Re: Proposal: Select ... AS OF Savepoint - Mailing list pgsql-hackers

From Gokulakannan Somasundaram
Subject Re: Proposal: Select ... AS OF Savepoint
Date
Msg-id 9362e74e0711020922t59da63ffra5f71776ceb21c03@mail.gmail.com
Whole thread Raw
In response to Re: Proposal: Select ... AS OF Savepoint  (Hans-Juergen Schoenig <postgres@cybertec.at>)
Responses Re: Proposal: Select ... AS OF Savepoint  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers


On 11/2/07, Hans-Juergen Schoenig <postgres@cybertec.at> wrote:

I think Simon Riggs is already working on that idea. This one is fairly easy to implement. I think these are some of the features only a time-stamp based database can implement. I think database standards were formed during the time, when the data consistency was provided with Lock based mechanisms. And moreover i have already committed on the indexes with snapshot and i am still waiting for its approval from hackers. If that does go through, then i need to work on the reverse mapping hash tables, which is really a long task. So i may not be able to take  up  time-travel now.


if i remember my last talk with Simon correctly the idea is to have timetravel across transactions.
having this feature inside a transaction will not make it into CVS as it is basically of no practical use.

I am just reminding my days of working with Oracle. The Flashback feature was allowed only for DBAs, and they never let the developers access that feature, unless there is a strong reason for it. It was more thought of as a feature for recovery and they never let deveopers use that in the application. Also it was designed as a optional feature. If its switched off, it cannot be used. If someone comes up with the time travel feature across transactions and if it is designed as non-optional feature and if it happens to be a feature, which DBA can let the developers use freely, then this feature should be rolled back. The feature i am talking about is very simple and it won't even add 100 lines of code into the Postgres source code base.

i would suggest to put some effort into making it work across transactions. just saving the snapshot is not enough
here - there are a couple of other things which have to be taken into consideration (transaction wraparound, etc.)

When i think about it, Timetravel always look big for me and i don't have the bandwidth to take that up.
 

if you want to work on timetravel my team and i can provide some assistance as we wanted to help in this area anyway.

Thanks. Please send me your findings.




--
Thanks,
Gokul.
CertoSQL Project,
Allied Solution Group.
(www.alliedgroups.com)

pgsql-hackers by date:

Previous
From: Heikki Linnakangas
Date:
Subject: Re: Clarification about HOT
Next
From: "Joshua D. Drake"
Date:
Subject: Re: Intel x64 vs AMD x64 pgdata