0

I’m looking for some information about whether PostGreSQL EDB can provide balance between nodes similarly to the way Oracle RAC does this.

AFAIK, it does not.

The closest question / answer combination I could find to this subject follows:

https://stackoverflow.com/questions/1498793/does-oracle-rac-allow-completely-transparent-failovers-between-nodes

However, that question and its answers at the above link discuss failovers, not load balancing.

I've also reviewed the e-booklet here:

http://www.enterprisedb.com/postgres-plus-edb-blog/gary-carter/comparing-edb-postgres-and-oracle

The ol' internet and a search on serverfault.com have not given me much joy, so far. I keep coming up with generic guidelines or corporate leaflets (without detailed examples of load balancing behavior) on either product.

So, for any of you who might have used PostGreSQL EDB, does it provide load balancing among nodes in a similar way to Oracle RAC?

Craig Ringer
  • 10,553
  • 9
  • 38
  • 59
Nathan Basanese
  • 321
  • 1
  • 4
  • 19
  • // , To me, it's pretty obvious if tool A provides feature α in a similar way to the way that tool B provides feature α. I think this is sufficiently specific (and brief). However, if it's not, at risk of being long winded, I would love to provide a list and one or two behavior examples. – Nathan Basanese Nov 18 '16 at 20:40

1 Answers1

1

Oracle RAC is shared-storage clustering. PostgreSQL doesn't do that.

As far as I know neither does PPAS, EDB's product. I work for a competing vendor and I don't use their products routinely, but I'm pretty sure I would've heard if they'd added shared-storage clustering.

The best way to be sure is to ask EDB directly. They have forums, they don't really hang out here.

There are other products and tools that provide various load-balancing, sharding and distribution options for PostgreSQL. Look into Postgres-XL, Citus, BDR, PgPool-II, EDB's failover manager, AWS Redshift, EDB's xDB, etc. All have advantages and disadvantages. None use the same model as Oracle RAC.

Shared-storage clustering is in my opinion not great anyway. It places a huge amount of reliance on a SAN, and its performance scaling is mediocre at best. Most PostgreSQL clustering work is focusing on replication based clustering with query routing and scatter/gather querying, sharding, distributed lock management/snapshot management/transaction management, etc. But then, I work on a replication based clustering solution (BDR) so I would say that.

Craig Ringer
  • 10,553
  • 9
  • 38
  • 59
  • // , Great to hear this from the other horse's mouth. I will consider this my answer. I don't know that much about either PPAS or products by competing vendors. I think this will give me a base. I only ask here because EDB has yet to respond, and I think someone like me may need an answer to this question in the future. I'll check the EDB forums. – Nathan Basanese Nov 21 '16 at 21:39
  • Why do you prefix everything with `// , `? It's ... very confusing. – Craig Ringer Nov 22 '16 at 00:42