SET PASSWORD is indeed not replicated. However, GRANT commands are
replicated and this is not consistent behavior, so we probably need to
implement SET PASSWORD replication. I will file a LP bug tracker for
this task.
This is a bit more complicated to deal with, as the SET command can
have several variables and we are supposed to replicate only PASSWORD
part of the command.
Thanks for your answer.
Yes, I was thinking of such a problem since SET can be use in many different cases.
Could you please post the LP bug entry URL when you create it so we will be able to know when the problem will be fixed ?
Thks a lot !
Laurent
Le mardi 29 mai 2012 16:10:59 UTC+2, Seppo Jaakola a écrit :
> SET PASSWORD is indeed not replicated. However, GRANT commands are > replicated and this is not consistent behavior, so we probably need to > implement SET PASSWORD replication. I will file a LP bug tracker for > this task. > This is a bit more complicated to deal with, as the SET command can > have several variables and we are supposed to replicate only PASSWORD > part of the command.
> Thanks for your answer.
> Yes, I was thinking of such a problem since SET can be use in many
> different cases.
> Could you please post the LP bug entry URL when you create it so we will be
> able to know when the problem will be fixed ?
> Thks a lot !
> Laurent
> Le mardi 29 mai 2012 16:10:59 UTC+2, Seppo Jaakola a écrit :
>> Hi Laurent,
>> SET PASSWORD is indeed not replicated. However, GRANT commands are
>> replicated and this is not consistent behavior, so we probably need to
>> implement SET PASSWORD replication. I will file a LP bug tracker for
>> this task.
>> This is a bit more complicated to deal with, as the SET command can
>> have several variables and we are supposed to replicate only PASSWORD
>> part of the command.
Do you know when this bug will be fixed please or at least do you know of a workaround to avoid it please ?
In a future production environment and to do a presentation of the product with such case in the test list, this is a little bit problematic that such a "simple" feature is not yet supported/implemented ... :)
Do you think it could be integrated in the next release ?
Thanks ! Regards,
Laurent
Le jeudi 31 mai 2012 10:03:16 UTC+2, Seppo Jaakola a écrit :
> > Thanks for your answer. > > Yes, I was thinking of such a problem since SET can be use in many > > different cases. > > Could you please post the LP bug entry URL when you create it so we will > be > > able to know when the problem will be fixed ?
> > Thks a lot !
> > Laurent
> > Le mardi 29 mai 2012 16:10:59 UTC+2, Seppo Jaakola a écrit :
> >> Hi Laurent,
> >> SET PASSWORD is indeed not replicated. However, GRANT commands are > >> replicated and this is not consistent behavior, so we probably need to > >> implement SET PASSWORD replication. I will file a LP bug tracker for > >> this task. > >> This is a bit more complicated to deal with, as the SET command can > >> have several variables and we are supposed to replicate only PASSWORD > >> part of the command.
SET PASSWORD replication has been pushed in development branch and
will be part of next regular release build, which now looks to happen
during next week.
> Do you know when this bug will be fixed please or at least do you know of a
> workaround to avoid it please ?
> In a future production environment and to do a presentation of the product
> with such case in the test list, this is a little bit problematic that such
> a "simple" feature is not yet supported/implemented ... :)
> Do you think it could be integrated in the next release ?
> Thanks !
> Regards,
> Laurent
> Le jeudi 31 mai 2012 10:03:16 UTC+2, Seppo Jaakola a écrit :
>> > Thanks for your answer.
>> > Yes, I was thinking of such a problem since SET can be use in many
>> > different cases.
>> > Could you please post the LP bug entry URL when you create it so we will
>> be
>> > able to know when the problem will be fixed ?
>> > Thks a lot !
>> > Laurent
>> > Le mardi 29 mai 2012 16:10:59 UTC+2, Seppo Jaakola a écrit :
>> >> Hi Laurent,
>> >> SET PASSWORD is indeed not replicated. However, GRANT commands are
>> >> replicated and this is not consistent behavior, so we probably need to
>> >> implement SET PASSWORD replication. I will file a LP bug tracker for
>> >> this task.
>> >> This is a bit more complicated to deal with, as the SET command can
>> >> have several variables and we are supposed to replicate only PASSWORD
>> >> part of the command.
>> >> > Is this normal that this query is not propagated/replicated over the
>> >> whole
>> >> > cluster please :
>> >> > SET PASSWORD FOR 'root'@'localhost' = PASSWORD( '********' )
>> >> > This change the local node password on which I'm currently connected
>> but
>> >> > does not update the password on the others connected node ?
> SET PASSWORD replication has been pushed in development branch and > will be part of next regular release build, which now looks to happen > during next week.
> > Do you know when this bug will be fixed please or at least do you know > of a > > workaround to avoid it please ? > > In a future production environment and to do a presentation of the > product > > with such case in the test list, this is a little bit problematic that > such > > a "simple" feature is not yet supported/implemented ... :) > > Do you think it could be integrated in the next release ?
> > Thanks ! > > Regards,
> > Laurent
> > Le jeudi 31 mai 2012 10:03:16 UTC+2, Seppo Jaakola a écrit :
> >> > Thanks for your answer. > >> > Yes, I was thinking of such a problem since SET can be use in many > >> > different cases. > >> > Could you please post the LP bug entry URL when you create it so we > will > >> be > >> > able to know when the problem will be fixed ?
> >> > Thks a lot !
> >> > Laurent
> >> > Le mardi 29 mai 2012 16:10:59 UTC+2, Seppo Jaakola a écrit :
> >> >> Hi Laurent,
> >> >> SET PASSWORD is indeed not replicated. However, GRANT commands are > >> >> replicated and this is not consistent behavior, so we probably need > to > >> >> implement SET PASSWORD replication. I will file a LP bug tracker for > >> >> this task. > >> >> This is a bit more complicated to deal with, as the SET command can > >> >> have several variables and we are supposed to replicate only > PASSWORD > >> >> part of the command.
> >> >> > Is this normal that this query is not propagated/replicated over > the > >> >> whole > >> >> > cluster please :
> >> >> > SET PASSWORD FOR 'root'@'localhost' = PASSWORD( '********' )
> >> >> > This change the local node password on which I'm currently > connected > >> but > >> >> > does not update the password on the others connected node ?
