Re: root cause of corruption in hot standby

Hi,

On 2018-09-06 12:28:18 -0500, Mike Broers wrote:
> So I have discovered corruption in a postgres 9.5.12 read replica, yay
> checksums!
> 
> 2018-09-06 12:00:53 CDT [1563]: [4-1] user=postgres,db=production WARNING:
> page verification failed, calculated checksum 3482 but expected 32232
> 
> 2018-09-06 12:00:53 CDT [1563]: [5-1] user=postgres,db=production ERROR:
> invalid page in block 15962 of relation base/16384/464832386

Have you compared whether that block looks significantly different on
the primary from the standby with a hex editor?

Greetings,

Andres Freund