Emmanuel, Thanks for the explanation. I am using sequoia-2.10.9 build in which I had reported the problem. This seems to be the latest build. But you suggested this problem is fixed in the latest build.
Am I missing something here ? I am looking for the builds at https://forge.continuent.org/frs/?group_id=6. thanks & regards, rathi On Tue, Apr 22, 2008 at 5:32 AM, Emmanuel Cecchet <[EMAIL PROTECTED]> wrote: > Rathi, > > Thanks for the response. But I had a doubt. Why do we need add an entry > > for BEGIN statement. We have the transaction id which should be enough for > > starting the transaction. > > > Actually, when a checkpoint is inserted in the recovery log some > transactions might have to finish. When we replay the log, we don't replay > the queries in transactions for which we have not replayed the begin > (because these queries were played before the backend was disabled). > Therefore we need the begin marker to know which transactions have started > before or after a checkpoint. > > Thanks for your interest in Sequoia, > Emmanuel > > > > > On Thu, Apr 17, 2008 at 10:38 PM, Emmanuel Cecchet > > <[EMAIL PROTECTED]> wrote: > > > > Rathi, > > > > This was due to > > https://forge.continuent.org/jira/browse/SEQUOIA-1063 that has > > just been recently fixed. > > Using the latest version from the 2.10 branch should solve the issue. > > > > Best regards, > > Emmanuel > > > > I am sorry I ment backends instead of VirtualDatabases. > > My setup is one controller with one virtual database (myDB) > > and two backends (locahhost and localhost2). > > > > My commands for disabling backend > > > > admin myDB > > admin > > expert on > > initialize localhost > > disable localhost > > quit > > > > > > My commands for enabling backend > > > > admin myDB > > admin > > expert on > > initialize localhost > > enable localhost > > quit > > > > My commands for inserting data > > > > sql client jdbc:sequoia://localhost:25322,localhost:25323/myDB > > user > > begin > > insert into demo values (1) > > commit > > insert into demo values (2) > > quit > > > > This is my setup if you need anyother details please let me know. > > > > Would like to understand this comment in the RecoverThread code > > /* > > * if the task transaction id does not have a corresponding begin > > * (it is not in the tids list), then this task has already been > > * played when the backend was disabled. So we can skip it. > > */ > > > > > > thanks & regards, > > Ritesh Rathi > > > > On Thu, Apr 17, 2008 at 9:38 PM, Emmanuel Cecchet > > <[EMAIL PROTECTED]> wrote: > > > > Hi, > > > > What is your setup? How many controllers? How many backends? > > What do you call enable/disable virtual database? These > > operations > > are only performed on backends. > > Please provide a detailed scenario with actual command > > consoles > > and configuration to reproduce the issue. > > > > Thanks for your feedback, > > Emmanuel > > > > I do the following steps > > 1. Disable a Virtual Database. > > 2. Insert some entries with begin and commit. > > 3. Entries are added to the enabled virtual database. > > 4. Enable the disabled virtual database. > > 5. It tries to recover. All the entries which were added > > in > > auto commit mode are added to the enabled virtual > > database but > > not the transacted inserts. > > 6. While analyzing saw the below comment in the > > RecoverThread > > code which results in this behavior. > > > > /* > > * if the task transaction id does not have a > > corresponding begin > > * (it is not in the tids list), then this task has > > already been > > * played when the backend was disabled. So we can skip > > it. > > */ > > > > Can someone please explain this behavior. > > > > thanks & regards, > > Ritesh Rathi > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Sequoia mailing list > > [email protected] > > <mailto:[email protected]> > > <mailto:[email protected] > > <mailto:[email protected]>> > > > > https://forge.continuent.org/mailman/listinfo/sequoia > > > > > > > > > > -- Emmanuel Cecchet - Research scientist > > EPFL - LABOS/DSLAB - IN.N 317 > > Phone: +41-21-693-7558 > > > > _______________________________________________ > > Sequoia mailing list > > [email protected] > > <mailto:[email protected]> > > https://forge.continuent.org/mailman/listinfo/sequoia > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Sequoia mailing list > > [email protected] > > https://forge.continuent.org/mailman/listinfo/sequoia > > > > > -- > Emmanuel Cecchet - Research scientist > EPFL - LABOS/DSLAB - IN.N 317 > Phone: +41-21-693-7558 > > _______________________________________________ > Sequoia mailing list > [email protected] > https://forge.continuent.org/mailman/listinfo/sequoia >
_______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia
