| From: | Joachim Wieland <joe(at)mcknight(dot)de> |
|---|---|
| To: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
| Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Snapshot synchronization, again... |
| Date: | 2010-12-31 14:38:51 |
| Message-ID: | AANLkTiktRBU5M6=Zgfmodn=ZX5VkZm=K2+_TdLxhYkC0@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Fri, Dec 31, 2010 at 8:28 AM, Alvaro Herrera
<alvherre(at)commandprompt(dot)com> wrote:
> A backend can have any number of snapshots registered, and those don't
> allow GlobalXmin to advance. Consider an open cursor, for example.
> Even if the rest of the transaction is read committed, the snapshot
> registered by the open cursor still holds back GlobalXmin. My
> (handwavy) idea is that whenever the transaction calls
> pg_publish_snapshot(), said snapshot is registered, which makes it safe
> to use even if the transaction continues to operate and obtain newer
> snapshots.
Cool, even better that this is taken care of already :-)
Thanks,
Joachim
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Simon Riggs | 2010-12-31 16:56:38 | Re: Sync Rep Design |
| Previous Message | Magnus Hagander | 2010-12-31 14:38:29 | Re: Streaming replication as a separate permissions |