.5 TODO: any info on "Create dump tool for write-ahead logs..." in PITR section (1.4)?

Can anyone elaborate (or point me to some additional info) on the 8.5
TODO item in the "Point-In-Time Recover (PITR) section (1.4)": Create dump tool for write-ahead logs for use in
determining
transaction id for point-in-time recovery    This is useful for checking PITR recovery.

I poked around a bit and found some code that "walks" the WAL logs (in
src/backend/access/transam/xlog.c, I think) and I could probably
figure out how to display a WAL log file contents, but I'm hoping
someone can provide some context as to what issue this TODO item is
trying to address (i.e., what output would be useful).

Hi,

On Sat, Oct 3, 2009 at 7:30 AM, shakahshakah@gmail.com
<shakahshakah@gmail.com> wrote:
> Can anyone elaborate (or point me to some additional info) on the 8.5
> TODO item in the "Point-In-Time Recover (PITR) section (1.4)":
>  Create dump tool for write-ahead logs for use in determining
> transaction id for point-in-time recovery
>     This is useful for checking PITR recovery.
>
> I poked around a bit and found some code that "walks" the WAL logs (in
> src/backend/access/transam/xlog.c, I think) and I could probably
> figure out how to display a WAL log file contents, but I'm hoping
> someone can provide some context as to what issue this TODO item is
> trying to address (i.e., what output would be useful).

I think that xlogdump (http://xlogviewer.projects.postgresql.org/) is
the first step to address that TODO item. Though I'm not sure if the
xlogdump project is still active.

Regards,

--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center

shakahshakah@gmail.com wrote:
> Can anyone elaborate (or point me to some additional info) on the 8.5
> TODO item in the "Point-In-Time Recover (PITR) section (1.4)":
>   Create dump tool for write-ahead logs for use in determining
> transaction id for point-in-time recovery
>      This is useful for checking PITR recovery.
> 
> I poked around a bit and found some code that "walks" the WAL logs (in
> src/backend/access/transam/xlog.c, I think) and I could probably
> figure out how to display a WAL log file contents, but I'm hoping
> someone can provide some context as to what issue this TODO item is
> trying to address (i.e., what output would be useful).

That TODO item is a lot less important after we have Hot Standby. It
contains functions that allow you to pause and continue WAL replay, and
step through the WAL one transaction at a time. It won't let you go
backwards, though, so it won't completely replace a tool like xlogdump,
which you can use to look into the WAL before applying it.

--  Heikki Linnakangas EnterpriseDB   http://www.enterprisedb.com

Hi,

On Tue, Oct 6, 2009 at 7:47 AM, Heikki Linnakangas
<heikki.linnakangas@enterprisedb.com> wrote:
> That TODO item is a lot less important after we have Hot Standby. It
> contains functions that allow you to pause and continue WAL replay, and
> step through the WAL one transaction at a time.

I don't think this is practical for PITR purpose, since applying one transaction
at a time up to the recovery target point might take very long time. So the tool
to quickly determine the recovery target point is necessary even after we have
HS, I think. Am I missing something?

Regards,

-- 
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center

Fujii Masao escreveu:
> On Tue, Oct 6, 2009 at 7:47 AM, Heikki Linnakangas
> <heikki.linnakangas@enterprisedb.com> wrote:
>> That TODO item is a lot less important after we have Hot Standby. It
>> contains functions that allow you to pause and continue WAL replay, and
>> step through the WAL one transaction at a time.
> 
> I don't think this is practical for PITR purpose, since applying one transaction
> at a time up to the recovery target point might take very long time. So the tool
> to quickly determine the recovery target point is necessary even after we have
> HS, I think. Am I missing something?
> 
No. BTW, I have an unpublished version here that I didn't commit because I
need to clean it up. :(


--  Euler Taveira de Oliveira http://www.timbira.com/