Hacker News Reader

Comment by zzzeek | Hacker News Reader

Great points. I'd love to see someone tackle how to get a corporate environment to consider Postgres over MS-SQL. We weren't able to, due to MS-SQL's great amount of "out of the box" functionality regarding automatic failover, mirroring, clustering, etc. PG of course can do all of this but to convince the DBAs to throw away their point-and-click interfaces so that we could start writing custom scripts and listeners from scratch wasn't really feasible.

View on HN

I don't have an answer to the issue of getting PostgreSQL in corporate environments - but I did run into a couple other hurdles in trying.

1. It's surprisingly difficult to find an experienced PostgreSQL DBA who wants to work in a corporate environment.

2. It's even more difficult to get a large corporation to pay a reasonable salary (why would we pay more for a PostgreSQL DBA, when we can just throw that money at Oracle/Microsoft and get a cheaper DBA with all the certs...) which might entice an experienced PostgreSQL DBA to join the organization.

I would guess that #1 is caused by #2... does #2 mean that Oracle+DBA (or MSSQL+DBA) is cheaper than a Postgres DBA, or just that the database and the DBA are generally paid for out of different department's budgets?

I think it's correct that (2) directly influences (1). And I think that most corporate groups are willing to hire an Oracle/MS SQL DBA with small amounts of experience but the carrying the right certifications.

To expand a bit, IMO - that scenario doesn't really exist for PostgreSQL DBAs, you are either experienced or inexperienced (with no weight-carrying certificate to complement credentials). Hence the corporate environment won't hire an entry-level DBA who has no certs to validate the hire, and they won't justify the seemingly added cost of an experienced PostgreSQL DBA.

Right now, it's incredibly difficult to find good PostgreSQL talent. One of those reasons is because a lot of postgres skills are interchangeable with MS-SQL, and Oracle (or insert enterprise RDBMS of choice here), and they pay much better for less complex work.

You could always hire someone like command prompt to do it for you (i'm pretty sure they already have all this stuff pre-written).

heh..actually..it's the first thing on their product list: http://www.commandprompt.com/products/hot_standby_drbd/ ... they also do clusters, etc.

You won't have much luck. MS based corporate outfits like to hire ready trained monkeys so you will get very little buy in from management.

I've taken 5 years and several critical arguments to get a couple of MySQL instances in one of them to run some internal stuff. It's a start at least.

repmgr is one option but its still command line. In my opinion Postgres is more likely to gain traction in Unix environments where it can replace MySQL pretty much anywhere and Oracle in quite a few use cases.

I like Postgres, however in my experience, where money is not a consideration, Oracles RAC with its (fairly) easy multi node DB clusters is often the competitive advantage that prevents Postgres from being used more widely in the enterprise - maybe that could be something that the Postgres team look at replicating.

Ignoring the rip on 'point and click' UIs, MSSQL provides a lot more out of the box that PG does not do. MSAS, SSIS, and SSRS are 3 big ones right off the top. For data analysis the plugin that lets Excel be a pivot table UI backed by MSAS is very cool.

we could start writing custom scripts and listeners from scratch

Let me explain the DBA mindset: "when this falls over at 2am the day before <major, non-negotiable deadline>, who is responsible for fixing it?"

Now I am not saying Postgres is flaky; I am saying that the DBAs operate under a different set of constraints than developers perhaps do. One common scenario is that devs are incentivized on new features delivered and DBAs/SAs are incentivized on uptime.

Reply on news.ycombinator.com