On Fri, Jun 11, 2010 at 1:48 AM, Josh Berkus <josh@agliodbs.com> wrote:
> On 06/09/2010 07:36 PM, Mark Kirkwood wrote:
>>
>> On 10/06/10 14:07, Tatsuo Ishii wrote:
>>>
>>> The one of top 3 questions I got
>>> when we propose them our HA solution is, "how long will it take to
>>> do failover when the master DB crashes?"
>>>
>>
>> Same here +1
>
> In that case, wouldn't they set max_standby_delay to 0? In which case the
> failover problem goes away, no?
Yes, but I guess they'd also like to run read only queries on the standby.
Setting max_standby_delay to 0 would prevent them from doing that because
the conflict with the replay of the VACUUM or HOT record would often happen.
vacuum_defer_cleanup_age would be helpful for that case, but it seems to be
hard to tune that.
Regards,
--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center