> I know for several months, I have recommended the > latest developer build of osol, as long as you're not using it in production. > But if you *are* using it in production, I recommend paying for sol10. > (Better yet, sun hardware ith sol10.)
I do wish people would remember that there's a spectrum in between "top tier support" and "no support at all". I have Sun gear running Solaris with Oracle databases on them. I want, and have, full support contracts on that gear. It only makes sense. However, before the full Oracle takeover, I was deploying Solaris onto the minor stuff... my mail hubs, for example, or the low-end web servers. I certainly hope I wouldn't need to explain to anyone on this list why Solaris would be the bee's knees for these tasks. It helps me maintain some homogeneity of environment, and more importantly it gives my coworkers a place where they can practice, a place where it's reasonably important that they watch what they do, but crashing the mail server isn't quite the same disaster as wiping out the payroll database server. I DON'T need full Oracle support on those machines. Because they're net-facing, I DO need updates. I can't justify $1k per year per machine (Oracle's quoted price) for access to security patches... my managers will - and have - told me to just go with Windows 2008R2, which costs us ~$500 one-time charge with OS updates for free for the rest of the serviceable lifetime of the product, or my coworkers tell me to switch to Redhat Linux, for half that again. So Oracle has priced itself out of the small/medium internet services market. I can cope with that; that market is what OpenSolaris is for. However, before I can start deploying OSol, I need *proof* that it's going to continue. It's important to remember that - from a certain point of view - Oracle has NEVER put out an OpenSolaris release. Have they said they plan to? Yes, in uncertain terms. Have others said they're going to? Yes, in uncertain, "I can't speak authoritatively for Oracle" terms. I've seen the words "have faith" thrown around, which is completely ridiculous... nobody makes business decisions based on "faith" except a church. So I need concrete proof of a future OpenSolaris release, and I need it in the form of a binary release, even to /dev. My coworkers are by no means dummies, but expecting them to build the distribution from source is unreasonable, not the least of which because it's a messy, convoluted process. They've got things that need to be done, and they can flatten and rebuild using Linux (or Win2008) in less time. And to explain why that's a bad thing: once that happens, not only do I lose a *Solaris server, *they* lose a reason to learn Solaris at all. And why is that bad? Because when it comes time to refresh our database kit, my very-upper-management - who is already very pro-Microsoft as it is - is going to look at the fact that they have ONE Solaris expert, a bunch of Linux experts, a bunch of Windows experts, weigh the balances, and switch us over to SQL Server. I'm not speaking from FUD here... this is *exactly* the same circumstances that caused my organization to kick IBM to the curb and switch us to Sun kit eight years ago. But Oracle has one advantage IBM doesn't; a very-near-free version of their enterprise OS. *All* I need to stave off the switch is a binary OpenSolaris release. It doesn't even need to be a *good* release. It can be buggy as hell; I simply won't upgrade to it. But it will be a tactile, hold-it-in-your-hands proof of Oracle's commitment. I don't need / repo updated, I need /dev updated. To stave off a depressing series of events, I need to deploy a small, unimportant server using OSol, *without* my manager and coworkers looking at me as if I've just suggested deploying a server running BeOS. I haven't ranted about this here before, because I *know* this list is filled with engineers who can't make the release happen, who can't adjust Oracle's marketing efforts. I've been enjoying the fruits of their labours for a long time now, and I definitely don't want to increase their frustration. But I've seen multiple people present three options: build OSol from source; run "old" OSol; buy a full Solaris support contract. In my environment, none of those is particularly viable. "Have faith" isn't a choice at all. -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
