From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | jd(at)commandprompt(dot)com |
Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Kevin Grittner <Kevin(dot)Grittner(at)wicourts(dot)gov>, Markus Wanner <markus(at)bluegap(dot)ch>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: determine snapshot after obtaining locks for first statement |
Date: | 2009-12-17 18:13:05 |
Message-ID: | 4765.1261073585@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
"Joshua D. Drake" <jd(at)commandprompt(dot)com> writes:
> What is needed here is a layman's context of what isolation modes are
> good for what type of operation. Neither your explanation or Tom's is
> particularly useful except to say, "Crap, I might be screwed but I don't
> know if I am... how do I find out?"
If we had a simple way to characterize that, we'd not be having this
discussion :-(
One possibility is to try to list the risky cases. So far I can think
of:
* updates using a WHERE clause that tests columns being changed by other
transactions
* updates using subqueries/joins so that the result depends on other rows
besides the one directly updated/deleted, and those other rows are
subject to concurrent changes
But I'm not sure this is a complete list, and an incomplete one might do
more harm than good ...
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Kevin Grittner | 2009-12-17 18:16:57 | Re: determine snapshot after obtaining locks for first statement |
Previous Message | Kevin Grittner | 2009-12-17 18:12:02 | Re: determine snapshot after obtaining locks for first statement |