| From: | Greg Stark <stark(at)enterprisedb(dot)com> |
|---|---|
| To: | Robert Haas <robertmhaas(at)gmail(dot)com> |
| Cc: | Alvaro Herrera <alvherre(at)commandprompt(dot)com>, Andrew Dunstan <andrew(at)dunslane(dot)net>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Pavel Golub <pavel(at)gf(dot)microolap(dot)com>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Named transaction |
| Date: | 2009-06-17 22:26:00 |
| Message-ID: | 407d949e0906171526u56517109v7e548dbe2161cd96@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Jun 17, 2009 at 8:09 PM, Robert Haas<robertmhaas(at)gmail(dot)com> wrote:
>
>> I have no idea what they are in Firebird but the name conjured up a
>> different (interesting) idea for me. I had the image of naming a
>> transaction and then being able to have other sessions join that same
>> transaction. We've discussed this before for connection-pooled systems
>> which want to be able to return their connection to the pool in the
>> middle of their transaction. It would also possibly be useful for
>> parallel data dumps and loads.
>
> At the risk of veering off-topic, wouldn't this present some awfully
> nasty issues vis-a-vis the command counter?
I didn't say it would be easy :)
I think the command counter might be ok (though I'm not sure we could
pull off the same "phantom cid" trick we do now). But locking and all
the per-transaction information stored in the pgproc info would be a
problem. Basically anywhere in the code where we used "session" as a
proxy for "transaction"...
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Greg Stark | 2009-06-17 22:32:40 | Re: Determining client_encoding from client locale |
| Previous Message | Andrew Dunstan | 2009-06-17 21:16:24 | Re: Named transaction |