Re: trying again to get incremental backup

From: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
To: Robert Haas <robertmhaas(at)gmail(dot)com>
Cc: Dilip Kumar <dilipbalaut(at)gmail(dot)com>, Andres Freund <andres(at)anarazel(dot)de>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: trying again to get incremental backup
Date: 2023-11-20 19:03:01
Message-ID: 202311201903.j3r4qidaxtkc@alvherre.pgsql
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2023-Nov-16, Robert Haas wrote:

> On Thu, Nov 16, 2023 at 12:23 PM Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:

> > I don't understand this point. Currently, the protocol is that
> > UPLOAD_MANIFEST is used to send the manifest prior to requesting the
> > backup. You seem to be saying that you're thinking of removing support
> > for UPLOAD_MANIFEST and instead just give the LSN as an option to the
> > BASE_BACKUP command?
>
> I don't think I'd want to do exactly that, because then you could only
> send one LSN, and I do think we want to send a set of LSN ranges with
> the corresponding TLI for each. I was thinking about dumping
> UPLOAD_MANIFEST and instead having a command like:
>
> INCREMENTAL_WAL_RANGE 1 2/462AC48 2/462C698
>
> The client would execute this command one or more times before
> starting an incremental backup.

That sounds good to me. Not having to parse the manifest server-side
sounds like a win, as does saving the transfer, for the cases where the
manifest is large.

Is this meant to support multiple timelines each with non-overlapping
adjacent ranges, rather than multiple non-adjacent ranges?

Do I have it right that you want to rewrite this bit before considering
this ready to commit?

--
Álvaro Herrera Breisgau, Deutschland — https://www.EnterpriseDB.com/
"No nos atrevemos a muchas cosas porque son difíciles,
pero son difíciles porque no nos atrevemos a hacerlas" (Séneca)

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2023-11-20 19:03:28 Re: Use of backup_label not noted in log
Previous Message Jubilee Young 2023-11-20 18:50:36 Re: Hide exposed impl detail of wchar.c