January 11, 2017 at 7:18 am
Coming at this from more a production DBA and less a developer DBA point-of-view, I'm rather ambivalent about SQL on Linux. To a large extent, I don't really care what the OS underpinning my instances is, as long as my customers get what they need out of the system. I will admit though, if we were to move to SQL-on-Linux, I'd want someone else to handle keeping the OS patched as my Linux skills are quite minimal (currently I'm responsible for also applying MS updates to the OS on my servers as well as SQL updates.)
I think where the lack of excitement might be coming from is more the "great, now I need to figure out how to get X working on a completely different OS when I just got it working the way I want on Windows."
As for MS taking a bottom-line hit on not selling OS licenses, a quick Google search finds Windows Server 2012 R2 Standard, 2 processor licenses, going for ~$600. I doubt they'll even notice the difference, even if every SQL instance was immediately switched over to SQL-on-Linux, considering all the other MS products that do still require Windows for the OS (SharePoint, Exchange, AD, etc.)
What ought to be interesting is to take a look at the breakdown of what OS SQL is installed on a couple years after Linux SQL is released (non-beta / CTP,) I'd be willing to bet there'll be a decent percentage of Linux installs.
January 11, 2017 at 7:19 am
SQL Server on Linux makes lots of sense for Azure databases, I don't see tons of use elsewhere.
January 11, 2017 at 8:05 am
A couple additional thoughts..
When I put my DBA hat on there is no way I'd even think about using something like this until somewhere around Rev. 3 and everyone else, including upper management, are on board. That has been the observed pattern right? 😉 That requires "bubble-up buzz" that reaches the point where the marketing folks see opportunity.
So, in order to get more quickly to Rev. 3 in order to gain (and preserve) market share we need to compress time. Money compresses time in most cases but this project would seem to be quite large and expensive; perhaps prohibitive in terms of catching up. Thomas Sowell in his fabulous book, "The Vision of The Anointed" which addresses the practical application of the economic truths of Adam Smith's invisible hand of the market noted that (my words) "300 million pairs of eyeballs will always be more correct that 535 pairs (Congress) of eyeballs."
Regarding the stability-related issues I'd also be willing to wager that MSFT's decision to put SQL 'out there' is really a great way to vet the product and speed up the development cycle time, correspondingly lowering development costs. Microsoft understands good wine comes from crushing grapes. What better place than the open source community who love not very many things more than taking Microsoft down a notch. This falls under the 'iron sharpens iron' or 'what does not destroy me makes me stronger' category.
With respect to market share this looks like Microsoft's time honored and well-honed strategy of redefining the market by "drawing a larger circle". Think Novell when they redefined 'network'. n'est–ce pas?
Microsoft has a long history of goofs, right Bob? 😛 Contrary to the very fiber and core of the DBA they are risk takers. They understand that turtles, in order to get anywhere must stick out their necks. Ask a good DBA to do this. :ermm:
With respect to Oracle, and I'd add specifically RAC they have a loooong way to go yet.
January 11, 2017 at 8:11 am
GeorgeCopeland (1/11/2017)
SQL Server on Linux makes lots of sense for Azure databases, I don't see tons of use elsewhere.
No doubt someone at this moment is trying to install SQL Server Express edition on an Android phone or a cluster of networked Raspberry Pies.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
January 11, 2017 at 8:14 am
My initial reaction and thought on SQL Server on *nix was along the same lines of Azure.
SQL Server engine on *nix opens up an easier transition for customers to begin stretching their services from the local *nix systems to Azure services. Ultimate end game currently for MS, imho, is all their customers on Azure. That's the long term achievement I see. That's were they're hanging their hat as the next bulk revenue source. They've realized OS domination is a futile venture in the future. There's a large *nix community out there and MS products on *nix via Core will expose all *nix devs/users to the Azure platform via an easier integration path. The battle is for cloud services.
I do believe eventually all data-stores for all businesses will eventually be in the cloud in some form or other, and the cost of data-stores in the cloud will become so cheap it will make no sense to build and maintain your own. Let the experts who built the data-stores manage it for you. We'll see. Interesting times. I like it :).
January 11, 2017 at 8:32 am
My reaction is totally "Meh". It might be fun to spin up a linux VM and install express to mess around with or do some minor website things.
I'm sure it's nifty for people who like Linux but I don't particularly care for linux and I don't see any advantages to it besides cost saving. If it helps some people migrate away from Oracle and MySql that's great but losing all the integration with the windows stack doesn't seem worth it to me in an enterprise setting.
January 11, 2017 at 8:42 am
My company is mostly Microsoft shop. We bought a company that ran Oracle. This was before my time. They moved Oracle to a Windows platform which is supported. Just not as well as Oracle running on Linux. When you call support, you don't always get someone who knows windows very well. Also best practices for one system are not best for all systems. Finally Microsoft and 3rd parties are going to focus their energies and money on the meat and potatoes which will be SQL on Windows for a long time. I don't think it is a bad idea for Microsoft to build it. I just worry about the poor people who have to support it in production 5-10 years from now if it doesn't take off.
January 11, 2017 at 8:55 am
IMHO, Microsoft is focused on tools, and not the OS. Early on, DOS and then Windows provided a platform to sell tools. Multiplan anyone? Much like razor blades and the free handle. And the focus on tools was the reason Bill Gates put money into Apple to save it- to sell more office tools. And now we have Office tools for iPads. Office 365 is in the cloud. Power BI -> cloud (but soon supported on-premises via SSRS).
All indications are that MSFT sees the cloud, and thus the tools as key to long term success. And SQL Server is a great tool - leverage it, don't contain it. And it is getting great placement in the Gartner reviews (like them or love them, CIO's use Gartner reports to validate their decision). It makes sense for them to go all out.
The SQLSvr team posted a write-up about how they did it.
SQL Server on Linux: How? Introduction
Way above my paygrade, but they are doing some great things. Let's support them.:-D
The more you are prepared, the less you need it.
January 11, 2017 at 10:53 am
Now, does SSMS or Toad work on Linux?
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
January 11, 2017 at 11:00 am
jasona.work (1/11/2017)
Coming at this from more a production DBA and less a developer DBA point-of-view, I'm rather ambivalent about SQL on Linux. To a large extent, I don't really care what the OS underpinning my instances is, as long as my customers get what they need out of the system.
I hope I made this point. SSoL is just SQL Server. From a DBA perspective (or even Dev), it mostly just looks the same.
January 11, 2017 at 11:04 am
Eric M Russell (1/11/2017)
Now, does SSMS or Toad work on Linux?
SSMS works from Win to a SSoL. It doesn't run (yet) on anything other than Windows. Native Linux tooling is very light (cmd line) right now). Not sure if we'll see SSMS ported. With VS on Mac, perhaps, but I don't know what the plans are.
There are other tools: http://alternativeto.net/software/sql-server-management-studio/?platform=linux
January 11, 2017 at 11:13 am
Whether people like it or not, this is going to open up new opportunities for DB's and IT professionals as a whole. For Microsoft, this is a stepping stone to bigger and better things.
January 11, 2017 at 11:18 am
Steve Jones - SSC Editor (1/11/2017)
jasona.work (1/11/2017)
Coming at this from more a production DBA and less a developer DBA point-of-view, I'm rather ambivalent about SQL on Linux. To a large extent, I don't really care what the OS underpinning my instances is, as long as my customers get what they need out of the system.I hope I made this point. SSoL is just SQL Server. From a DBA perspective (or even Dev), it mostly just looks the same.
Linux/Wine/SSMS looks like a maybe.
https://www.winehq.org/search?q=ssms
Also, perhaps Linux/VMware/Windows/SSMS
Ironically, some may end up using Oracle's Java based SQL Developer tool on Linux just to manage to SQL Server. I've used SQL Developer in the past to manage Oracle, and I experimented for a about one day with it's SQL Server management features before going back to SSMS.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
January 11, 2017 at 11:21 am
Eric M Russell (1/11/2017)
GeorgeCopeland (1/11/2017)
SQL Server on Linux makes lots of sense for Azure databases, I don't see tons of use elsewhere.No doubt someone at this moment is trying to install SQL Server Express edition on an Android phone or a cluster of networked Raspberry Pies.
I have two Raspberry Pis, A long-term project is to get SQL Server (developer edition probably, not Express) running on one. Just to say that I have.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
Viewing 15 posts - 16 through 30 (of 67 total)
You must be logged in to reply to this topic. Login to reply