Re: [mysql-ha-devel] alpha-6.3-1

[Fernando I. <fi...@ad...>]

Yes, that would be great.

The things we use from the system shouldn't change too much from FC5 to=20
RHEL4.

As for MySQL 4.1, it will be good to know how this works with that=20
version. The only thing that changes from version to version and that cou=
ld
be broken in our code is the way we detect if the slave thread is running=
.

My recommendation would be to:

    * Skip the install.sh script, since it probably won't work on that
      system.
    * Build and install fake
    * Create a mysql-ha.conf file (you can use configuration-wrapper.sh
      for this)
    * Test the cluster by running ./configurator.sh directly on the
      slave, as I'm still working on the init script (it starts
      properly, but the stop is not reliable).


I don't have access to RHEL4, but I can get a test system with CentOS 4=20
by next week, and that  should be binary compatible.
If you're willing to test it on your setup I'll install CentOS on my=20
boxes and start working out the bugs as we come across them.

I can also write documentation on how to perform a manual installation=20
(skipping install.sh) if you think that will help you (I was planning to=20
do this anyway and this would
be a great opportunity to test the results).

Regards,

    Fernando.

Leandro Guimar=C3=A3es Faria Corcete DUTRA wrote:
> Em Mon, 30 Oct 2006 18:07:37 -0200, Fernando Ipar escreveu:
>
>  =20
>> 6.3-1 is out and two different failure scenarios are handler properly=20
>> under my FC5 w/MySQL5 test boxes.
>>    =20
>
> 	Do you think I could test it with RHEL4 and MySQL 4.1?
>
>  =20