Comment 3 for bug 527899

Yes, that needs to be added as well.

On Feb 25, 2010, at 7:55 PM, Martin Scholl wrote:

> Hello Paul,
>
>
> there is already a commit where I simply call
> xn_xres_run_recovery_thread in the pbxt initialization function; that
> looks like an unclean solution in the light of your proposal.
>
> The recovery thread in xn_xres_run_recovery_thread waits for the
> sweeper to terminate before returning. Should I add this too?
>
>
> Martin
>
>
> On Thu, Feb 25, 2010 at 7:29 PM, Paul McCullagh
> <email address hidden> wrote:
>> Hi Martin,
>>
>> There is probably no need to do async recovery in EPBXT.
>>
>> Just call the following:
>>
>> xt_open_database(self, mysql_real_data_home, TRUE);
>> pbxt_database = self->st_database;
>> xt_heap_reference(self, pbxt_database);
>>
>> instead of xt_xres_start_database_recovery().
>>
>> The only reason this is done asynchronously is to work around
>> problems
>> in MySQL. Namely: the MySQL plugin environment is not completely
>> initialized and creation of a THD can fail.
>>
>> --
>> TODO wait for database recovery on EPBXT initialization
>> https://bugs.launchpad.net/bugs/527899
>> You received this bug notification because you are a direct
>> subscriber
>> of the bug.
>>
>
> --
> TODO wait for database recovery on EPBXT initialization
> https://bugs.launchpad.net/bugs/527899
> You received this bug notification because you are subscribed to
> embedded PBXT.
>
> Status in Embedded PBXT (EPBXT): In Progress
>
> Bug description:
> Do a blocking wait for database recovery upon EPBXT initialization
> in EPBXT::init_epbxt()
>
>

--
Paul McCullagh
PrimeBase Technologies
www.primebase.org
www.blobstreaming.org
pbxt.blogspot.com