a.mitani@xxxxxxxxxxxxxx wrote:
> About PGCluster-II,
> It is a shared storage clustering solution like Oracle RAC in roughly.
>
Yes, PGCluster-II will be a very nice solution as first disk-shared
opensource solution. It is to be regretted that i must start my project
now and not in a year =(
> ------------------------------
> At.Mitani
>
>> Pgpool-II, how i understand, can do parallel select query but not
>> update. May be pgpool2 can manage pools of clusters with assignment
>> different tables to different clusters? I don't find it on pgpool's
>> site. In any case, i must start solution in production in few months and
>> can't wait something stable =(
>> Now i try to make choose - multimaster PGCluster or master-slave Slony
>> =)
>> A little question about PGCluster. Is it work fine with OpenFTS or
>> other full index solutions?
>>
>>
>> a.mitani@xxxxxxxxxxxxxx ?????:
>>> Christian,
>>>
>>> Thank you for explain it.
>>>
>>> Ordinary, a replication is not help for improve performance of the data
>>> writing load.
>>>
>>> You can choice the parallel query as solution of it.
>>> I heard that "pgpool-II" is the tool for it.
>>>
>>> However, it is a mutual exclusive between HA and divide database (like a
>>> parallel puery).
>>> So, please make carefull choice.
>>>
>>> I hope that PGCluster-II might be a one of solutions for them.
>>> But it's still in progress.
>>>
>>> Regards,
>>> ---------------
>>> At.Mitani
>>>
>>>> Yes, you are right. If you have a lot of update queries, a cluster will
>>>> actually perform worse than a single server. You only get added
>>>> performance if you have more select than update queries. Same with
>>>> slony.
>>>>
>>>> If you want to improve performance on a database with lots of updates,
>>>> use a single db server and add some iron :-)
>>>>
>>>> Cheers,
>>>>
>>>> Christian
>>>>
>>>>
>>>> Dmitriy MiksIr wrote:
>>>>> Hello!
>>>>> At this time i'm try to make a choice between cluster solutions
>>>>> for
>>>>> Postgresql. The main aims is h.a. and scalability. With h.a. all
>>>>> clear,
>>>>> with scalability for select queries is too. But i'm not understand
>>>>> scalability on update queries. For Slony update server can be only one
>>>>> and it's bottleneck, for pgpool and pgcluster too, because each update
>>>>> going through all nodes... am i right? Is any solutions for increase
>>>>> update rate except hardwares upgrade?
>>>>> Thanks.
>>>>>
>>>>> _______________________________________________
>>>>> Pgcluster-general mailing list
>>>>> Pgcluster-general@xxxxxxxxxxxxx
>>>>> http://pgfoundry.org/mailman/listinfo/pgcluster-general
>>>>>
>>>> _______________________________________________
>>>> Pgcluster-general mailing list
>>>> Pgcluster-general@xxxxxxxxxxxxx
>>>> http://pgfoundry.org/mailman/listinfo/pgcluster-general
>>>>
>> _______________________________________________
>> Pgcluster-general mailing list
>> Pgcluster-general@xxxxxxxxxxxxx
>> http://pgfoundry.org/mailman/listinfo/pgcluster-general
>>
>
|
