| From: | "David G(dot) Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com> |
|---|---|
| To: | "jacktby(at)gmail(dot)com" <jacktby(at)gmail(dot)com> |
| Cc: | Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com>, pgsql-hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: Re: Give me more details of some bits in infomask!! |
| Date: | 2023-02-26 15:44:45 |
| Message-ID: | CAKFQuwYO0FmBS7TDTXbaTxHD9D+Jh6BN63oURAnpt6XziLsy9Q@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Sun, Feb 26, 2023 at 8:36 AM jacktby(at)gmail(dot)com <jacktby(at)gmail(dot)com> wrote:
> > CID means "command ID" i.e. sequential ID assigned to commands in a
> > single session (for visibility checks, so that a query doesn't see data
> > deleted by earlier commands in the same session). See
> > src/backend/utils/time/combocid.c for basic explanation of what "combo
> > CID" is.
> I think if cid is used for visibility checks in one session, that's
> meaingless, beacause we can use the t_xmin and t_xmax to
> get this goal. Is tis
>
>
I think the word "session" is wrong. It should be "transaction".
IIUC, it is what is changed when one issues CommandCounterIncrement within
a transaction. And you need somewhere to save which CCI step deletes rows,
in particular due to the savepoint feature.
David J.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Melanie Plageman | 2023-02-26 16:03:43 | Track IO times in pg_stat_io |
| Previous Message | David G. Johnston | 2023-02-26 15:40:18 | Re: Re: Why the lp_len is 28 not 32? |