Re: Why we lost Uber as a user

Something I didn't see mentioned that I think is a critical point: last 
I looked, HOT standby (and presumably SR) replays full page writes. That 
means that *any* kind of corruption on the master is *guaranteed* to 
replicate to the slave the next time that block is touched. That's 
completely the opposite of trigger-based replication.

On 8/3/16 3:51 PM, Kevin Grittner wrote:
> Personally, I can't imagine running logical replication of
> supposedly matching sets of data without something equivalent.

I think it depends heavily on the replication solution. I ran londiste 
for 6+ years with no major issues, but of course there was at least one 
other major company running that. I also took the time to completely 
read all the source code; something that's a reasonable prospect with a 
few thousand lines of python. For streaming rep it's difficult just to 
draw the line at where the code is.

Ultimately, people really need to understand the trade-offs to the 
different solutions so they can make an informed decision on which ones 
(yes, plural) they want to use. The same can be said about pg_upgrade vs 
something else, and the different ways of doing backups.

Something I think a lot of folks fail to understand is the value of 
having a system that has simple technology in the mix. Keeping something 
like londiste running has a non-zero cost, but the day you discover 
corruption has replicated through your entire infrastructure you'll 
probably be REALLY happy you have it. Similarly, I always encourage 
people to run a weekly or monthly pg_dump if it's at all feasible... 
just to be safe.
-- 
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Experts in Analytics, Data Architecture and PostgreSQL
Data in Trouble? Get it in Treble! http://BlueTreble.com
855-TREBLE2 (855-873-2532)   mobile: 512-569-9461