Re: Replication mode
От | Mark Kirkwood |
---|---|
Тема | Re: Replication mode |
Дата | |
Msg-id | 486c6844-6e55-0151-449c-6ccf5dcc3cf6@catalyst.net.nz обсуждение исходный текст |
Ответ на | Re: Replication mode (Azimuddin Mohammed <azimeiu@gmail.com>) |
Список | pgsql-admin |
On 11/01/18 14:10, Azimuddin Mohammed wrote: > :-) > Our fear is using open source tools for HA creates dependency and > single point of failure(pgpool). Or say pacemaker which our enterprise > do not recommend. > We are planning to use postgres to use for our hadoop metastore which > is presently on Oracle. > Ok, I understand. The usual way this single point of failure thing is handled is to use something like a pair of HAProxy servers with redundant network switches (i.e make the HA manager itself HA). While this is a bit fiddly it is reasonably well documented and standard (experienced system and network admins should have no difficultly getting it to work). There are also specialized Postgres related solutions for HA (e.g repmgr from 2ndquadrant) that might be worth a look as well. regards Mark
В списке pgsql-admin по дате отправления: